Image processing methods for quantitatively detecting soybean rust from multispectral imagesSoybean rust is one of the most destructive foliar diseases of soybean, and can cause significant yield loss. Timely application of fungicide in the early stage of rust infection, which is critically important for effective control of the disease, is heavily dependent upon the capability to quantitatively detect the infection. This paper reports research outcomes from developing image processing methods for quantitatively detecting rust severity from multi-spectral images. A fast manual threshold-setting method was originally developed based on HSI (Hue Saturation Intensity) colour model for segmenting infected areas from plant leaves. Two disease diagnostic parameters, ratio of infected area (RIA) and rust colour index (RCI), were extracted and used as symptom indicators for quantifying rust severity. To achieve automatic rust detection, an alternative method of analysing the centroid of leaf colour distribution in the polar coordinate system was investigated. Leaf images with various levels of rust severity were collected and analyzed. Validation results showed that the threshold-setting method was capable of detecting soybean rust severity under laboratory conditions, whereas the centroid-locating method had the potential to be applied in the field.

Soybean rust, caused by Phakopsora pachyrhizi, is one of the most destructive diseases for soybean production. It often causes significant yield loss and may rapidly spread from field to field through airborne urediniospores. In order to implement timely fungicide treatments for the most effective control of the disease, it is essential to detect the infection and severity of soybean rust. This research explored feasible methods for detecting soybean rust and quantifying severity. In this study, images of soybean leaves with different rust severity were collected using both a portable spectroradiometer and a multispectral CDD camera. Different forms of vegetation indices were used to investigate the possibility of detecting rust infection. Results indicated that both leaf development stage and rust infection severity changed the surface reflectance within a wide band of spectrum. In general, old leaves with most severe rust infection resulted in lowest reflectance. A difference vegetation index (DVI) showed a positive correlation with reflectance differences. However, it lacks solid evidence to identify such reflectance change was solely caused by rust. As an alternative, three parameters, i.e. ratio of infected area (RIA), lesion color index (LCI) and rust severity index (RSI), were extracted from the multispectral images and used to detect leaf infection and severity of infection. The preliminary results obtained from this laboratory-scale research demonstrated that this multispectral imaging method could quantitatively detect soybean rust. Further tests of field scale are needed to verify the effectiveness and reliability of this sensing method to detect and quantify soybean rust infection in real time field scouting.

An image-processing based algorithm to automatically identify plant disease visual symptoms

This study describes an image-processing based method that identifies the visual symptoms of plant diseases, from an analysis of coloured images. The processing algorithm developed starts by converting the RGB image of the diseased plant or leaf, into the H, I3a and I3b colour transformations. The I3a and I3b transformations are developed from a modification of the original I1I2I3 colour transformation to meet the requirements of the plant disease data set. The transformed image is then segmented by analysing the distribution of intensities in a histogram. Rather than using the traditional approach of selecting the local minimum as the threshold cut-off, the set of local maximums are located and the threshold cut-off value is determined according to their position in the histogram. This technique is particularly useful when the target in the image data set is one with a large distribution of intensities. In tests, once the image was segmented, the extracted region was post-processed to remove pixel regions not considered part of the target region. This procedure was accomplished by analysing the neighbourhood of each pixel and the gradient of change between them. To test the accuracy of the algorithm, manually segmented images were compared with those segmented automatically. Results showed that the developed algorithm was able to identify a diseased region even when that region was represented by a wide range of intensities.

Image pattern classification for the identification of disease causing agents in plants

This study reports a machine vision system for the identification of the visual symptoms of plant diseases, from coloured images. Diseased regions shown in digital pictures of cotton crops were enhanced, segmented, and a set of features were extracted from each of them. Features were then used as inputs to a Support Vector Machine (SVM) classifier and tests were performed to identify the best classification model. We hypothesised that given the characteristics of the images, there should be a subset of features more informative of the image domain. To test this hypothesis, several classification models were assessed via cross-validation. The results of this study suggested that: texture-related features might be used as discriminators when the target images do not follow a well defined colour or shape domain pattern; and that machine vision systems might lead to the successful discrimination of targets when fed with appropriate information.

