在步态识别中,常用轮廓信息来描述人体运动的细节,而单像素人体轮廓是决定步态识别精度的一个关键环节.提出了一种单像素人体轮廓提取方法.首先,使用灰度图像背景差法获取人体区域图像;然后利用数学形态学的方法对人体区域图像进行修复处理;最后,利用边缘像素邻域信息提取单像素的人体轮廓.仿真实验结果表明,提出的方法能够有效地提取不同步态图像的单像素人体轮廓.
Human body contour is commonly used to describe the detail of human motion information in gait recognition,so the extraction of single pixel body contour is the key for the accurate ratio of gait recognition.A method of single pixel body contour extraction is proposed.Firstly,the human area image is obtained by using background subtraction method; secondly,the human are image is repaired based on mathematical morphology; finally,the single pixel body contour is extracted based on the adjacent information of the edge pixels.The simulation experimental results show that the method proposed can extract single pixel body contour for different gait image.