Crop pests are the major reason and responsible for social, economic, and environmental losses around the world. Agricultural scholars are focused on deep learning models because farmers have shown prominent trust in image recognition. In this research work, a new pest detection framework is proposed to enhance high quality crop production. Initially, IoT-based data collection process is held to properly gather the necessary agricultural data and it is fed into a Vision Transformer with a Mask Region-based Convolutional Neural Network (ViT-MR-CNN) for the segmentation process. Crop image segmentation provides highly accurate pest classification for improving the yield of crops. Finally, classification is performed with the help of Adaptive Multi-dilated EfficientNet (A-MENet) to timely classify the pest in the crop. It helps farmers to lessen crop damage and increase the yield of crops for mitigating economic loss. An Enhanced Gazelle Optimization Algorithm (EGO) model is designed to optimize the attributes from the designed A-MENet for maximizing the reliability and convergence speed. Finally, the classified outcome is contrasted over several conventional pest classification frameworks to ensure the effectiveness of the pest classification process. In the experimental performance, the recommended model has achieved reliable outcomes like 96% sensitivity, 96% specificity, and 99% NPV value in terms of 1st K-fold to significantly prove its better classification performance.
Madhu et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: