Insect pest infestations are serious threats to global crop production and agricultural quality, so they need effective monitoring systems. Traditional machine learning algorithms like Support Vector Machines (SVM), k-Nearest Neighbors (k-NN), and Random Forests classify data based on manually crafted geometric, textural and color features. These methods struggle with changing lighting and background noise in the field. To address these issues, we propose a unified edge computed precision framework based on the YOLOv5 object detection architecture. This lightweight deep learning model overcomes the limits of manual feature extraction, and achieves feature extraction, multi-scale localization and classification in one go. Results on the prepared IP102 benchmark subset show that the trained YOLOv5s detector achieved an overall precision of 60.6%, recall of 58.4%, mAP@0.5 of 64.7%, and mAP@0.5:0.95 of 41.3% after 30 training epochs. The proposed system integrates this model with a low-cost embedded platform and a conditional response mechanism to assist farmers in detecting pests early, reducing excessive pesticide use, and promoting sustainable farming practices.

Lovepreet Kaur, Kuldeep Kumar, "Low-Cost Real-Time Pest Detection and Conditional Response System" International Journal of Computer Science and Engineering Development, V1(4): Page(115-121) September - October 2026. ISSN: 3139-0862. www.ijcsed.com. Published by Scientific and Academic Research Publishing.