Terravision Technical Papers.pdf

Embed Size (px)

Citation preview

  • 8/9/2019 Terravision Technical Papers.pdf

    1/55

  • 8/9/2019 Terravision Technical Papers.pdf

    2/55

    Pacifico Minerals 22 Sept 2014 2

    When deriving an equation for subsurface radio location we will neglect the member BV in (7), but in equation (4) we will consider 1. Practically for any subsurface

    areas these assumptions are true. Wave equation for electric multiplier E can beobtained in the following manner. Let us differentiate equations (1) and (7) accordingto time t , and equation (2) according to spatial values, using differential operator ,after this we will obtain:

    t E

    t E

    t H

    2

    2

    0

    t H

    E 02

    Out of this equations follows a wave equation for electric multiplier E ofelectromagnetic wave in the environment, characterized by permittivity and

    conductivity :

    0022

    22

    t E

    t E

    c E

    (8)

    Here00

    1

    c - speed of light in vacuum.

    3. Total-differential approximation of wave equation.Let us use a total-differential approximation of differential operators and write fordigital function U , which will correspond to the multiplier of electric vector E , adigital analogue of 1-D wave equation (8) in t x, coordinates.

    02)(

    2)(

    2 1102

    11

    2211

    t

    U U

    t

    U U U

    c x

    U U U t xt

    xt

    xt

    xt

    xt

    xt

    x

    (9)

    Equation (9) corresponds to a clear conditionally stable finite-difference scheme ofsecond-order precision on analytical grid with a step x along solid axis x and with astep t along time axis t . Superscripts on wave function U represent number of mesh

    points on time axis, subscripts – numbers of mesh points on spatial axis.

    Wave function on time level 1t is represented as a function in previous timemoments like:

    2)(

    122)()(

    2)(

    12

    01112

    22201 ct U U U U U

    xt cct

    U t xt

    xt

    xt

    xt

    xt

    x (10)

    The solution of equation (10) is stable upon fulfillment of condition [2]:

  • 8/9/2019 Terravision Technical Papers.pdf

    3/55

    Pacifico Minerals 22 Sept 2014 3

    c

    xt

    )( (11)

    Below we provide a program for calculation of wave field U , made in programmingsystem “MATLAB”, which makes calculations according to the formula (10) with anautomatic selection of step (11).

    %------- ONE DIMENTION WAVE EQUATION%------- FOR ELECTRIC FIELD E(x,t)clear

    packclose all%---------PARAMETERS FOR NUMERICAL CALCULATIONX=3; %---- max depth, mT=40e-9; %---- max time, secdt=0.05e-9; %---- step on T, secXP=0.8; %---- disdant of crossection, m%-----------PARAMETERS OF THE SOIL

    D1=1.5; %---- depth of the 1-st layer, meps1=6; %---- permitivity of the 1-st layersig1=0.; %---- conductivity of the 1-st layer, sim/mDelta=0.; %---- width of the boundary, meps2=12; %---- permitivity of the 2-nd layersig2=0.; %---- conductivity of the 2-nd layer, sim/m%------ calculation of the dx-step from stability conditionc=3e8; %---- light velocity, m/secmu0=4*pi*1e-7; %---- magnetic permeabilityemin=eps1;if eps2

  • 8/9/2019 Terravision Technical Papers.pdf

    4/55

    Pacifico Minerals 22 Sept 2014 4

    sig(cnx)=sig1;endfor cnx=1:nD

    eps(cnx+n1)=eps1+(eps2-eps1)/nD*cnx;sig(cnx+n1)=sig1+(sig2-sig1)/nD*cnx;

    endfor cnx=n1+nD:NX

    eps(cnx)=eps2;sig(cnx)=sig2;end%----------------------------- count of the wave equationfor cnt=2:NT-1for cnx=2:NX-1

    E(cnt+1,cnx)=c^2/eps(cnx)*dt^2/dx^2*(E(cnt,cnx+1)-2*E(cnt,cnx)+E(cnt,cnx-1))...+2*E(cnt,cnx)-E(cnt-1,cnx)*(1-sig(cnx)*mu0*dt*c^2/2/eps(cnx));

    E(cnt+1,cnx)=E(cnt+1,cnx)/(1+sig(cnx)*mu0*dt*c^2/2/eps(cnx));endend

    %-----------visualisation of the results%-------- Fig.1 - 2D electrical field E(x,t), eps(x)*1e-9%-------- Fig.2 - electrtrical field E(t) at distant XP

    t=(0:NT-1)*dt;x=(0:NX-1)*dx;

    figure pcolor(x,t,E(:,:))shading interpcolorbarhold on

    plot(x,eps*1e-9,'w')xlabel('X, m')ylabel('T, sec,eps*1e-9')hold off

    pause(3)amp=zeros(NT);xs=round(XP/dx);

    for cnt=1:NTamp(cnt)=E(cnt,xs);

    endfigure

    plot(t,amp,'k')hold onxlabel('T. sec')hold off

  • 8/9/2019 Terravision Technical Papers.pdf

    5/55

    Pacifico Minerals 22 Sept 2014 5

    Pic.1

    Reflection of signal from sharp edge between a layer with 61 and a layer with 122 .

    Pic.2Wave function of electric field is at depth of 0.8 m. The second signal, reflected from a more dense

    layer, has a reverse polarity against the former decreasing one.

  • 8/9/2019 Terravision Technical Papers.pdf

    6/55

    Pacifico Minerals 22 Sept 2014 6

    The results of calculations are provided in a graph in pic.1-2.

    Pic.1 shows a wave function in xt , coordinates, function amplitude is represented by acolor. A ratio between function values and color is shown I the right part of the

    picture. On the same picture a white line additionally indicates to a dependence of permittivity with depth. In this case, time axis t represents its values with re-

    calculation coefficient, shown on this axis.

    Pic.2 shows a section of bivariate function of amplitude in relative units for the depth,assigned in the program by a parameter XP.The program allows assign two layers with different environment parameters , .Width of edge between layers is defined by a parameter Delta. In and around the edge

    permittivity and conductivity is assigned by means of linear interpolation. Undercondition when Delta=0 environment parameters change suddenly.

    In order to necessarily limit calculation zone along x -axis, zero limit conditions areassigned on edges under 0 x and x X conditions, which is equivalent to assigningan ideal reflecting edge on these intervals. Signals reflected from these edges can also

    be viewed on the pictures.

    4. Reflection and transition of radio-impulse signal through sharp edge of sectionof two non-conducting environments.

    Advancing wave experiences reflection only from those sections of subsurfaceenvironment, where space variation of its parameters are observed: permittivity ,conductivity or both values simultaneously.

    Upon normal fall of wave to a sharp edge reflection coefficient in non-conducting hasthe following look

    21

    21

    R (12)

    A value n is called a refractive index. It defines velocity of wave propagation V in material environment

    nc

    V (13)

    It is derived from (12) that refractive index can be both positive and negative. If21 , which means that wave falls from a less transmission dense environment to a

    more dense one (for instance, from dry sand to clay), refractive index will be negative.Upon 21 it has a positive value. In the first case a signal reflected from an edge

  • 8/9/2019 Terravision Technical Papers.pdf

    7/55

    Pacifico Minerals 22 Sept 2014 7

    “turns over”, that is, multiplied by a negative value, and in the second case, it repeatsthe shape, which it had before reflection.

    (A term “transmission density” in optics means refractive index. We will use this termhereinafter, omitting a definition “optical”.)

    The first case (fall to the edge with more dense environment) is shown on pic.1-2. On pic.2 one can clearly see that the second (reflected from the edge at the depth of 1.5m) signal is “turned over” with respect to the first falling signal. The third signalappeared because of reflection from an upper edge and is “turned over” twice, that is,it is located in phase with the falling one. The results shown on these pictures fullycorrespond to the text of the program provided above, including variable parameters.

    Pic.3-4 show the second case, which is a wave fall on the edge from more denseenvironment to less dense environment. Here layers change places: 6,12 21 . It is

    seen on picture 4 that reflected signal has the same polarity (or “phase”) as the fallingone.

    Pic.3Reflection of signal from a sharp edge between layer with 121 and a layer with 62 .

  • 8/9/2019 Terravision Technical Papers.pdf

    8/55

    Pacifico Minerals 22 Sept 2014 8

    Pic.4Wave function of electric field at the depth of 0.8 m. Upon falling of wave from more dense

    environment to less dense one, a signal reflected from the edge (the second one in the picture) hasthe same polarity as the falling one.

    Transmission coefficient T upon vertical fall in non-conducting environment has thefollowing look:

    21

    22

    T (14)

    It should be noted that refractive index R (12) and transmission coefficient T (14) areconnected with a correlation 1T R .

    From the formula (14) it should be noted that a passed impulse always keeps the polarity of the falling one, wherein one can make sure on pictures 5 and 6 for the twocases of wave fall reviewed earlier.

