An Improved Deep Learning Model for Farm Pests Detection in Agriculture

The agriculture field is the most advanced technique and the backbone of society. The nation's economy and wealth rely significantly on agricultural production. Agriculture supplies both food and raw materials for industry. In this paper, the model works on the agriculture datasets and performs evaluations using different methods. Overall, the success of agriculture influences a broader culture and is important for long-term stability and progress. Maintaining agricultural health requires ongoing attention to potential threats from pests and diseases. The crop production data was used to analyze the current environment. This work is compared with which has applied the ReLU activation function and the ADAM optimizer for performance evaluation. This paperwork has applied different combinations of activation function and optimization function to evaluate performance for farm pests' detection against the considered peer method [18]. The paper depicts the results of deep learning architectures such as ResNet50V2, ResNet152V2, MobileNet, Xception, and EfficientNetV2B0 used to get the best accuracy and make the model more accurate to balance the class weight.