An Image-Based Diagnostic Expert System for Corn DiseasesThe annual worldwide yield losses due to pests are estimated to be billions of dollars. Integrated pest management (IPM) is one of the most important components of crop production in most agricultural areas of the world, and the effectiveness of crop protection depends on accurate and timely diagnosis of phytosanitary problems. Accurately identifying and treatment depends on the method which used in disease and insect pests diagnosis. Identifying plant diseases is usually difficult and requires a plant pathologist or well-trained technician to accurately describe the case. Moreover, quite a few diseases have similar symptoms making it difficult for non-experts to distinguish disease correctly. Another method of diagnosis depends on comparison of the concerned case with similar ones through one image or more of the symptoms and helps enormously in overcoming difficulties of non-experts. The old adage a picture is worth a thousand words is crucially relevant. Considering the user's capability to deal and interact with the expert system easily and clearly, a web-based diagnostic expert-system shell based on production rules (i.e., IF < effects > THEN < causes >) and frames with a color image database was developed and applied to corn disease diagnosis as a case study. The expert-system shell was made on a 32-bit multimedia desktop microcomputer. The knowledge base had frames, production rules and synonym words as the result of interview and arrangement. It was desired that 80% of total frames used visual color image data to explain the meaning of observations and conclusions. Visual color image displays with the phrases of questions and answers from the expert system, enables users to identify any disease, makes the right decision, and chooses the right treatment. This may increase their level of understanding of corn disease diagnosis. The expert system can be applied to diagnosis of other plant pests or diseases by easy changes to the knowledge base.

Use of leaf color images to identify nitrogen and potassium deficient tomatoesSoilless culture has been popular in modern agriculture. However, plants in soilless culture often appear to be nutrient deficient. Therefore the intellective diagnostic system of plants disease of nutrients deficiency is important. In this paper, a novel idea based on computer vision is presented. Color and texture features of leaves are extracted by some methods such as percent intensity histogram, percent differential histogram, Fourier transform, and wavelet packet. Moreover, Genetic Algorithm (GA) has been used to select features to get the best information for diagnosing the disease. Experiments showed that the accuracy of this diagnostic system is above 82.5% and it can diagnose disease about 610 days before experts could determine.Grape leaf disease detection from color imagery using hybrid intelligent system

Vegetables and fruits are the most important export agricultural products of Thailand. In order to obtain more value-added products, a product quality control is essentially required. Many studies show that quality of agricultural products may be reduced from many causes. One of the most important factors of such quality is plant diseases. Consequently, minimizing plant diseases allows substantially improving quality of the products. This work presents automatic plant disease diagnosis using multiple artificial intelligent techniques. The system can diagnose plant leaf disease without maintaining any expertise once the system is trained. Mainly, the grape leaf disease is focused in this work. The proposed system consists of three main parts: (i) grape leaf color segmentation, (ii) grape leaf disease segmentation, and (iii) analysis & classification of diseases. The grape leaf color segmentation is pre-processing module which segments out any irrelevant background information. A self-organizing feature map together with a back-propagation neural network is deployed to recognize colors of grape leaf. This information is used to segment grape leaf pixels within the image. Then the grape leaf disease segmentation is performed using modified self-organizing feature map with genetic algorithms for optimization and support vector machines for classification. Finally, the resulting segmented image is filtered by Gabor wavelet which allows the system to analyze leaf disease color features more efficient. The support vector machines are then again applied to classify types of grape leaf diseases. The system can be able to categorize the image of grape leaf into three classes: scab disease, rust disease and no disease. The proposed system shows desirable results which can be further developed for any agricultural product analysis/inspection system.

Application of neural networks to image recognition of plant diseases

Digital image recognition of plant diseases could reduce the dependence of agricultural production on the professional and technical personnel in plant protection field and is conducive to the development of plant protection informatization. In order to find out a method to realize image recognition of plant diseases, four kinds of neural networks including backpropagation (BP) networks, radial basis function (RBF) neural networks, generalized regression networks (GRNNs) and probabilistic neural networks (PNNs) were used to distinguish wheat stripe rust from wheat leaf rust and to distinguish grape downy mildew from grape powdery mildew based on color features, shape features and texture features extracted from the disease images. The results showed that identification and diagnosis of the plant diseases could be effectively achieved using BP networks, RBF neural networks, GRNNs and PNNs based on image processing. For the two kinds of wheat diseases, the best prediction accuracy was 100% with the fitting accuracy equal to 100% while BP networks, GRNNs or PNNs were used, and the best prediction accuracy was 97.50% with the fitting accuracy equal to 100% while RBF neural networks were used. For the two kinds of grape diseases, the best prediction accuracy was 100% with the fitting accuracy equal to 100% while BP networks, GRNNs or PNNs were used, and the best prediction accuracy was 94.29% with the fitting accuracy equal to 100% while RBF neural networks were used.