  • 8/9/2019 Terravision Technical Papers.pdf

    9/55

    Pacifico Minerals 22 Sept 2014 9

    Pic.5

    Wave function of an impulse passed through the edge at the depth of 2 m. Falling from less dense tomore dense environment (pic.1). Impulse polarity did not change.

    Pic.6Wave function of impulse passed through the edge at the depth of 2 m. Falling from more dense to

    less dense environment (pic.3). Impulse polarity did not change.

  • 8/9/2019 Terravision Technical Papers.pdf

    10/55

    Pacifico Minerals 22 Sept 2014 10

    5. Reflection and transition of radio-impulse signal through smooth edge ofsection of two non-conducting environments.

    A definition of sharp or smooth transmission edge between layers is defined as a ratio between linear size of transition region l and typical spatial oscillation size of directimpulse at the moment of transition through the edge. If l , we have a case of

    sharp edge. If the dimension of transition region is commensurable or higher thantypical oscillation size of direct impulse – it is the case of smooth edge. In geo-radiolocation, when has a value of unit – tens of centimeters, for geological layersthe case of smooth edge often occurs (for instance, transition of sand into clay andvice versa), which is conditioned by geologic processes. A sharp edge is observedmainly upon reflection from artificial subsurface items and constructions.

    In the case of smooth edge, even upon normal wave fall, simple well-known formulas(12), (14) are not always applicable even for evaluative calculations. A theory of

    transmission and reflection of impulse from a smooth edge of arbitrary shape is quitecomplicated, it can be found, for example, in monograph [3], we will not go beyondverification of major quality properties on the basis of computation modeling. Let ustake already reviewed case of wave fall on sharp edge between layers with 61 and

    122 as a basis (pic.1-2, 5). All parameters of the problem remain former, we willonly change the width of edge, defined by the parameter Delta. Let Delta=0.3 m.

    Pic.7Reflection and transition of signal through smooth edge with a width of 0.3 m.

    It follows from the pic.8 that a smooth edge reflects a signal, transforming its shape.These distortions are defined as a specific type of transition function. A common trait

  • 8/9/2019 Terravision Technical Papers.pdf

    11/55

    Pacifico Minerals 22 Sept 2014 11

    for all types of smooth edge functions is “stretching” of reflected impulse over time,which should be considered important in geo-radar data.

    Practically always signals coming from under the ground are longer than directimpulse. Most often the reason of this event is a reflection from smooth edge. Othermechanisms of impulse “stretching”, which will be mentioned below, are also

    possible.

    Comparing pic.9 for impulse passed through smooth edge and pic.5 for impulse passed through sharp edge, we can talk about their practical identity by shape. It isextremely important during analysis of geo-radar data. The point is that direct impulsein the general case may in series reflect from multiple layers. Keeping the shape ofimpulse, which passed through the next layer, allows analyze the look of specificreflecting edge irrespective of neighboring edges (we are not talking about signalamplitude and time delay here, which are defined not only by the edge, but also by

    parameters of homogeneous layer).

    Pic.8Wave function of signal reflected from a smooth edge at the depth of 0.8 m. The second signal

    reflected from a smooth edge has a reverse polarity as in a sharp edge between layers. The shape ofreflected signal has changed, it has “stretched” over time.

  • 8/9/2019 Terravision Technical Papers.pdf

    12/55

    Pacifico Minerals 22 Sept 2014 12

    Pic.9Wave function of signal passed through smooth edge at the depth of 2 m. Polarity and shape of the

    impulse almost did not change.

    6. Influence of environment conductivity to radio-wave propagation.

    Environment conductivity exists in wave equation (8) as a coefficient before thefirst derivative of time, which, as it is well-known, determines dissipation (absorption)of wave energy. As a matter of fact, it is the dissipative term, which limits allcapabilities of subsurface radiolocation, including penetration depth and spatialresolution (upon 0 radio-waves in the environment do not attenuate, which meansa capability of reaching any depth and any resolution).

    In order to clarify the features of radio-wave propagation in conductive environmentlet us move on from a time domain to frequency domain, since it simplifies furthercomputations. With this purpose let us seek the solution of wave equation (8) in theform of attenuating monochromatic wave

    )exp( t iikx px E (15)Here p - attenuation coefficient, k - wave number, - circular frequency.

    Let us put (15) into initial equation (8) and we will obtain a new equation, connectingwave parameters (15) with environment conductivity .

  • 8/9/2019 Terravision Technical Papers.pdf

    13/55

    Pacifico Minerals 22 Sept 2014 13

    02 022

    22

    i

    cipk k p (16)

    Equating of real and imaginary parts (16) to zero leads a biquadratic equation

    04

    220

    22

    2

    24 p

    c p (17)

    The solution of equation (17) is written as follows

    442

    2220

    4

    42

    2

    2

    cc p (18)

    Marks before radicals are selected in the condition of physical feasibility of results.

    Let us consider two asymptotics of function (18): low-frequency ( 0 ) and high-frequency ( ).

    In the area of low frequencies attenuation coefficient is expressed by the formula

    20 p (19)

    In the area of high frequencies attenuation coefficient is expressed by the formula

    20 c p (20)

    In these two extreme cases environment conductivity differently influences radio-wave propagation.

    In the area of low frequencies attenuation depends on frequency, and while itsreduction can be arbitrarily small. But in this range broadband signals experience

    dispersion, that is, distortion of their shape.

    In the area of high frequencies attenuation does not depend on frequency, whichduring radio-impulse propagation leads to reduction of its amplitude only withoutdistortion of shape.

    Frequency differentiating these two radio-wave propagation mechanisms depends onconductivity and permittivity of environment:

    0

    2

    c (21)

  • 8/9/2019 Terravision Technical Papers.pdf

    14/55

    Pacifico Minerals 22 Sept 2014 14

    or for cyclic frequency 2/ f

    91018

    f (21 а )

    For example, when msim /10 2 and 10 , differentiating frequency is MHz f 18 .Such parameters of environment are typical for the vicinities of Moscow.

    Let us bring formula (20) to more usual look, expressing long attenuation W indecibels per meter.

    16372W (20a)

    In the real situation, rocks constituting geological layers may have parameters , dependent on frequency, which should be taken into account while evaluating wave

    attenuation [4].

    Let us show on numerical examples the influence of environment conductivity to propagation of electromagnetic impulse. Let us assign a uniform environment in thecalculation area with permittivity 6 and watch the shape of impulse, passing at thedepth of 0.8 m, against conductivity .

    Pic.10-11 give an example of unattenuated propagation ( 0 ).

    Pic.12-13 show “high-frequency” feature of influence of conductivity to impulseshape. Here conductivity influences mainly its amplitude, impulse shape remains practically without changes.

    Pic.14-15 demonstrate influence of strong conductivity of “low-frequency” type,leading not only to amplitude reduction, but also to a strong impulse distortion, which

    becomes apparent in the form of appearing long “stretches”.

    Such practice is quite often observed in practice, when conductivity sharply increases,

    for example in damp saline soils. A long “tail” from upper edges of conducting layersin this case may disguise signals from lower edges, if geo-radar does not havesufficient number of quantizing levels over amplitude. One of the possible ways outfrom this situation – transition to shorter sounding impulse, for which a “high-frequency” propagation mechanism is performed, when “tails” do not occur. Inmajority of geo-radars it is performed by a replacement of antennas with higherfrequency.

  • 8/9/2019 Terravision Technical Papers.pdf

    15/55

    Pacifico Minerals 22 Sept 2014 15

    Pic.10

    Impulse advancing in uniform environment without attenuation ( 0 )

    Pic.11Impulse shape in uniform environment without attenuation at the depth of 0.8 m.

  • 8/9/2019 Terravision Technical Papers.pdf

    16/55

    Pacifico Minerals 22 Sept 2014 16

    Pic.12

    Impulse advancing in uniform environment with conductivity msim /01.0 .

    Pic.13

    Impulse shape at the depth of 0.8 m. Conductivity msim /01.0 for the assigned impulse leadsonly to reduction of amplitude, its shape practically does not change.

  • 8/9/2019 Terravision Technical Papers.pdf

    17/55

    Pacifico Minerals 22 Sept 2014 17

    Pic.14

    Impulse advancing in uniform environment with msim /05.0

    Pic.15Impulse shape at the depth of 0.8 m. Conductivity msim /05.0 severely distorts impulse shape,

    low-frequency “tail” appeared on it.

  • 8/9/2019 Terravision Technical Papers.pdf

    18/55

    Pacifico Minerals 22 Sept 2014 18

    As we already mentioned, change of conductivity just like the change of permittivity leads to reflection of radio-waves.

    Here, just as during change of permittivity, similar laws act: the impulse passedthrough the edge does not change its polarity, the impulse reflected from the edgewith more conducting environment changes its polarity, and the impulse reflected

    from the edge with lower conductivity also does not change its polarity.

