Optimization based hybrid approach for grape leaf disease classification and pesticide detection.

Grape is one of the most cultivated fruit crops in India and widely used to produce the food items. Among the world, grapes are considered as most significant fruit and it comprises various nutrients, namely Vitamin C. However, grape plants suffer with diseases which extremely affect good quality grape yield. The major six general grape leaf disease and pests are brown spot, leaf blight, downy mildew, anthracnose, and black rot, which produces severe economic suffers to grape commerce. Thus, the early detection of grape leaf disease is highly needed to produce healthy grape yield. Moreover, computer vision driven approaches have been employed effectively in modern years for detecting and classifying plants diseases. In this paper, hybrid approach with fuzzy logic and neural network has been proposed to classify the grape leaf disease. The work is further extended to identify whether leaf is affected with pesticide or not and percentages of pesticide is calculated if it is affected with pesticides. Optimization approach has been used to find accurate results. This optimization based deep learning approach achieved the accuracy of pesticide identification 93.43%. [ABSTRACT FROM AUTHOR]