CAG-YOLO:轻量级网球检测  

CAG-YOLO:Lightweight tennis detection

作　　者：赵雨欣 杨武[1] 李迎江 卢玲[1] ZHAO Yu-xin;YANG Wu;LI Ying-jiang;LU Ling(College of Computer Science and Engineering,Chongqing University of Technology,Chongqing 400054,China)

机构地区：[1]重庆理工大学计算机科学与工程学院,重庆400054

出　　处：《计算机工程与设计》2024年第6期1822-1828,共7页Computer Engineering and Design

基　　金：国家社会科学基金项目(2017CG29);重庆市教育科学规划基金项目(2021-GX-363);重庆市研究生科研创新基金项目(CYS22661)。

摘　　要：为实现智能网球回收机器人的高精度实时网球检测,提出一种轻量级网球检测算法CAG-YOLO。提出融合坐标注意力的Ghost残差块(coordinate attention ghostbottleneck, CAG),构建轻量型骨干网络CAG-Backbone,采用加权双向特征金字塔网络加强特征融合。采用SCYLLA-IoU计算坐标回归损失,改进非极大值抑制的后处理方法解决网球重叠问题。算法在Wtennis数据集上的实验结果表明,CAG-YOLO较基线方法的精度提高8.6%且模型体积减少31.7%,检测速度为21 ms,性能优于其它算法。CAG-YOLO能够用小规模参数提升检测精度,易于移植至移动智能设备。To realize the high-precision and real-time tennis detection of intelligent tennis recycling robot,a lightweight tennis detection algorithm CAG-YOLO was proposed.Coordinate attention ghostbottleneck(CAG)was proposed and a lightweight backbone network CAG-Backbone was constructed.Bidirectional feature pyramid network was used for feature fusion.SCYLLA-IoU was used to calculate the coordinate regression loss,and an improved post-processing method of non-maximum suppression was proposed to solve the problem of tennis overlap.Results of the experiment on Wtennis dataset show that the accuracy of CAG-YOLO is improved by 8.6%and the model volume is reduced by 31.7%compared with baseline method.The detection speed is 21 ms,outperforming that of the competitive algorithms.It verifies that CAG-YOLO can improve the detection accuracy with small-scale parameters and is easy to be transplanted to mobile intelligent device hardware.

关 键 词：目标检测 网球回收 深度学习 鬼影残差块 坐标注意力机制 双向特征金字塔 非极大值抑制 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]