    But one should remember that processes of reflection and passing through the edgehere occur in the background of processes of attenuation and dispersive distortionsmentioned above.

    Let us show numerical examples of wave reflections from edges of conductivitychange. In these examples permittivity of both layers is assigned by identical 6 .The edge between layers runs at the depth of 1.5 m.

    Pic.16Fall of impulse from the environment with conductivity msim /01.0 into non-conductive

    environment.

    It is seen on pic.16-17 that reflected signal has a polarity of the falling one, since itis reflected from less conductive environment.

    Pic.18-19 demonstrate a different case – falling onto the layer with highconductivity. Here reflected signal has changed its polarity with respect to the

    falling one.

  • 8/9/2019 Terravision Technical Papers.pdf

    19/55

    Pacifico Minerals 22 Sept 2014 19

    Pic.17

    Shape of reflected impulse at the depth of 0.8 m. During the fall of impulse from conductiveenvironment into non-conductive, its polarity remains unchanged. Reflected signal is more

    distorted than the first falling one, since it interacts longer with the conductive environment.

    Pic.18Fall of impulse from non-conductive environment into conductive with msim /05.0 .

  • 8/9/2019 Terravision Technical Papers.pdf

    20/55

    Pacifico Minerals 22 Sept 2014 20

    Pic.19

    Shape of reflected impulse at the depth of 0.8 m. It has reverse polarity as compared to thefalling (first) impulse.

    7. Conclusion.Let us formulate in brief main laws of electrodynamics, viewed in the article.

    a. Electromagnetic wave reflects only from those sections of environment (edges)where change of permittivity, conductivity or both parameters together exist.

    b. Signal passed through the edge has lower amplitude, but keeps its polarity andshape.

    c. Signal reflected from more dense or more conductive environment changes its polarity.

    d. Signal reflected from less dense or less conductive environment does notchange its polarity.e. Signal reflected from smooth edge “stretches” over time. f. Low conductivity leads to signal attenuation only, while high conductivity leads

    additionally to their dispersive distortion. In particular, low-frequency “tails”appear.

  • 8/9/2019 Terravision Technical Papers.pdf

    21/55

    Pacifico Minerals 22 Sept 2014 21

    Preprocessing of Terravision Radar signals.

    This chapter reviews features of Terravision Radar signals and methods of their preprocessing.

    1. Introduction.

    Ordinary radiolocation of air targets uses radio impulses in the capacity of directimpulses. These radio impulses have carrier frequency and are modulated in oneway or another. In geo-radiolocation such type of signals cannot be used due tostrong attenuation of radio-waves in underlying terrain. Spatial dimensions of radioimpulses in ordinary locators constitutes units – hundreds of meters, and this is themaximum sounding depth of Radars, resulting from attenuation. Here it is notnecessary to talk about time resolution of items, since signals reflected fromsubsurface items attenuate earlier than direct impulse ends.

    In order to increase time resolution it will be necessary to decrease radio-impulseenvelope. On the other hand, in order to increase sounding depth it is necessary todecrease carrier frequency. The limit of these tendencies is super-broadbandimpulse, which has from one to several current oscillations (voltage, electric ormagnetic field intensity – depending on in which part of the hardware orenvironment it is considered).

    In domestic literature such impulses are sometimes called radio-impulses without

    carrier or video-impulses. The latter definition comes from ordinary radiolocation,since they resemble video-impulses by shape (detected radio-impulses), used invisualization channels of these radio locators.

    In western literature such super-broadband impulses are called “simple pulse”. Wewill further call them simply direct impulses, meaning that a different type ofsignal in geo-radiolocation is not applicable.

    In this article we will review ways of evaluation of signal parameters, reflected

    from subsurface heterogeneities while sounding with video-impulses. We willname this processing as preprocessing, since it is performed over wave functions ofreflected signals obtained in one site of profile.

    2. Shape of direct impulse.

    For generating direct impulses in Terravision Radars, as a rule, a so called “methodof impact excitation” is used. In accordance with this method a condenser whichconnects directly to the transmitting aerial through the key element (usuallyavalanche transistor or discharger) is charged up to a specific voltage. Transmittingaerial (just like receiving aerial) in the general case is a resonance system, whichhas its own frequency and good quality, defined by the geometry of antenna

  • 8/9/2019 Terravision Technical Papers.pdf

    22/55

    Pacifico Minerals 22 Sept 2014 22

    (length, shape) and properties of underlying terrain. Like any resonance system, itcan be either oscillating or aperiodic, depending on energy losses of the system onradiation and heat [1].

    If in Terravision Radars one uses antennas without additional artificial energydissipation, as it takes place in the majority of devices (for instance, antennas of

    “butterfly” type), then a signal radiated by such antennas has an oscillating naturewith attenuation. The shape of such signal is shown on pic.1.

    Pic.1Oscillating impulse with attenuation. The function intersects zero-axis multiple times.

    The reason of oscillations is that in ordinary conditions (on soils with averagestatistical parameters) the effectiveness of antenna radiation and energy dissipationon heat do not switch resonance antenna system to aperiodic mode. Thereforeimpulse function intersects zero-axis multiple times, and theoretically – infinitelyoften with lower amplitude, since asymptotic of this function is

    )cos(~ t eF t (1)

    Here - attenuation coefficient, - own frequency of oscillating system.

    For ensuring aperiodic mode of radiation it is necessary to insert additionalartificial energy dissipation into the antenna, as it is done in resistive-fraughtdipoles, sometimes called Wu-King antennas after the names of their developers.Aperiodic direct impulse is shown in pic.2.

  • 8/9/2019 Terravision Technical Papers.pdf

    23/55

    Pacifico Minerals 22 Sept 2014 23

    Pic.2

    Aperiodic impulse. Function intersects zero-axis only once.

    In aperiodic signal the “tail” asymptotically tends to zero, not intersecting axes.Asymptotics of this function is

    t eF ~ (2)

    Such function can be considered practically ideal for direct impulse of TerravisionRadar, taking into consideration limitation of frequency line from below. Due tothis reason direct impulse cannot be uni-polar, since it lacks a zero frequencymultiplier (permanent multiplier), for the radiation of which an infinitely longantenna will be require.

    Radargrams, obtained with the help of aperiodic signal, have significantly betterquality of resolution then under oscillating signal, when reflection from each edgeis accompanied by periodic iterations (rings). However, using aperiodic signal isconnected with additional energy losses, which reduce radar’s capability.Therefore, Wu-King’s antennas can be used in Terravision Radars with morecapabilities, when it is possible to spare some energy of the device forimprovement of data quality.

    3. Shape of reflected signals.The aerial system of a Terravision Radar is located, as a rule, directly on the

    surface of earth. Separation of antennas from the surface to the distance of morethan several centimeters leads to remarkable worsening of data quality. Practicallyalways the rule: “the closer – the better” is true. Because of this, surface layers of

  • 8/9/2019 Terravision Technical Papers.pdf

    24/55

    Pacifico Minerals 22 Sept 2014 24

    soil are located in the near-field region of antenna and can be considered as a partof the antenna itself, which forms its major characteristics: input resistance,directional diagram, frequency range.

    When relocating radar along a line, sections of earth adjacent to the antenna canchange their permittivity and conductivity , and this may lead to a change in

    form of transmitted direct impulse. For example, low damped resistive dipole ondamp soil can radiate aperiodic impulse, and on dry sand – oscillating.

    When reflecting from the edges a direct impulse changes its polarity if wave passesfrom environment with lower permittivity and conductivity to the environmentwith higher values of these parameters, and does not change its polarity in theopposite case. The shape of reflected impulse (for example, its width) is defined bymaterial dispersion and width of intermediate region (edge) between layers.

    It follows from the above stated that a stable shape of direct and reflected impulsescannot be guaranteed, since they are substantially formed by the surveyedenvironment itself. Due to this reason, it is not possible to be based on somespecific shape of impulse (for example, its width or number of oscillations) when

    processing data, it is necessary to provide for possibility of variations of its parameters.

    Major information about environment is concluded in the amplitude of impulse, its polarity, time delay and width of oscillations. Geometric dimensions of layers andtheir parameters , influence amplitude, polarity and time delay, and width ofedge – to the width of oscillations. These parameters can be evaluated for a largeclass of functions, both oscillating and aperiodic. Reliability of evaluations,nevertheless, will depend on noise level and specific impulse function. Upon otherequal conditions aperiodic signals will deliver more reliable evaluations thanoscillating.

    A real radargram in the general case consists of superposition of multiple reflectedalternating signals, which can lay on each other. If direct impulse were “ideal” uni-

    polar, then detection of reflected impulses would add up to seeking a maximum ofobtained realization module. Under alternating function it is impossible todistinguish a signal, reflected by a new layer, from a “tail” of previous signal underthis algorithm. A kind of detecting method of reflected signals, invariant withrespect to different impulse functions, is required. Such method is provided below.