Grading Method of Leaf Spot Disease Based on Image ProcessingSince current grading of plant diseases is mainly based on eyeballing, a new method is developed based on computer image processing. All influencing factors existed in the process of image segmentation was analyzed and leaf region was segmented by using Otsu method. In the HSI color system, H component was chosen to segment disease spot to reduce the disturbance of illumination changes and the vein. Then, disease spot regions were segmented by using Sobel operator to examine disease spot edges. Finally, plant diseases are graded by calculating the quotient of disease spot and leaf areas. Researches indicate that this method to grade plant leaf spot diseases is fast and accurate.

Image recognition of plant diseases based on principal component analysis and neural networks

Plant disease identification based on image processing could quickly and accurately provide useful information for the prediction and control of plant diseases. In this study, 21 color features, 4 shape features and 25 texture features were extracted from the images of two kinds wheat diseases (wheat stripe rust and wheat leaf rust) and two kinds of grape diseases (grape downy mildew and grape powdery mildew), principal component analysis (PCA) was performed for reducing dimensions in feature data processing, and then neural networks including backpropagation (BP) networks, radial basis function (RBF) neural networks, generalized regression networks (GRNNs) and probabilistic neural networks (PNNs) were used as the classifiers to identify wheat diseases and grape diseases, respectively. The results showed that these neural networks could be used for image recognition of these diseases based on reducing dimensions using PCA and acceptable fitting accuracies and prediction accuracies could be obtained. For the two kinds of wheat diseases, the optimal recognition result was obtained when image recognition was conducted based on PCA and BP networks, and the fitting accuracy and the prediction accuracy were both 100%. For the two kinds of grape diseases, the optimal recognition results were obtained when GRNNs and PNNs were used as the classifiers after reducing the dimensions of feature data with PCA, and the prediction accuracies were 94.29% with the fitting accuracies equal to 100%.A framework for detection and classification of plant leaf and stem diseases

We propose and evaluate a framework for detection of plant leaf/stem diseases. Studies show that relying on pure naked-eye observation of experts to detect such diseases can be prohibitively expensive, especially in developing countries. Providing fast, automatic, cheap and accurate image-processing-based solutions for that task can be of great realistic significance. The proposed framework is image-processing-based and is composed of the following main steps; in the first step the images at hand are segmented using the K-Means technique, in the second step the segmented images are passed through a pre-trained neural network. As a testbed, we use a set of leaf images taken from Al-Ghor area in Jordan. Our experimental results indicate that the proposed approach can significantly support accurate and automatic detection of leaf diseases. The developed Neural Network classifier that is based on statistical classification perform well and could successfully detect and classify the tested diseases with a precision of around 93%.Research on the color image segmentation of plant disease in the greenhouse

Image segmentation is critical to image processing and pattern recognition. Basically, color image segmentation techniques are based on monochrome ones operating in different color spaces. A color image segmentation method in the RGB color space is reported in the paper, the image segmentation is used in the image about the disease spot and the normal spot of cucumber in the greenhouse, we determine the RGB threshold value, through counting the characteristic parameter of stochastic picture element in the different image spot. The combined methods of the main factor three threshold value R, G B decide the division. this method not only retain the image real color, but also have very good division effect.

Classification of cotton leaf spot diseases using image processing edge detection techniques

This Proposed Work exposes, a advance computing technology that has been developed to help the farmer to take superior decision about many aspects of crop development process. Suitable evaluation and diagnosis of crop disease in the field is very critical for the increased production. Foliar is the major important fungal disease of cotton and occurs in all growing Indian regions. In this work we express new technological strategies using mobile captured symptoms of cotton leaf spot images and categorize the diseases using HPCCDD Proposed Algorithm. The classifier is being trained to achieve intelligent farming, including early Identification of diseases in the groves, selective fungicide application, etc. This proposed work is based on Image RGB feature ranging techniques used to identify the diseases (using Ranging values) in which, the captured images are processed for enhancement first. Then color image segmentation is carried out to get target regions (disease spots). Next Homogenize techniques like Sobel and Canny filter are used to Identify the edges, these extracted edge features are used in classification to identify the disease spots. Finally, pest recommendation is given to the farmers to ensure their crop and reduce the yeildloss.