A Prediction of Rice Leaf Disease using KSC

Agriculture is the primary means of providing food for all the people. According to a comprehensive study, the spread of plant diseases has a significant impact on agriculture. As a result, in order to predict the disease, automatic disease detection techniques should focus more on the early indicators of rice plant diseases. This study intends to perform an extensive evaluation of numerous image processing techniques to detect rice plant diseases. This study develops a novel technique for mechanically classifying diseases in rice plants by utilizing image processing techniques. To achieve accurate feature extraction, this study suggests a K-means clustering for segmenting the diseased section from a leaf image and a Support Vector Machine (SVM) is used to extract different characteristics for multi-class classification under the three categories of colour, shape, and texture. The proposed system detects leaves and probably disease-induced lesions using image processing algorithms. With a test accuracy of 100%, the XceptionNet deep feature with SVM classifier outperforms the other existing models that have undergone DMD preprocessing. The effectiveness of the proposed DMD-based attention-driven pre-processing was also examined by using the images of real rice leaves. Comparing the XceptionNet model to other transfer-learned models, it has a higher classification accuracy of 94.33%.