科技创新与应用2025，Vol.15Issue(27)：34-37,41,5.DOI:10.19981/j.CN23-1581/G3.2025.27.007
基于改进YOLOv8的遮挡车辆目标检测算法研究
刘广宇 1郑冠天 2严伟 3雷新卓3
作者信息
- 1. 华中科技大学 集成电路学院,武汉 430074||江苏北方湖光光电有限公司,江苏 无锡 214194
- 2. 华中科技大学 集成电路学院,武汉 430074
- 3. 江苏北方湖光光电有限公司,江苏 无锡 214194
- 折叠
摘要
Abstract
The improved YOLOv8 algorithm uses a combination of Puzzle Mix and Mosaic to replace the original algorithm Mosaic data enhancement method,which improves the problem of low accuracy of the YOLOv8 algorithm in detecting occluded objects.Two attention modules,CBAM and SENet,are introduced into Backbone to improve the problem of difficulty in accurately detecting under complex weather and mutual occlusion.When calculating regression loss,the EIoU+Soft_NMS method is used to solve the problem of missing occluded objects and slow convergence speed.The improved YOLOv8 algorithm inherits the efficiency of the original algorithm and improves the detection accuracy and accuracy of vehicle targets.关键词
目标检测/YOLOv8/注意力/准确率/遮挡车辆Key words
target detection/YOLOv8/attention/accuracy/blocking vehicle分类
信息技术与安全科学引用本文复制引用
刘广宇,郑冠天,严伟,雷新卓..基于改进YOLOv8的遮挡车辆目标检测算法研究[J].科技创新与应用,2025,15(27):34-37,41,5.
