Classification of Cotton Diseases Using Cross Information Gain_minimal Resource Allocation Network Classifier with Particle Swarm Optimization

This paper is developed based on machine vision system and data mining techniques to identify the cotton leaf spot diseases. The leaves are most probably affected by the fungi, viral and bacterial diseases in the leaf spot areas which plays a vital role of crop situation. This paper clarifies six types of diseases in the cotton plant. The significance of this research work design is based on advanced computational techniques to reduce the complexity, cost and time. The proposed techniques correctly identify the diseases. In preprocessing, the image resolution value is resized to the 150* 150 pixels. The paper uses Enhanced Particle Swarm Optimization [EPSO] for feature selection to identify the affected region of a leaf. The Proposed Skew divergence (statistical method) is based on calculating the edge, color, texture variance features for analysis of the affected part of a cotton leaf. The Proposed Cross Information Gain of Minimal Resource Allocation (CIG-MRAN) Classifier has been used to classify the six types of diseases and increases the accuracy of the classification system.