    4. Detection and evaluation of parameters of reflected signals.It is proposed to use Hilbert transform as a basis of algorithm of detecting super-

    broadband Terravision Radar signals and defining their features. Hilbert transformmodule for direct and reflected Terravision Radar impulses is of quite arbitrary

    shape – smooth analytical uni-polar function having only one maximum.Therefore, detection of signal on time-axis adds up to seeking maximums of

  • 8/9/2019 Terravision Technical Papers.pdf

    25/55

    Pacifico Minerals 22 Sept 2014 25

    transform module. Impulse polarity is defined by values of wave function at themoment of time corresponding to the maximum of module.

    Hilbert transform from a real function )(t x consists of calculation of someadditional function )(t y , whose all spectral components have the same module, butare turned over on phase at 90 , that is transform realizes a function of ideal phase

    changer. For example, for function )sin()( t At x additional function is)cos()( t At y .

    Hilbert transform is usually represented in the complex form )()()( t yit xt h . Inthe theory of analytical signals, impulse envelope )(t A is defined through moduleof Hilbert transform )(t h

    )()()()( 22 t yt xt ht A (3)

    In our case, additional function )(t y has a quite definite physical meaning – it is“recovered” magnetic multiplier of electromagnetic impulse. Although we even donot register magnetic multiplier of direct impulse, according to the laws ofelectrodynamics, it certainly exists in radio-wave, its spectral components arereally turned over on the phase at 90 with respect to electric multiplier.

    Thus, a square of Hilbert transform module can be considered as normalizeddensity function of full electromagnetic energy of direct and reflected impulses.

    Let us provide below a definition of Hilbert transform for digital function )(t x under number of counts N with a step t through spectral density )( f S with a stepover frequency )/(1 t N f

    1

    0

    .2/,.....,1,0),/2exp()()( N

    k

    N n N knit k xt f nS (4)

    2/

    1 )/2exp()(Im2)(

    N

    n N kn f nS f t k y (5)

    For definition of polarity of reflected impulses one of transform properties is used:maximum module )(t h corresponds to a maximum of module )(t x of real alternatefunction [2]. For a maximum of Hilbert transform module a function sign )(t x at acorresponding time moment is analyzed. Function sign corresponds to the polarityof impulse.

    Pic.3 and 4 show oscillating alternate impulses of identical shape, but different polarity and Hilbert transform module (impulse envelope) calculated with formulas(4-5), which has only one maximum, agreeing with the extremum of impulse.

  • 8/9/2019 Terravision Technical Papers.pdf

    26/55

    Pacifico Minerals 22 Sept 2014 26

    Let us formulate algorithm of preprocessing of Terravision Radar data on the basisof Hilbert transform.4.1 Preliminary digital filtering (not necessary).4.2 Calculation of Hilbert transform.4.3 Search of local maximums of Hilbert transform module.4.4 Determination of amplitude and sign of reflected impulses by wave function

    for time moments, corresponding to local maximums of transform.4.5 Determination of width of oscillations over intersection of zero from the leftand the right of detected extremum of reflected impulse.

    Pic.3Oscillating impulse of positive (by maximum) polarity and Hilbert transform module.

  • 8/9/2019 Terravision Technical Papers.pdf

    27/55

    Pacifico Minerals 22 Sept 2014 27

    Pic.4

    Oscillating impulse of negative polarity and Hilbert transform module.

    4. Conclusion.

    Let us formulate in brief main conclusions of the article on discussed issues.

    a. For geo-radiolocation because of a large wave absorption in the environmentonly video impulses (radio impulses without carrier) can be used in thecapacity of direct impulses.

    b. The best quality of data is provided by aperiodic direct impulse. For itsradiation and receiving antennas with additional artificial energy dissipationare required.

    c. Specific character of Terravision Radar devices and their operatingconditions do not ensure stability of characteristics of radiated and reflectedsignals.

    d. For preprocessing of data it is proposed to use Hilbert transform, whichtransforms different types of impulses into invariant form. Describedalgorithm of preprocessing is used in the software “Siluet”.

    In computer processing packages of Terravision Radar information, preprocessing,as a rule, is not used. The issue of classification of reflected impulses is resolved

    by profile (record of wave functions upon line movement). If layers are not plane- parallel, then appears an opportunity to observe interference figure and mark out“coherent lineup”, which is interpreted as a location of reflecting layers. But, as

  • 8/9/2019 Terravision Technical Papers.pdf

    28/55

    Pacifico Minerals 22 Sept 2014 28

    practice shows, such mark out badly algorithmizes and significantly depends onoperator’s experience.

  • 8/9/2019 Terravision Technical Papers.pdf

    29/55

    Pacifico Minerals 22 Sept 2014 29

    Inverse problem of geo ‐radiolocation.

    This chapter reviews two major methods of obtaining Terravision Radarinformation – profiling and sounding – and conditionality for solving inverse

    problem – recovering geologic structure of subsurface environment according to

    radargrams.

    1. Introduction.

    Inverse problem of geo-radiolocation – recovering structure of subsurfaceenvironment according to Terravision Radar data – commonly, like all inverse

    problems is incorrect (ill-conditioned) and ambiguous. Possibility of its solution isdetermined by a specific problem definition which includes a priori data aboutenvironment, as well as by a method of Terravision Radar data take-off.

    Under inverse problem we will further interpret recovering in flat-layeredgeometric dimensions case (deposition depth, capacity) of geologic layers and theirelectric features: permittivity and conductivity , and probably determination ofwidth of intermediate region (edge) between layers with the help of TerravisionRadar data.

    Flat-layerness of environment means that it consists of finite number of layers, inwhich electric parameters are constant. Layers have preferred horizontal

    stratification and are divided by an intermediate region – a boundary, width ofwhich, as a rule, is significantly smaller than the width of layer. In degeneratedcase flat-layer environment may consist of one wide boundary, in which verticalchange of electric features is significantly higher than horizontal [1]. Sand soil, inwhich dampness smoothly increases with the depth, can serve as an example.

    Two main methods of Terravision Radar surveys exist: “profiling” and “sounding”.When profiling radar moves along a line, at each measurement transmitting andreceiving antennas are located in one site of line, or at least the distance betweenthem is constant and significantly smaller than the length of line.

    During sounding one site of line, for which it will be made, is selected. Further anumber of registrations of reflected signals when moving antennas of transmitterand receiver to different sides at equal distances is carried out. As a result weobtain locus – a function of time delay of reflected signals from the distance

    between transmitting and receiving antennas [2]. Is there is a guarantee that in thevicinity of sounding site reflecting boundaries are flat-parallel, then it is possible tosimplify surveying procedure – leave the receiver on its place, and move onlytransmitter (or vice versa).

    A combined method is also possible, when sounding is performed in every site of profile, and also using these methods during areal mapping. Below we will review

  • 8/9/2019 Terravision Technical Papers.pdf

    30/55

    Pacifico Minerals 22 Sept 2014 30

    two major methods of obtaining Terravision Radar data – profiling and sounding interms of possibility of solving inverse problem.

    2. Profiling.Let us consider a simplified 1-D profiling layout for three boundaries, shown inpic.1 .

    Pic.1Simplified layout of impulse reflection and advancing for a separate site of line during profiling.

    Main simplification of the problem is the absence of multiple re-reflections between sharp edges. in fact, such situation is more realistic and is defined by astrong attenuation in layers. Multiple re-reflections are observed quite rarely, moreoften on profiles of pure fresh ponds under absence of wind, on which bottomrelief, while registering reflected signals, can repeat over and over again. Windcauses roughness of water surface and re-reflection from such surface becomesdiffusive.

    In the observed environment unknown parameters (independent variables), whichwe would like to define, are: capacities of layers 1h , 2h , permittivity 1 , 2 , 3 ,conductivity 1 , 2 .

    Слой 11 , 1

    Слой 22 , 2

    21

    2112

    R

    21

    212

    2

    T

    Слой 33 , 3

    32

    3223

    R

    21

    121

    2

    T 1h

    2h

  • 8/9/2019 Terravision Technical Papers.pdf

    31/55

    Pacifico Minerals 22 Sept 2014 31

    All these seven independent variables influence parameters of two reflectedimpulses. Is we work in high-frequency range, then shape of reflected impulsesaccurate within sign and constant coefficient repeats the shape of direct impulse.Therefore, from received reflected signals we can determine only four parameters:time delay 1t , 2t and amplitudes 1 A , 2 A .

    For these four parameters we can write four independent equations, connectingthem with environment parameters.

    c

    ht 111

    2 (1)

    c

    h

    c

    ht 22112

    22 (2)

    )2exp( 111201 h p R A A (3)

    )22exp( 221121231202 h ph pT RT A A (4)

    Here: 12 R - coefficient of reflection from the edge between the first and the secondlayers, 12T - coefficient of advancing through the edge between the first and thesecond layers, 23 R - coefficient of reflection from the edge between the second and

    the third layers, 21T - coefficient of advancing through the second and the firstlayers, c - speed of light in vacuum, 0 A - initial amplitude.

