基于改进YOLOv5算法和ResNet50网络的行人检测与识别系统

Pedestrian Detection and Recognition System Based on YOLOv5 Algorithm and ResNet50 Network

行人检测与识别技术在交通管理、智能监控等领域具有重要的应用价值。针对现有行人检测与识别存在的检测精度低、识别困难等问题,提出了一种融合SE注意力模块的YOLOv5算法和ResNet50网络的行人检测与识别系统。在Backbone网络中引入SE注意力模块,以捕获更加丰富的特征信息,从而提升模型的检测精度;采用ResNet50网络对裁剪图片进行识别检索。实验结果表明,该算法的检测精度较高,能够识别复杂场景下的行人,基本满足不同场景下的行人检测与识别要求。

Pedestrian detection and recognition has an important application value in traffic management,intelligent monitoring and other fields.Aiming at the problems of low detection accuracy and recognition difficulty faced by the existing pedestrian detection and recognition,this study proposes a model of pedestrian detection and recognition system that integrates the YOLOv5 algorithm of attention mechanism and ResNet50 network.By introducing the SE attention mechanism into the YOLOv5 algorithm Backbone network,richer feature information is captured,thus improving the detection accuracy of the model.The pedestrian recognition system uses ResNet50 network to recognize and retrieve the detected cropped images.The experimental results show that the algorithm has high detection accuracy and can recognize pedestrians in complex scenes,basically meeting the requirements of pedestrian detection and recognition in different scenes.