基于下肢关节点运动关系的特定人群步态识别

Gait Recognition for Specific Population Based on the Movement Relationship of Lower Limbs Joint Points

在视频犯罪侦查领域,肢体关节点运动是视频监控中人体行走的重要特征,客观量化关节点空间运动关系是视频侦查领域的热点问题。本文以高犯罪年龄段人群为研究对象,进行了正常行走状态的视频采集,然后以Fitts定律为基础提出新型综合对称指标,验证了视频中步态行走过程的稳定性,最后利用变异系数法和正太性检验确定了个体行走过程中6个稳定的关节关系。

In the field of video crime detection,the movement of joints is an important feature of human walking in video surveillance,and the objective quantification of the spatial motion relationship of joints is a hot issue in the field of video surveillance.Taking the high crime age group as the object,video acquisition of normal walking state is carried out.And then a new symmetry index is proposed based on Fitts Law,the stability of walking process in video is verified.At last,six stable joint relationships in the same person’s walking process were determined by the method of coefficient of variation and the test of normality.