    In the region of high frequencies attenuation coefficient p is connected withconductivity in the following manner:

    20 c p

    Expression for coefficients of reflection and refraction are provided in pic.1.

    It is clear from the formulas provided above that it is impossible to solve an inverse problem while profiling: seven independent variables cannot be unambiguouslydetermined from four equations. The problem is solved only under presence ofadditional information, for example, drilling data, or under presence of multiplereflections, which allows increase the number of equations.

    3. Sounding.

    Sounding layout is shown on pic.2 .

  • 8/9/2019 Terravision Technical Papers.pdf

    32/55

    Pacifico Minerals 22 Sept 2014 32

    Pic.2Sounding layout. Initially transmitter and receiver are located on position 1, then on position 2.

    It is seen on the provided layout that number of independent equations duringsounding is doubled. Increasing spacing sites to more than two does not increasethe number of independent equations, however, during sounding usually more thantwo sites are used. It is explained with the following reasons: firstly, larger numberof sites allows better identify locus; secondly, accuracy of calculations increase.Here principally one can obtain accuracy of calculations higher than samplingincrement of corresponding parameter.

    Let us write below equations for time of impulse receipt and amplitude, reflectedfrom the edge between the first and the second layer.

    c

    lt 111

    2 (5)

    c

    lt 122

    2 (6)

    Layer 11 , 1

    Layer 22 , 2

    Layer 33 , 3

    1h

    Transmitter

    2 pos. 1pos.

    Receiver

    1pos. 2pos.

    2h

  • 8/9/2019 Terravision Technical Papers.pdf

    33/55

    Pacifico Minerals 22 Sept 2014 33

    )2exp()( 111201 l p R A A (7)

    )2exp()( 211202 l p R A A (8)

    Here l - geometrical length of impulse distance in the environment from the pointof radiation to the point of reflection, equal to the distance from the point ofreflection to the point of receipt

    4

    22 d hl (9)

    It is determined according to Pythagorean theorem by means of indefinitethickness of layer h and definite distance between transmitter and receiver d .

    Reflection coefficient )( R under oblique incidence depends on hade . It can becomputed by means of Snell law, or, if d h , accept )0()( R R .

    Equations (5-8) for the first impulse allow determine all parameters for the firstlayer: 1 , 1 , 1h , as well as permittivity of the second layer 2 by means ofreflection coefficient 12 R . It is evident that similar equations can be also written forall subsequent reflected impulses. Thus, sounding method allows unambiguouslysolve inverse problem of geo-radiolocation in flat-layer arrangement.

    When processing sounding data we most often use only equations (5-6) for timedelay, which are defined by locus. From these equations one can obtain a truedepth of layer 1h and permittivity 1 . Often this data is sufficient for solvingmajority of practical problems.

    As it was already mentioned – locus – is a dependence of signal time delay t fromthe distance d between receiver and transmitter. It is assigned by the formula

    422

    2d

    hct

    (10)

    Formula (10) assigns a locus for absolute synchronization, when receiver startscountdown immediately after radiation of impulse. In Terravision Radar “Loza”synchronization is carried out over air wave from direct impulse of transmitter,therefore, commencement of registration delays from radiation time for cd t / .

    cd d

    hct 422

    2

    (11)

  • 8/9/2019 Terravision Technical Papers.pdf

    34/55

  • 8/9/2019 Terravision Technical Papers.pdf

    35/55

    Pacifico Minerals 22 Sept 2014 35

    results, and then conduct processing of records, initial data is usually multiplied bysome often exponential function in order to strengthen later according to the delaysignals. Since strengthening function is determined “approximately” by theoperator and a little connected with unknown properties of the environment, then itis reasoned more than binary “leveling”, when an equal amplitude is attached to allsignals intersecting the edge. In both cases signal time delays remain the same, and

    this enables to use the same data processing algorithms for binary records that wereused during registration of wave shapes.

    Nevertheless, during binary registration, more often in low resistance environment,sometimes situations occur, when from upper edges appears low-frequency “tail”,which takes out low signals from threshold region and they do not intersect it. Anexperienced operator can set them by means of changing main registrationthreshold or connecting two additional thresholds of different level. Otherwise,threshold registration leads to partial loss of signal information.

    Pic.3Trajectory of propagation of impulses during profiling of single objects. Because of final width

    of directional diagram the object is observed at some area from the depositing point.

    x

    h l

  • 8/9/2019 Terravision Technical Papers.pdf

    36/55

    Pacifico Minerals 22 Sept 2014 36

    4. Determination of deposit depth of localized objects.During profiling there exists an opportunity to determine the deposit depth ofseparate localized objects and permittivity of environment covering them due tosignificantly wide directional diagram of transmitting and receiving antenna. Whilemoving over line a separate localized object is observed not only in the site directlyabove it, but also at some remoteness to both directions. In this connection the

    value of this remoteness is determined by angular width of antennas directionaldiagram. Pic.3 shows a layout of propagation of impulses during profiling of asingle object.

    Time delay t is determined from geometry shown in pic.3.

    2222 xhcc

    lt (14)

    Formula (14) has the same look as the locus formula, therefore, with twomeasurements of time delay at two sites, depth of object depositing and

    permittivity of covering environment can be calculated from it. Let 0t be a signaltime delay above object, 1t - time delay at remoteness 1 x , then depth and

    permittivity will be determined as

    2021

    21

    20

    t t

    xt h (15)

    2

    220

    4hct

    (16)

    Pipes and cables often act as single objects. But it should be remembered that theseobjects can be considered single and use formulas (15-16) only upon their

    perpendicular intersection by profile. Intersection at any other angle violatesgeometry of the problem ( pic.3 ) and brings up incorrect results.

    In this method of determining depth and permittivity we may not necessarily talkabout using signal amplitude, since evaluation of reflection coefficient from asingle object is itself of complex task, even if its exact structure is known.

    5. Conclusion.The most informative method of Terravision Radar survey is sounding. For flat-layer environment this method allows determine all environment parameters:capacities and depth of layer deposits, their permittivity and conductivity.

    Profiling method allow unambiguously determine none of environment parameterswithout additional assumptions or without attracting additional data.

  • 8/9/2019 Terravision Technical Papers.pdf

    37/55

    Pacifico Minerals 22 Sept 2014 37

    More optimal method of Terravision Radar survey in terms of time expenditure isinitial profiling of line and determination of sounding sites. These sites are selectedon flat-layer regions of line. If with regard to a selected region on the remaining

    part of line new layers do not appear, then one site of sounding is sufficient. Dataobtained during sounding will further serve as additional data for profiling,

    allowing solve inverse problem on the whole line of profiling.

    Making a conclusion about ambiguity of solving inverse problem during profiling,we reviewed the process of impulse propagation in high-frequency asymptotic,when dispersive distortions of their shape are absent. In the region of lowfrequencies, except for impulse attenuation, dispersive distortions will also beobserved. Perhaps, laws of impulse shape changes will be managed to be used asadditional information for solving inverse problem in the region of lowfrequencies.

    Signals of Terravision Radar in conductive environments.

    Radio waves belong to electromagnetic oscillations, but not all electromagneticoscillations belong to radio waves. A distinctive feature of radio waves is the

    presence of bias current. This current is introduced into electrodynamics byMaxwell. Its density in vacuum is determined by the formula:

    t E

    J S 0 (1)

    Here S J - density of bias current, 0 - vacuum permittivity, E - electric fieldvector, t - time.

    An example of bias current may be a passing of alternating current throughcondenser, which has no charge carriers between its facings.

    Current density conductivity P J in the environment, where charge carriers exist, isdetermined by the formula:

    E J P , (2)

    where - environment conductivity.

    In the law of total current, which is a constituent part of a total system ofMaxwell’s equations, both currents participate on equal terms:

    E t

    E H rot 0 (3)

  • 8/9/2019 Terravision Technical Papers.pdf

    38/55

    Pacifico Minerals 22 Sept 2014 38

    However, in the wave equation, obtained from the system of electrodynamicequations, they participate differently and determine various propagationmechanisms of electromagnetic energy.

    A wave equation for electric field intensity in environments with conductivity has

    the following look:

    0022

    22

    t E

    t E

    c E

    (4)

    Here - relative environment permittivity, 0 - magnetic conductivity of vacuum,c - speed of light.

    Differential Equation in Quotient Derivatives (4) (in students terminology –

    “DEQD”) has the signs of two equations: Dalamber’s equation, in which the firsttime derivative is lacking, and diffusing equation, in which the second timederivative is lacking. These equations are considered classical and have beenstudied long ago.

