Enhanced weighted sum back propagation neural network for leaf disease classification

Abstract In an agricultural country like India, farmers have facilities to cultivate different crops. However, cultivation of the crops for better yield and quality production can be made easy with the support of technology. Image processing plays a vital role in detecting a plant disease and will provide a support to the research. In reality, we can find fast, automatic, cheap and accurate solutions by using image processing techniques. This work provides a significant and an efficient system for the classification of leaf diseases. During segmentation, the infected part of the leaf is recognized by using Otsu’s threshold method. The extraction of features, from the segmented image is performed using Gray Level Co-occurrence Matrix (GLCM). The features hence derived are used as the input to the classifier. Enhanced Weighted Sum Back Propagation (EWSBP) Algorithm is used in the Multilayer Perceptron Neural Network (MPNN) for classification. This classifier provides better performance compared to other leaf disease classification methods.