Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Vineyard pruning weight assessment by machine visionÉvaluation du poids de bois de taille des vignes par un système de vision par ordinateur
Javier TardaguilaProfessor of Precision Viticulture
University of La Rioja. Spain
televitis.unirioja.es
From visual assessment to data-driven viticulture
New and emerging non-invasive technologies
Thermal imaging
Spectroscopy
Machine vision
Hyperspectral imaging
New technologies in precision viticulture
Thermal manual camera
Aerial platform
Terrestrial platform
Vineyard robot
7
Televitis mobile lab
Computer
Sensor 2
Sensor 1
GPS
@ 5 km/h
What viticultural parameters can
be assessed using non-invasive
technologies in the vineyards?
• Canopy status
• Water status
• Grape composition
• Cluster compactness
• Yield components
• Trunk diseases
• Pruning weight
Key viticultural parameters
Canopy status
y = 0.7427x + 8.1439R² = 0.91
0
10
20
30
40
50
60
0 10 20 30 40 50 60
% p
oro
sit
yfr
om
imag
ean
aly
sis
on
-th
e-g
o
% porosity with Point Quadrat (reference value)
On-the-go assessment of canopy porosity by machine vision
Diago et al. 2019 AJGWR
Vineyard water status
Vineyard water assessment using thermal imaging on-the-go
Gutierrez et al. 2018 PlosOne
14
Gutierrez et al. 2018 PlosOne
Machine learning and thermal imaging for vineyard water status monitoring
Grape composition
Hyperspectral imaging (HSI) working under field conditions
Assessment grape composition under field conditions by HSI
Gutierrez et al. 2018 AJGWR
Cluster compactness
Cluster compactness assessment in commercial vineyards
Palacios et al. 2019 Sensors
Yield components
Yield, a key factor for grape and wine sector
Diago et al. 2015 JSFA
Berry number per cluster by image analysis under
lab conditions
y = 1.4112x + 15.114R² = 0.81**
0
50
100
150
200
250
300
0 20 40 60 80 100 120 140 160 180
Ac
tua
l n
um
be
ro
f b
err
ies
pe
r c
lus
ter
Number of berries visible on images acquired at fruit set
Estimation of berry number per cluster by machine vision
Aquino et al. 2017
Visión por computador o Visión artificial Yield assessment from on-the-go image acquisition
Wood pruning weight
Traditional evaluation of wood pruning weight in the vineyard
• Manual pruning + Wood weighing
• High labour demanding
• Time consuming
• No comercial assessment
Machine vision in viticulture
Computer vision is a new technology for assessing key viticultural parameters
Machine vision in viticulture
Manually acquired image with background and segmented image
On-the-go assessment of vine pruning weight using machine vision
Original image taken on-the-go by Televitis mobile lab
Segmented image for pruning weight assessment
Multi-step automated image processing algorithm
Vineyard in Rioja cv. Tempranillo
y = 431775x + 24408
R² = 0.82**
0
100000
200000
300000
400000
500000
600000
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Pix
el n
um
be
ro
f w
oo
d
Pruning weight (Kg/vine)
Millan et al. 2019 OENO One
On-the-go assessment of pruning weight:
cv Tempranillo. Rioja
Mapping of pruning weight in commercial vineyards
Millan et al. 2019 OENO One
Conclusions
❑Machine vision can be applied for assessing wood pruning weight
❑ Vine vigour can be determined and map using a mobile sensing platform and image analysis
Merci beaucoup pour votre attention !