Title: A lightweight pest detection model based on improved YOLOv8

Authors: Zhoujian Yi; Zhen Yang; Junya Liu; Zhijian Yin; Xin Zhou

Addresses: Jiangxi Engineering Research Center of Unattended Perception System and Artificial Intelligence Technology, Jiangxi Science and Technology Normal University, Nanchang, 330038, China ' Jiangxi Engineering Research Center of Unattended Perception System and Artificial Intelligence Technology, Jiangxi Science and Technology Normal University, Nanchang, 330038, China ' Jiangxi Engineering Research Center of Unattended Perception System and Artificial Intelligence Technology, Jiangxi Science and Technology Normal University, Nanchang, 330038, China ' Jiangxi Engineering Research Center of Unattended Perception System and Artificial Intelligence Technology, Jiangxi Science and Technology Normal University, Nanchang, 330038, China ' Jiangxi Engineering Research Center of Unattended Perception System and Artificial Intelligence Technology, Jiangxi Science and Technology Normal University, Nanchang, 330038, China

Abstract: Crops are vital resources for countries worldwide, as they provide food and energy for most people on a daily basis. The yield of crops is greatly affected by pests, which can reduce the output of infected crops and spread over large areas of crops. Therefore, pest detection is crucial for crop management and protection. In this work, we propose an improved Yolov8n model for pest detection, which is based on the original Yolov8n model but modifies and enhances the network structure. We conduct experiments on the IP102 public dataset, which contains 75,000 images of 102 insect pest species. Our results show that our model achieves a 2.2% increase in mean average precision (mAP) compared to the original Yolov8n model, and has a lower number of parameters and computational cost.

Keywords: agricultural pests; object detection; YOLOv8; deep learning; lightweight model; smart agriculture.

DOI: 10.1504/IJSCIP.2025.150932

International Journal of System Control and Information Processing, 2025 Vol.4 No.4, pp.340 - 355

Received: 20 Mar 2024
Accepted: 31 Jul 2025
Published online: 05 Jan 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article