计算机与数字工程2025，Vol.53Issue(3)：901-906,6.DOI:10.3969/j.issn.1672-9722.2025.03.048
基于改进轻量级YOLOv4-tiny的轮胎缺陷检测
Tire Defect Detection Based on Improved Lightweight YOLOv4-tiny
赵蒙蒙 1张岩1
作者信息
- 1. 青岛科技大学机电工程学院 青岛 266061
- 折叠
摘要
Abstract
Against the problem that the current model is hard to achieve a balance between tire defect detection accuracy and speed,this paper proposes an improved lightweight YOLOv4-tiny network for tire defect detection.The ability to represent multi-scale objects is improved by improving the feature fusion part of the network without adding too much computational cost.In the meantime,combined with the super channel attention ECA-Net,the convolutional neural network can better pay attention to im-portant features,enhance the expression of tire defect features,and weaken irrelevant features such as tire texture background.Ex-perimental results show that the proposed method achieves an mAP of 95.12%on the tire defect dataset.The average detection time of tires is 18.97 ms,and the model parameters are small and can be easily deployed,so this method can meet the needs of industrial real-time detection.关键词
轮胎缺陷检测/YOLOv4-tiny/注意力机制/目标检测Key words
tire defect detection/YOLOv4-tiny/attention mechanism/target detection分类
计算机与自动化引用本文复制引用
赵蒙蒙,张岩..基于改进轻量级YOLOv4-tiny的轮胎缺陷检测[J].计算机与数字工程,2025,53(3):901-906,6.
