Machine-Learning Methods for Plant Leaf Disease for Improving Agricultural Production

Agricultural production is an essential component in India’s economy. Agricultural production ensures that everyone can eat, even in cases of rapid population increase. In particular, tomatoes have emerged as India’s most valuable crop. It is recommended to anticipate plant diseases when they are in their early phases of growth so that food can be provided for all residents. Predicting disease in immature crops is a sad reality, though. The research aims to inform farmers about state-of-the-art methods for decreasing tomato leaf diseases. When machine-learning techniques applied, it can detect leaf diseases in tomato plants. The samples of sick tomato leaves are considered in this study. By analysing these diseased tomato leaf samples, farmers would be able to identify infections based on their early symptoms. After down sampling the tomato leaf samples to 256 × 256 pixels, histogram equalisation is applied to boost the tomato section’s quality. In order to partition the data into Voronoi cells, the K-means clustering algorithm is employed. The procedure of contour tracing is used to determine the boundaries of leaf samples. Discrete wavelet transforms, principal component analysis, and greyscale co-occurrence matrices are among the descriptors that can be used to extract the informative features of a leaf sample. A model is proposed, which is combination of SVM and CNN for classify the disease for tomato leaves. With an accuracy of over 99% in both the training and test datasets, the results demonstrate that the suggested model performs admirably when it comes to detecting and classifying illnesses in tomato leaves.