基于行人重识别(ReID)技术的乘客出行特征研究

Research on Passenger Travel Characteristics Based on Person Re-identification (ReID) Technology

随着大数据时代的来临,海量的非结构化视频数据激增,传统的人工处理方法已经难以满足现有的数据处理需求。随着人工智能、神经网络等技术的发展和算力的提升,视频分析技术应运而生。行人重识别(ReID)算法是智能视频分析中的一项重要技术,标注人员首先对视频样本进行标注作业,然后将标注结果应用到算法模型中,可以快速从海量的监控视频中识别出不同摄像头网络中的同一个行人,最后通过自动检索目标,精确定位每一位乘客的运行轨迹,实现对客流的精确划分。

In recent years,with the development of technologies,such as artificial intelligence and neural networks,and the continuous improvement of computing power,the number of cameras has increased rapidly,which generating massive amounts of video data.The traditional manual processing and manual monitoring methods can no longer meet the needs of many data processing.Therefore,video analysis technology emerges as the times require.Due to the characteristics of closed underground,such as poor lighting,the large number of stairs,narrow space,and large passenger flow.The algorithm accuracy tends to drop significantly model algorithms once applied to the subway scene,although those algorithms can perform well in other fields.Person Re-Identification(ReID)algorithm is an important technology in intelligent video analysis.Annotators first perform labeling operations on video samples,and then apply the labeling results to the algorithm model.This technology can identify the same person in different camera networks,accurately locate the running trajectory of each passenger,track the passenger trajectory,and achieve accurate division of passenger flow from massive surveillance videos by automatically retrieving the query target.