    Dalamber’s equation describes electromagnetic wave processes, in which biascurrents prevails over condition currents. For a Terravision Radar in the suburbs ofMoscow city – these are usually frequencies over 50 MHz. But with the decreaseof frequency and/or increase of conductivity, processes connected with

    environment conductivity that are described by the diffusion equation (or heatconductivity equation, which has the same look), start being revealed. In geo- physics the diffusion process of electromagnetic field into conductive environmentis used in the “field formation method”. Physical processes occurring in theformation method have a different physical nature and are substantially (for severaltimes!) slower than radio wave processes.

    The effect of conductivity becomes apparent in the ohm radio wave absorption andin the appearance of a diffusion component in the form of “tails” (or “taffies”)

    directly after the radio signal, the origination process of which is shown on pic. 1-4.

  • 8/9/2019 Terravision Technical Papers.pdf

    39/55

    Pacifico Minerals 22 Sept 2014 39

    Pic. 1 . Signal in the environment without conductivity.

    Pic. 2 . Environment conductivity - 0.02 Sim/m.

  • 8/9/2019 Terravision Technical Papers.pdf

    40/55

    Pacifico Minerals 22 Sept 2014 40

    Pic. 3 . Environment conductivity - 0.04 Sim/m.

    Pic. 4 . Environment conductivity - 0.06 Sim/m.

  • 8/9/2019 Terravision Technical Papers.pdf

    41/55

    Pacifico Minerals 22 Sept 2014 41

    Diffusion “tail” of the signal has a maximum sign of preceding radio signal. Asymmetric radio signal, when its maximum and minimum values are equal by size,has no “tail”.

    The size of “tail” is so lengthy in terms of time (for example milliseconds), that

    Terravision Radar is simply unable to record it due to a limited duration ofregistration, even if its input circuits were passed through to the receiver input. Butour gears have a high frequency filter on the input, which cuts frequencies atdurations of about a microsecond. Therefore, we record only the beginning of“tail”, which has real amplitude. But signal fall function for all “tails” is similarand is defined by frequency features of the gear only, including its antennas.

    Radio and diffusion signals are described by a linear differential equation (4), forwhich a principle of superposition is performed: all of them are simply summed up

    along the time scale. It is demonstrated on pic. 5-8.

    Pic.51 1 2 29 0 4 0

  • 8/9/2019 Terravision Technical Papers.pdf

    42/55

    Pacifico Minerals 22 Sept 2014 42

    Pic. 6 .

    Pic.7 1 1 2 29 0 16 0

    1 1 2 29 0.01 4 0.01

  • 8/9/2019 Terravision Technical Papers.pdf

    43/55

    Pacifico Minerals 22 Sept 2014 43

    Pic. 8 .

    As it can be seen from the pictures, conductivity of layers does not affect time position and shape of radio signals, which occur as a result of reflection of twolayers at the edge. In contrast to radio signal, a diffusion signal is formed not bythe edge, but by the layer mass itself.

    In geo-radiolocation the appearance of “tails” in itself is not considered a kind of principal obstacle for the solution of a problem, although it suggests thatconductivity in the environment has been increased, which is an evidence of signalattenuation, leading to the limitation of depth.

    Radio signal can be discovered well enough against a background of “tails”. As anexample of this, we can provide the results of seabed sounding in our works on

    mapping of Kimmeriyskiy swell on the bed of Taman gulf. Seawater and seabedhave a high level of conductivity, which bring up very powerful “tails” right at the

    beginning of time-base. But selection of binary registration threshold practically onthe level of its size, allowed obtain results with a high-quality spatial resolution.

    In the case of our gear the “tail” has a partial negative impact, since it “lifts up” thesignal to the area of high values of field intensity, where sensitiveness is lower thanin the area of low fields: due to the necessity of registering the signal within a hugedynamic range we apply signal “compression” at its high values. This small

    deficiency can be compensated by a digitizer with a low sampling increment.

    1 1 2 29 0.01 16 0.01

  • 8/9/2019 Terravision Technical Papers.pdf

    44/55

    Pacifico Minerals 22 Sept 2014 44

    However the existence of “tails” itself allows expand the capabilities of the device:a principal opportunity to determine conductivity of subsurface layers appears.

    Pic. 9 shows a layout of receiving and transmitting antennas, which we used duringmapping of Kimmeriyskiy swell on the bed of Taman gulf.

    Pic.9.

    Terravision Radar Measurement modes

    Terravision Radars of “LOZA” series have three main measurement modes: Terravision Radar “LOZA-M” (not produced at present) makes measurements in binary

    registration mode; Terravision Radar “LOZA-B, В 1, Н , Н 1” enable making measurements in the “binary

    mode” and in the mode of wave shape registration in logarithmic scale; Terravision Radar LOZA- В 2” operate in the binary mode and record the wave shape in

    logarithmic and linear scale.

    3.1. Binary registration mode.Enter measurement mode – PROFILE (BINARY MODE)- check measurement menu settings,

    14h 15m 16s

    REGISTER

    PROFILE ROUTE

    Seabed

    Current linesin seawater

    Containerfrom

    dielectricMetal spheres-

    electrodes

  • 8/9/2019 Terravision Technical Papers.pdf

    45/55

    Pacifico Minerals 22 Sept 2014 45

    (title)

    SHOT 001

    MANUAL START

    --

    0 threshold - 02 threshold - +163 threshold - -16FORMAT 128 256

    AVERAGING ON ПАМ – 100%BAT – 12,6V

    t - 25 C

    Menu items in measurement mode highlighted in the blue color are informative and are notsubject to changing.

    Adjustable parameters: Start.

    Format.

    - “MANUAL START” (measurement comes from pressing a button),Other available modes

    – “AUTOMATIC START” with 1-9 sec. period , (period adjustable) start ofautomatic measurement comes from pressing a button, switch modes – long buttonclick. In the option “ В 1” in the binary mode there is an opportunity to conductregistration in automatic mode of 10, 20 and 50 measurements in second.- “START FROM SENSOR” (start of end switch on measuring wheel (odometer),sensor is connected to remote control connector from the right lower part of theregistrar). Triggering from sensor with adjustable recurrence – one triggering for1,2,4… rotations of wheel.

    Duration of time-base – receiver’s time of operation on reception ( it is neither power nordeepness ).Options

    256 nanoseconds (most occasionally used time-base). For medium soils such time- base corresponds to the depth of 14-15 meters.

    512 nanoseconds. For medium soils such time-base corresponds to the depth of 28-30meters. This time-base should be used only in the case if soil properties allow expectreflected signal from this depth. If informative part of profile occupies upper one thirdof the screen, and lower is a noise field, it should be necessary to switch to time-base256 ns. Time-base 512 ns is most often used while operating on the surface of water.

    It is not recommended to reconfigure “FORMAT”. If you, while working on one of the

    time-bases, decide to switch to another one – you should assign a new file and set new“FORMAT” in it.

  • 8/9/2019 Terravision Technical Papers.pdf

    46/55

    Pacifico Minerals 22 Sept 2014 46

    Thresholds.

    Averaging.

    Display of measurement results.

    “LOZA-M” and “LOZA-B, H” Terravision Radar screen is divided into two panels sized 128 pixels (horizontal) and 256 pixels (vertical).

    The left panel displays an image under “0” threshold. Displayed image represents a result ofthreshold processing of wave shape of reflected sounding signal.

    Each measurement is represented by a column of black and white pixels. Each new measurementappears in the left position, and the image moves to the right.

    The right end of the image is the beginning of the profile, the left end – the end.

    The display can simultaneously show 128 measurements, upon 129 th pressing of the “start” button a special indicating signal is heard. This signal represents moving of full shot data to permanent memory.

    Each new measurement will move the image one column to the right, and the panel will displaythe last 128 measurements. After finishing measurements an incomplete part of shot must berecorded by pressing a button “REC” and only after doing so the device can be switched off.Otherwise, only complete shots with 128 measurements will remain in the memory, and the lastincomplete shot will be lost.

    The second (right) panel can display measurement data on 2 nd or 3 rd thresholds. For choosing athreshold, measurement results of which will be displayed on the second panel, it will benecessary to set the cursor to appropriate threshold (2 or 3) and press “ENTER” . In the upper

    part of the measurement menu an indication of threshold will appear.

    For the division of profiles saved into one single file, you can use a “MARK” function. In the binary mode of registration the “MARK” has a visual character and appears on the screen as acolumn of white and black pixels (10 white, 10 black etc.) the “MARK” occupies measurementcolumns and this should be taken into account during profile processing in order to not distort

    profile’s length scale.

    Saving measurement data to the memory.A complete shot (128 measurements) is moved to the permanent memory automatically, anincomplete shot should be saved before switching off. Saved data can be viewed in the “VIEW

    In case of switching on averaging mode an average data of 16 measurements will be sent tothe memory and the device display. In case of operating in external disturbance conditions,using averaging may increase sounding depth for 20-30%. In the binary mode of registrationaveraging procedure will slightly increase unit measurement time. Recommended position –“ON”.

