Plant Leaf Disease Identification Supported by Image Segmentation, Feature Extraction and Ensemble Classification

Abstract

Plant health remains a vital part of agricultural science. Production quality severely falls in such scenarios. Most often, the disease makes the leaves as their first victim as these are one of the delicate parts of the plant. An attempt has been made to create a system that identifies diseases with least human intervention. Computer vision does it all within seconds. This paper offers an insight into how different image processing techniques that comprise segmenting the image, extracting features and classifying can be employed for the aim identification of diseases in plant leaves.Several state-of-the-art techniques have been analyzed and hence problem formulation has been put forward and methodology has been proposed. In methodology, k means segmentation, law’s textural measures and ensemble classifiers were focused upon. For extracting better and more features, Laws’ textural mask will be applied. The strong classifier method can be used for the detection of plant disease detection. Ensemble based classifiers are the combination of two or more classifier. Plant Village dataset is focused to be considered. The paper will assist the investigators to grasp concepts of numerous techniques for leaf disease identificationand an insight into the problem and future opportunities in the stated field.