Smart pest detection for an agricultural field crop based on deep learning

Abstract

A study was conducted to develop smart pest detection for an agricultural field crop based on deep-learning object detection. The study selected the agricultural field crop pumpkin, and the red beetle pest was detected. The study developed a deep learning-based object detection model using the YOLOv8l. Roboflow was used as the conversion tool for customized data preparation. The performance and accuracy of the model were found to be satisfactory. The integration of the model with the web application was done for real-time pest detection. The proposed approach has the potential to aid farmers in identifying the existence of pests, thereby diminishing the duration and resources needed for farm inspection. The YOLOV8l object detection model was implemented for the purpose of pest classification, localization, and quantification. The proposed pest detection approach demonstrated a noteworthy increase in performance in terms of precision (P) 86%, mean average precision (mAP) .89, F1-score 86.9%, and recall 88%. A web application was developed to aid farmers and agricultural professionals in real- time pest detection. The study concluded that integrating deep learning techniques holds immense promise for revolutionizing smart pest detection in agriculture. By harnessing the power of artificial intelligence, farmers can transition towards more sustainable and efficient pest management practices, contributing to food security, environmental conservation, and economic prosperity.