    Recommended values – 0 – threshold “ 0”,2 threshold “-16 ”,3 threshold “+16 ”.

    It is not recommended to reconfigure “THRESHOLD” within a single file. Ifwhile working on one of the thresholds, you decide to switch to others – you shouldassign a new file and set new “THRESHOLDS” in it.

  • 8/9/2019 Terravision Technical Papers.pdf

    47/55

    Pacifico Minerals 22 Sept 2014 47

    MEMORY” mode. Resaving data to the PC (desktop) is carried out by means of RS232 cableconnected to the registrar socket. Resaving data to laptop is carried out by means of RS232 cableand additional RS232-USB adapter (included in the device set). Resaving data from TerravisionRadar to computer can be done using software “TRANSFER 1002” or “1005” or software“KROT” . Data concerning three thresholds will be saved to files *.geo, *.geo2, *.geo3 correspondingly. When opening file with the program “KROT” , all three files must be in thesame directory.

    Data structure of binary mode of registration.Binary mode of registration represents a threshold processing process of reflected signal’s waveshape. Received signal goes through a special device called “comparator”, which determines asignal phase “+” or ”-“ to each quantization step of time with respect to preset threshold level.The most sensible, in most cases, is the zero threshold, since all reflected signals, both powerfulones – reflected from surface layers and weak ones – reflected from deep layers, repeat the shapeof sounding signal and have positive and negative parts.

    Pic. 3.1.1. Binary mode of registration signal.

    Pic. 3.1.1. shows an example illustrating a principle of binary registration. Left part of picture“1” displays a real wave shape of one measurement made approximately at 50 meter point of

    profile (the profile is done in the valley of Dnepr river in Smolensk region). A zero threshold(blue vertical line) divides a wave shape to the right “+” (conditionally, black) and to the left “-” (conditionally, white) parts. When such wave shape goes through geo-radar comparator, at theoutput (on the display and in the device memory) we obtain a column of black and white pixels,which represent a section of wave shape on zero line “2” . At each new measurement there is avertical column of black and white pixels, displayed on the screen, corresponding to themeasured wave shape. Each new measurement is displayed in the left part of the screen, and

    previous measurements shift to the right along the screen. As the profile moves, an imageappears on the screen, on which black and white sections of unit measurements line up intostructures reflecting geological edges in soil “3” .

    Binary method of registration leads to loss of information about amplitude of reflected impulseand limits opportunities of further processing. Positive features of binary registration:

    1 2 3

  • 8/9/2019 Terravision Technical Papers.pdf

    48/55

    Pacifico Minerals 22 Sept 2014 48

    Results of measurements occupy very little memory (as compared to full-waveregistration),

    Measurements in binary mode of registration within a smaller period of time (ascompared to full-wave registration)

    A picture of Terravision Radar reflections (radargrams) in binary mode is formed withina short span of time and does not require significant microprocessor resources, whichallows use this mode for displaying the picture on registrar screen in all operation modes

    of geo-radar.

    A binary mode used in geo-radars of model “M, B, H, B2” is performed in a complicatedoptions. Information about received signal is registeredsynchronously for three thresholds. This data is saved inthree files with extensions *.geo, *.geo2 and *.geo3.

    Registration thresholds can be configured atdifferent levels, for example: main “0” , second +15 Db ,and third -15 Db .

    After moving the data to the computer, you canunite all three files in the processing software and obtain a

    picture with three gradations of signal amplitude.Registration under three thresholds insignificantlyincreases the volume of used memory and registrationtime, but significantly extends self-descriptiveness ofresult.

    Pic. 3.1.2. shows an example of binary registrationdata “1” – threshold 0 Db, “2” – second threshold +15Db, “3” – third threshold -15 Db. Pic. “ 4” shows a geo-radar profile assembled from files under three thresholds.Signal amplitude is represented by color box (4 colors).

    Pic. 3.1.2. Binary mode of signal registration under three thresholds.

    Examples of data under various configurations of thresholds in binary registration illustrateopportunities of changing screen self-descriptiveness for all types of signal registration in“LOZA” Terravision Radar. To enable evaluating Terravision Radar profile directly in the

    process of measurement or in the mode of memory view the screen displays a picturerepresenting a result of threshold processing. By changing configurations of main threshold inthe MAIN MENU you can refine the picture displayed on the screen. Modification of thresholdis performed by means of keyboard with a step 1 Db.

    3

    2

    1

    4

  • 8/9/2019 Terravision Technical Papers.pdf

    49/55

    Pacifico Minerals 22 Sept 2014 49

    3.2. Wave mode of registration.

    Wave mode of registration allows obtain more thorough information about reflected soundingsignal (delay time, amplitude and signal phase). For switching on wave mode you should select amode “WAVE SHAPE LOGARITHMIC” in the MAIN MENU.

    Enter measurement mode – “WAVE SHAPE LOGARITHMIC” .

    - check settings of measurement menu,

    14h 15m 16s

    REGISTERWAVE SHAPElogarithmic

    ROUTE(title)

    SHOT 001

    MANUAL START

    Mode 2 screens

    FORMAT 128 512.1

    AVERAGING OFF ПАМ – 100%

    BAT – 12,6Vt - 25 C

    Menu items in the measurement mode highlighted with blue are informative and are not subjectto changing. Adjustable parameters:

    Start.

    - “MANUAL START” (measurement comes from pressing a button),Other available modes

    – “AUTOMATIC START” with 1-9 sec. period , (period adjustable) start ofautomatic measurement comes from pressing a button, switch modes – long buttonclick - “START FROM SENSOR” (start of end switch on measuring wheel (odometer),sensor is connected to remote control connector from the right lower part of theregistrar). Triggering from sensor with adjustable recurrence – one triggering for1,2,4… rotations of wheel.

  • 8/9/2019 Terravision Technical Papers.pdf

    50/55

    Pacifico Minerals 22 Sept 2014 50

    2 screens/1 screen mode.

    Format.

    Averaging.

    Displaying of measurement results on the screen.The display of “LOZA-B and H” Terravision Radar in the mode of registration of wave shapes isdivided into two panels sized 128 pixels (horizontal) and 256 pixels (vertical). The left panelshows an image under main “0” threshold. Changing thresholds does not influence wave shapedata saved in the memory. An image displayed on the screen represents a result of threshold

    processing of wave shape of reflected sounding signal. Each measurement represents a columnof black and white pixels. Each new measurement appears in the left position, and the image

    slightly shifts right. The right part of the image is a beginning of the profile, the left part – theend. Each new measurement shifts the image for one column to the right, the last 128measurements will remain on the panel. The image is calculated for visualization of profiling

    Duration of time-base – receiver’s time of operation on reception ( it is neither power nordeepness ).

    Options - 256 ns,It is a time-base most often used in the medium line of Russia. For medium soils such time-

    base corresponds to the depth 14-15 meters.

    - 512.1 - 512 nsFor medium soils such time-base corresponds to the depth of 28-30 meters. This time-baseshould be used only in the case if soil properties allow expect reflected signal from this depth.If informative part of profile occupies upper one third of the screen, and lower is a noise field,it should be necessary to switch to time-base 256 ns. Time-base 512 ns is most often usedwhile operating on the surface of water

    - 512.2 – 1024 ns,- 512.4 – 2048 ns.

    Time-bases 1024 ns and 2048 ns in “LOZA-B” geo-radars are used while operating on thesurface of freshwater reservoirs, in “LOZA-H” – as main operation modes.It is not recommended to reconfigure “FORMAT” in one file. If you, while working on

    one of the time-bases, decide to switch to another one – you should assign a new file andset new “FORMAT” in it.

    In case of switching on averaging mode an average data of 16 measurements will be sent tothe memory and the device display. In case of operating in wave shapes, one measurement isdone in 0.6 seconds. When working with averaging duration of one measurement will reach2.4 seconds. Using averaging significantly slows down the process of movement along

    profile. When working outside the town and industrial objects a recommended position is“OFF” .

    In the mode “2 screens” screen of geo-radar “LOZA-B, H” is divided into two panels sized128 pixels (horizontal) and 256 pixels (vertical). Left panel displays a binary image undermain “0” threshold. The right panel displays a wave shape of the last measurement.In the mode “1 screen” the whole screen displays a binary image 256 and 256 pixels. Waveshape of each measurement is not displayed on the screen, but is saved directly in the

    memory. “1 screen” mode is recommended for using during geological surveys. On thescreen at operator’s eyes there is a performance of twice larger length, which makes analysisof geological structures easier.

