科技创新与应用2024，Vol.14Issue(5)：7-11,5.DOI:10.19981/j.CN23-1581/G3.2024.05.002
基于改进YOLOv5s的森林烟火检测算法研究
摘要
Abstract
This paper proposes an improved forest fire detection algorithm based on YOLOv5s.The algorithm enhances the original YOLOv5s model by introducing the GSConv lightweight convolution and a strategy to eliminate grid sensitivity.Extensive experiments are conducted on a forest fire dataset,and the proposed algorithm is successfully deployed on a drone for real-world testing.The experimental results demonstrate significant performance improvements achieved by the enhanced model in forest fire detection.The average accuracy of the model is 90.65%,and the detection time is only 4.1 ms,which meets the high precision and real-time requirements of pyrotechnic detection.This study provides a strong support for the practical application of forest fire detection algorithm,and has important practical significance and application value.关键词
森林烟火检测/YOLOv5s/GSConv轻量化卷积/消除网格敏感度/实时性Key words
forest fire detection/YOLOv5s/GSConv lightweight convolution/elimination of grid sensitivity/real-time分类
农业科技引用本文复制引用
李虹,纪任鑫,陈军鹏,耿荣妹,蔡骁,张艳迪..基于改进YOLOv5s的森林烟火检测算法研究[J].科技创新与应用,2024,14(5):7-11,5.基金项目
北京市科技新星计划(Z191100001119111,Z201100006820107) （Z191100001119111,Z201100006820107）
