Solving Current Limitations of GP-ELM-RNN based Plant Disease Detection and its Classification
Kavita Karambelkar, Nidhi Tomar, M. Ramkumar Prabhu, T V RAJA, S Srisathirapathy, Jalla Vamshi
Abstract
The destruction of crops by pathogens including bacteria, viruses, and fungi has been a worldwide issue plaguing farmers for centuries. To maximize production and ensure agricultural sustainability, early diagnosis and prevention of crop diseases is crucial throughout the growing, harvesting, and processing stages of the crop. This study provides a comprehensive overview of the direct and indirect methods currently used in agricultural disease diagnosis. Preprocessing, segmentation, feature extraction, and model training make up the four stages of the proposed technique. An AHE and CL-SH preprocessing. Kapur thresholding is used for segmentation. The model is trained with GP-ELM-RNN, and the features are extracted by a co-occurrence colour technique. Two competing methods, ELM and ELM-RNN, are compared and contrasted with the suggested method. When compared to more traditional methods, the proposed solution fares quite well.