Classification of Paddy Plant leaf diseases using optimized Support Vector Machine

Plant disease detection is important for several reasons, such as forecasting crop loss, tracking and forecasting infections, identifying host resistance, and researching fundamental biological interactions between hosts and pathogens. Yields can be increased by early identification of plant leaf disease. If automated methods exist to assist farmers in identifying rice illnesses based on images of plant leaves, then this may be feasible. Here, therefore Optimized SVM is proposed. Pre-processing, feature extraction, classification, and segmentation are the four steps of the suggested methodology. The input picture is first converted into red, green, and blue formats. As part of the pre-processing stage, the noise in the green band is eliminated by applying a median filter. Next, utilizing the Texture and Color histogram characteristics, significant features of the green band are retrieved. Following feature extraction, an image is classified as Healthy, Bacterial Blight, Tungro, or Leaf Blast by feeding the retrieved features into the Optimized Support Vector Machine (OSVM) classifier. Adaptive Sunflower Optimization (ASFO) method is used to determine SVM parameters best to maximize SVM performance. A level set segmentation technique is then used to further identify the affected area. Accuracy, sensitivity, and specificity are the metrics used to evaluate how effective our work is. For rice disease data sets and rice leaf disease images, the suggested method’s optimized SVM achieved a maximum accuracy of 97.01%.