Application of image processing and machine learning techniques for detection of diseases in brinjal plant

Abstract

The computer aided detection and classification of diseases in plants is a very essential tool in the agricultural field. Most of the diseases in plants arise due to the influence of bacteria, fungi or virus. It is very difficult to identify or detect the diseases during the early stage, manually. In many of the plants, the diseases can be easily identified by observing the variations or changes in color, texture and structural properties of the leaves of the plants. But in large cultivation areas with numerous plants, such type of manual observation is very difficult. Hence, if the system is automated, it would be more helpful for the farmers, agriculturalists and pathologists for detecting and classifying the diseases effectively. newlineIn this thesis work, an automatic detection and classification of diseases in brinjal plant is proposed. The image processing techniques are used for extracting the important attributes of the leaf images. The machine learning techniques are employed for the detection and classification purposes. Initially, the leaves of the brinjal plant are captured by means of digital camera. A large set of images are collected to form a dataset. These images are further preprocessed for removing the distortion and blur regions present in the images. In the preprocessing step, the images are first denoised by applying suitable filters. The segmentation methods are applied to separate the infected areas of the leaf images. In this thesis, different types of segmentation are discussed in detail. newlineThe fine details of the image are obtained by transforming the image in different domains. newline