摘要
针对边缘计算设备算力和存储空间有限的问题,提出了一种基于YOLOX改进的轻量级人体跌倒检测算法。首先采用GhostNet中的Ghost模块降低YOLOX中Neck和Prediction层的卷积参数冗余;其次,在Neck层中添加坐标注意力机制,增强关键信息提取能力,减少背景带来的噪音影响;最后,针对轻量级模型检测头检测能力不足问题,引入辅助头模块以加强轻量检测头的学习能力。通过算法检测性能以及在边缘计算端NVIDIA Jetson Xavier NX运行实验,结果显示,所提模型的mAP@0.5达到84.9%,且模型大小为25.6 MB。相较于YOLOX模型,仅以牺牲少量推理速度提升了4.6%的检测精度,减少了25.4%的模型大小,另外与一些主流目标检测算法相比,也具有一定的优越性。这些结果表明所提模型能够更好地满足边缘计算设备在人体跌倒检测中对轻量化和准确性的需求。
Aiming at the problem of limited computing power and storage space of edge computing devices,a lightweight hu-man fall detection algorithm based on YOLOX was proposed.Firstly,Ghost module in GhostNet is used to reduce the redundancy of convolution parameters in Neck and Prediction layers in YOLOX.Secondly,the coordinate attention mechanism is added to the Neck layer to enhance the ability of key information extraction and reduce the influence of background noise.Finally,aiming at the problem of insufficient detection ability of lightweight model detection head,an auxiliary head module is introduced to strengthen the learning ability of lightweight detection head.The performance of the proposed model is tested by the algorithm and the experimental results are run on the edge computing end of NVIDIA Jetson Xavier NX.The results show that the mAP@0.5 of the proposed model reaches 84.9%,and the model size is 25.6 MB.Compared with the YOLOX model,only a small amount of reasoning speed is sacrificed to improve the detection accuracy of 4.6%and reduce the model size of 25.4%.In addition,compared with some mainstream object detection algorithms,the proposed model also has certain advantages.These results show that the proposed model can better meet the requirements of lightweight and accuracy of the model for edge computing equipment in human fall detection.
作者
龙艺文
闵宣霖
陈奕兆
罗欢
刘洪
易军
Long Yiwen;Min Xuanlin;Chen Yizhao;Luo Huan;Liu Hong;Yi Jun(College of Intelligent Technology and Engineering,Chongqing University of Science and Technology,Chongqing 401331,China;College of Safety Engineering,Chongqing University of Science and Technology,Chongqing 401331,China)
出处
《电子测量技术》
北大核心
2023年第10期109-116,共8页
Electronic Measurement Technology
基金
油气生产安全与风险控制重庆市重点实验室开放基金(CQSRC202115)
重庆科技学院硕士研究生创新计划项目(YKJCX2120809)资助。