  • 8/9/2019 Terravision Technical Papers.pdf

    51/55

    Pacifico Minerals 22 Sept 2014 51

    process and is kept in operating memory. The image is not saved after switching off the device.Only wave shapes of signal are saved in the memory. The second (right) panel displays a waveshape of the last measurement. In the mode “1 screen” the image is displayed on the wholescreen. The function “MARK” in the mode of wave shape registration is constructed in the waythat it enters measurement file and does not appear on registrar screen. Only one mark can bemade in one measurement. There are three mark types in the processing software KROT:

    Single mark – it is marked with black in the profile in the program. It is used for marking

    special profile points (turns, sewer hatches, columns etc.), Double mark – it is marked with blue in the profile in the program. It is performed in thefollowing manner during profiling: measurement – marking – measurement –marking . Double marking is used in the program KROT for automatic division of

    profiles saved in a single file, Triple mark - it is marked with red in the profile in the program. It is performed in the

    following manner during profiling: measurement – marking – measurement –marking – measurement - marking . It is used as a reserve mark for marking out

    particularly important items.

    Saving measurement data to the memory.Each measurement in wave shapes is saved to the memory automatically. The last measurementmust be saved by pressing a button “REC” . Saved data can be viewed in the mode “MEMORYVIEW” . It is convenient to use the following functions while viewing:

    Resaving data to the PC (desktop) is carried out by means of RS232 cable connected to theregistrar socket. Resaving data to laptop is carried out by means of RS232 cable and additionalRS232-USB adaptor (included in the device set). Resaving data from Terravision Radar tocomputer can be done using software “TRANSFER 1002” or “1005” or using downloadingsoftware “KROT” .

    Data structure in the registration mode of wave shapes. In the wave mode of registration, digital values of each separate measurement are saved in thedevice memory. These values contain information about signal amplitude and phase. Resolutionof digitization and time duration are defined by device settings. Binary picture displayed on thedevice screen, is calculated for visualization of sounding process and is not saved in the device

    memory. Screen picture of LOZA Terravision Radar has double meaning:

    - Zooming of screen part twice along vertical line (along time delay axis).To do this, it is necessary to preset a cursor at 5-6 pixels (lower edge of

    bore) and to press zoom button.

    - Cursor . After pressing this button a vertical and horizontal line of cursorswill appear on the screen. Moving of cursors is carried out by means of

    buttons “right-left” and “up-down”. A number appearing at the verticalcursor is a coordinate from the last measurement in pixels. A numberappearing at the horizontal cursor is a coordinate along vertical-axis in

    pixels (along time delay). For determination of approximate depth oneshould subtract a value “duration of bore” from a value “verticalcoordinate” (4-6 for 1m antennas and 8-10 for 1.5m antennas). Bymultiplying the obtained time delay to the average signal velocity in the soilin this area, we will obtain an approximate depth of item.

  • 8/9/2019 Terravision Technical Papers.pdf

    52/55

    Pacifico Minerals 22 Sept 2014 52

    According to binary picture, one can solve engineer-geological problems regardingdetermination of soil structure and finding objects in the process of observation before moving tocomputer. According to the information displayed on the screen (binary picture and wave shape of

    the last measurement) one can evaluate device operation conditions (noise situation, presenceand amplitude of “air” reflections etc.). Visual control of measurement process will help notice

    profile sections, on which the device is found in the area of strong noises capable of breaking

    synchronization of Terravision Radar’s operation. Pic. 3.2.1 shows an example of image onTerravision Radar screen (1) case of synchronization failure (left part of image). On these partsof radargram in the upper part of time-base (first 8-10 ns) “BORE” properties have noticeablychanged (period and phase of introduction). It is the first sign that the receiver is synchronized bya powerful impulse noise. The second sign of “irregular” operation of the device is a sharpchange of binary image structure – alternating and continuous black and white stripes. Waveimage in this part of measurements (2) justifies a noise origin of sounding signal. Wave shape(3) reflects a normal structure or reflected signal. In such situation it is necessary to increasesynchronization level for 5-10 units, until a complete stopping of synchronization failure.Default section of profile must be remade.

    Pic. 3.2.1. Geo-radar profile with sections of synchronization failure.

    The results of sounding in wave shapes when opening in the program KROT are represented bya colorful colored image. Palette of colors in this case determines signal amplitude: maximumamplitude of “+” phase (conditionally) is shown with red color, minimum amplitude of “-“ phaseis shown with black-violet tints of palette. Compliance of definite amplitude and phase of signalwith the color is conditional, and does not witness any properties of geological structure. Color

    presentation allows obtain a clear picture of amplitude-phase properties of Terravision Radar profile. Pic. 3.2.2 . shows an example of representing signal amplitude by a 16 color palette

    without any processing.

    1 2 3

    Normal view of bore.

    Synchronization by noise.

  • 8/9/2019 Terravision Technical Papers.pdf

    53/55

    Pacifico Minerals 22 Sept 2014 53

    Pic. 3.2.2. Geo-radar profile (Neva river). Amplitude is represented by 16 palette colors.

    Analysis, processing of Terravision Radar profile and preparation of reporting is carried out inthe program KROT or REFLEX.

    3.3. Distinctive features of Terravision Radar “LOZA-1B (Linear scale)” settings.

    A model of Terravision Radar “1 В – linear scale” is a further development of model “B” series.The device has options of all measurement modes of “В ” and “1 В ” models:

    1. Binary three-threshold registration.2. Registration of signal wave shape in logarithmic scale.

    The following new options are added to “1B – linear scale” model devices:

    1. Registration of signal wave shape in linear scale.2. Possibility of filtering (smoothing) in the mode of sounding data view on LCD screen

    of Terravision Radar.3. Possibility of viewing sounding results in the mode of picking out signal minimums

    and maximums on LCD screen of Terravision Radar.

  • 8/9/2019 Terravision Technical Papers.pdf

    54/55

    Pacifico Minerals 22 Sept 2014 54

    Pic. 3.3.1. Terravision Radar “LOZA-1B – linear scale” panel in the memory view mode.

    In the mode of registration of signal wave shape in the linear scale a signal level “is digitized” by128 levels uniformly (linearly) allocated along the whole dynamic amplitude range. As a resultof measurement of signal wave shape in the linear scale in the mode “with open input” (withoutweakening) there will be obtained some data, which will be located “in limitation” over the most

    part of time-base. The results of such measurements in terms of self-descriptiveness iscomparable with binary threshold registration mode. Advantages of linear registration scale will

    be completely realized only in the case applying additional signal depression on receiver input.Signal depression on input will decrease signal range and improve signal digitization detail. By

    adjusting signal depression in the main menu of the device, it becomes possible to move locationof detailed analysis of signal amplitude. From the experience of operating “LOZA-1B – linearscale” the advantages of linear registration scale start to realize after applying weakening of 40Db and over.

    Filtering mode is performed in two options of vertical and horizontal one-dimensional filtering,which can be used both jointly and separately.

    one-dimensional filtering over time delay scale (vertical filtering) is performed by anoption “FILTER 1” . Filtering is performed according to standard procedure ofnonrecursive filtering by a rectangular filter (moving average). Filter length N=3,5,7,if N=0, filtering is not performed. As a result of filtering a profile is cut off for (N-1)/2quantization steps in the beginning and at the end.

    )1()2/)1((1

    i I A N I F N

    i

    For changing filter parameters, move the cursor to “FILTER 1” position and setrequired filter length by successive pressing of a button with “double arrows”. Theresults of processing (filtering) are kept in operating memory and are displayed on theright half of LCD screen. The results of filtering are not saved to “FLASH memory”.

    one-dimensional filtering over J points of each time packet I (horizontal filtering) is performed by the option “FILTER 2” . Filtering is performed according to a standard procedure of nonrecursive filtering by a rectangular filter (moving average). Filter

  • 8/9/2019 Terravision Technical Papers.pdf

    55/55

    length M=3,5,7, if M=0, filtering is not performed. As a result of filtering a profile iscut off for (M-1)/2 quantization steps in the beginning and at the end.

    M

    j

    j J I F M J I FF 1

    )1,()2/)1(,(

    For changing filter parameters, move the cursor to “FILTER 2” position and set required filterlength by successive pressing of a button with “double arrows”. The results of processing(filtering) are kept in operating memory and are displayed on the right half of LCD screen. Theresults of filtering are not saved to “FLASH memory”.

    In the mode of picking out signal maximums and minimums a search of all local“MAXIMUMS” and “MINIMUMS” over vertical line, that is over variable I, or over timedelay is performed. Found signal MAXIMUMS are represented by black pixels on a grey

    background of binary (according to zero threshold) radargram. MINIMUMS of signal arerepresented by white pixels on a grey background of binary (according to zero threshold)radargram. Representation of geo-radar profile in signal MAXIMUMS and MINIMUMS can beexposed to filtering procedure using options “FILTER 1” and “FILTER 2” .

    The results of processing (picking out MAXIMUMS and MINIMUMS) are kept in the operatingmemory and are displayed in the right half of LCD screen and act for analysis of geo-radar

    profile in the field conditions. The results of filtering are not saved in “FLASH memory”.