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.
