人体高速运动图像步态特征精准定位方法研究  

Research on precise gait feature location method in high-speed motion image of human body

作　　者：谢颖[1] 赵蕾[1] XIE Ying;ZHAO Lei(Xi'an Institute of Translation,Xi'an 710105,China)

机构地区：[1]西安翻译学院体育部

出　　处：《自动化与仪器仪表》2019年第11期25-28,共4页Automation & Instrumentation

基　　金：国家社会科学基金项目《我国学龄儿童(7-15岁)动作学习能力评级指标与发展常模研究》(No.17XTY007)

摘　　要：人体高速运动图像步态特征定位受到人体高速运动环境以及动态特征边界因素的影响,容易产生步态特征定位误差。为了提高人体高速运动图像步态特征定位能力,需要进行步态特征模糊计算和识别处理,提出一种基于全局步态拟合和动态分块匹配的人体高速运动步态特征精准定位方法。构建人体高速运动图像步态特征定位的几何网格区域模型,采用模糊动态特征分割方法进行人体高速运动步态特征定位特征点重构,提取人体高速运动步态特征定位动态约束特征量,构建灰度直方图,根据灰度像素集进行动态步态行为特征重建,根据灰度像素的分布概率密度进行步态特征量融合,实现人体高速运动步态特征定位和信息增强处理。仿真结果表明,采用该方法进行人体高速运动图像步态特征定位的精度较高,步态定位过程收敛性较好。The gait feature localization of human body high speed moving image is affected by the high speed movement environment and the dynamic feature boundary factor,which is easy to produce gait feature localization error.In order to improve the gait feature localization ability of high-speed moving human body image,it is necessary to carry out fuzzy calculation and recognition of gait feature.A precise gait feature location method based on global gait fitting and dynamic block matching is proposed.The geometric mesh region model of gait feature localization in high-speed moving human body image is constructed,and the feature points of human gait feature location are reconstructed by using fuzzy dynamic feature segmentation method.The gait feature of human body is extracted to locate the dynamic constraint feature,the gray histogram is constructed,the dynamic gait behavior feature is reconstructed according to the gray pixel set,and the gait feature quantity is fused according to the distribution probability density of gray pixel.The gait feature localization and information enhancement are realized.The simulation results show that the proposed method has high accuracy and good convergence in the gait localization process of high-speed moving human body images.

关 键 词：人体高速运动图像 步态特征 精准定位 灰度直方图 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]