Citrus diseases classification and area detection using image processing

Abstract

This paper presents a technique for discovering and classifying major citrus diseases of economic importance. Because of the slight difference in symptoms of various plant diseases, the diagnosis requires the expert opinion of the disease detection. An inappropriate diagnosis may result in huge economic losses for farmers in terms of inputs such as pesticides. For several decades, computers have been used to provide automatic solutions rather than a manual diagnosis of plant diseases that are costly and error prone. three classes of citrus diseases with healthy class are discussed by using the top-hat enhancement method and the K-mean segmentation method. Finally, the MSVM method used to classify the leaf class. Also, the area of the diseases is evaluated. The results show an accuracy with 93.18%