重庆科技学院学报（自然科学版）2024，Vol.26Issue(3)：87-92,6.DOI:10.19406/j.issn.1673-1980.2024.03.014
一种改进YOLOv7-GCA的车型快速识别方法
A Fast Identification Method for Vehicle Model Based on Improved YOLOv7-GCA
摘要
Abstract
Aiming at the problems of high road traffic flow and slow model identification,a fast identification method for vehicle model based on improved YOLOv7-GCA is proposed.Firstly,the lightweight GhostConv with smaller parameter count and faster detection speed is used to replace the ordinary convolutional Conv in the network to im-prove the speed of vehicle model identification;secondly,in order to ensure the recognition accuracy of the model,the CA attention mechanism module is introduced before entering the neck.The experimental results show that the YOLOv7-GCA model reduces the number of parameters and complexity of the model and improves the vehicle mod-el recognition speed while ensuring the identification accuracy.关键词
YOLOv7/轻量化/GhostConv/CA注意力机制Key words
YOLOv7/lightweight/GhostConv/CA Attentional Mechanism分类
信息技术与安全科学引用本文复制引用
斯洪云,苏盈盈,邓圆圆,阎垒,杨浩军..一种改进YOLOv7-GCA的车型快速识别方法[J].重庆科技学院学报（自然科学版）,2024,26(3):87-92,6.基金项目
重庆市教育委员会科学技术研究项目"面向智能化工厂转型的通用型指针式仪表识别方法及实现"(KJQN202101510) （KJQN202101510）
重庆科技大学硕士研究生创新计划项目"基于YOLOv7的智能交通监测系统"(YKJCX2220419),"基于改进YOLOv7的带钢缺陷检测系统设计及其实现"(YKJCX2220408) （YKJCX2220419）
重庆科技大学本科生创新计划项目"智能交通监测系统设计及实现"(2023010) （2023010）
