Paddy Crop Yield Prediction and Early Sheath Blight Disease Identification Using an Improved Grasshopper Optimization Algorithm-Based CNN (IGOA-CNN)

Abstract

Therefore, in order to prevent or increase crop production, it is crucial to identify and successfully detect paddy crop illnesses at every step of a plant's life cycle. Initially, image preprocessing is done using bilateral filtering (BF) techniques. The semantic segmentation method is then used to segment the pre-processed images. The extracted segments are subjected to feature extraction. The feature data is extracted using the Scale-Invariant Feature Transform (SIFT) model. In this performance, an proposed Improved Grasshopper Optimization Algorithm-based Convolutional Neural Network (IGOA-CNN) with stochastic optimization method is described for both classifications and detection system of datasets gathered from the Kaggle Database at different stages of life, images that enable early detection of paddy crop yield infection even before the development of severe disease symptoms and healthy and Sheath Blight Disease. In order to reduce the use of chemicals and improve crop protection and productivity, an automatic pesticide spraying system is activated based on the detection results to only apply pesticides in affected areas. Paddy Crop Yield accuracy IGOA-CNN model: The proposed model was used to carry out the detection achieves 99.2% accuracy rate was attained.