Plants have been cultivated for food, medicine, clothing, shelter, fiber and beauty for thousands of years. Fungus, bacterial and viruses are the cause of plant disease. Therefore, this problem requires an automatic detection of plant disease Because of pests; around 18% of crop yields are lost every year. A traditional method of plant diagnosis is not efficient and incredible. At present, a block is used on a farm, which has a number of different sensors and a device is used to convert and change data for monitoring and control purposes. In the proposed system, for the identification of leaf disease two phases are used. First phase is detection of leaf disease and the second phase is encryption techniques for security purpose. In first phase, the leaf disease detection phase has four steps, such as Preprocessing, Segmentation, Feature Extraction and Classification. The segmentation used here is the Otsu’s threshold based segmentation. While using the Otsu’s threshold based segmentation we get better result when compared to the previous method. In feature extraction here the feature is extracted using the ABCD feature. And classification is performed using the Support Vector Machine based particle swarm optimization (SVM-PSO), which is used to categorize the leaf disease separately. In the second phase, International Data Encryption Algorithm and Tiny Encryption Algorithm Encryption techniques are used to encrypt and decrypt our image in order to maintain privacy. Our experimental results show that accuracy level and performance is improved. The simulations are done on MATLAB application.
Volume 11 | 09-Special Issue
Pages: 738-750
DOI: 10.5373/JARDCS/V11/20192628