Pest Detection and Prevention for Agricultural Crops Using YoloV8 Algorithm

Agriculture's productivity depends on seasonal rains, weather, and pesticides. Due to a lack of knowledge about pests, farmers are unable to take preventive measures within the stipulated time, resulting in decreased crop productivity. Here, pest detection is very helpful in increasing food quality and also helps the country's economy. This paper describes the development of a deep learning-based agricultural insect detection system. A web-based application and an algorithm for vermin detection comprise the system. The YOLOV8 architecture uses convolutional neural networks (CNN), bottlenecks, convolution to flatten (C2F), and detection blocks for feature extraction, representation learning, and pest detection. Finally, the system was implemented in Python in the Visual Studio 2022 environment, and a web application was developed using the Streamlit library. The system detects the uploaded pest images, and identifies the diseases, and suggests remedies on the web application. The system constructed to detect 10 kinds of agricultural pests inhabiting the crops, such as citrus, paddy, cotton, mango, tomato, lady's finger, and pomegranate. The findings demonstrate 95%% efficacy. It is better than Resnet50,EfficientDet and Yolo's previous versions.