Classification of pest detection in paddy crop based on transfer learning approach

Pest recognition in the agriculture field plays a critical issue for farmers which diminishes economic growth. So far, traditional practices were followed by to increase yield production. Nowadays, researchers execute deep learning approach classify various kinds of images practically. In this paper, Deep convolutional neural networks (DCNN) are used recognise ten pests present paddy crop. The data repository contains around 3549 pest that affect crops, Since Learning supports well larger data-set so augmentation process is carried out. model build using DCNN architecture, interpretation was made over models based on accuracy rate and performance. transfer applied set fine-tuning hyperparameters layers ResNet-50 model. By comparing resultant value, fine-tuned produced better 95.012% among other models. obtained value describes effective performance disease classification.