Early Pest Identification in Greenhouse Crops using Image Processing Techniques

Early disease detection is a major challenge in agriculture field. Hence proper measures has to be taken to fight bioagressors of crops while minimizing the use of pesticides. The techniques of machine vision are extensively applied to agricultural science, and it has great perspective especially in the plant protection field,which ultimately leads to crops management. Our goal is early detection of bioagressors. The paper describes a software prototype system for pest detection on the infected images of different leaves. Images of the infected leaf are captured by digital camera and processed using image growing, image segmentation techniques to detect infected parts of the particular plants. Then the detected part is been processed for futher feature extraction which gives general idea about pests. This proposes automatic detection and calculating area of infection on leaves of a whitefly (Trialeurodes vaporariorum Westwood) at a mature stage.