基于步态触觉信息的运动轨迹图像分割方法

MOTION TRAJECTORY IMAGE SEGMENTATION METHOD BASED ON GAIT TACTILE INFORMATION

人体在运动过程中,由于运动幅度和速度并没有遵循特定的规律,无法提取出运动轨迹的实际特征。为此,提出了一种基于步态触觉信息的运动轨迹图像分割方法,利用人体足跟与地面接触的脚印信息,获取步态触觉信息,通过轮廓检测函数计算图像轮廓的内外平均灰度值,利用形状特征因子,确定图像区域,提取出运动轨迹图像的初始轮廓。将图像的像素梯度作为一个二维向量,计算像素点梯度的灰度值,通过差分处理得到运动轨迹图像在灰度等级区间内的直方图,在离散形式下,增强运动轨迹图像轮廓。将增强处理之后的图像轮廓作为分割边界,确定前景区域和背景区域,实现运动轨迹图像的分割。实验结果表明,该方法能够根据人体运动过程中的步态触觉信息,分割运动轨迹图像,并将分割均匀度提高到80%以上。

In the process of human movement,because the motion amplitude and speed do not follow the specific law,the actual features of the motion track can’t be extracted.Therefore,a motion track image segmentation method based on gait tactile information is proposed.The gait tactile information was obtained by using the footprint information of human heel contact with the ground.The average gray value of the image contour was calculated by the contour detection function.The image region was determined by the shape feature factor,and the initial contour of the motion trajectory image was extracted.The pixel gradient of the image is taken as a two-dimensional vector,the gray value of the pixel gradient is calculated,and the histogram of the motion track image within the gray level interval is obtained through differential processing.In the discrete form,the contour of the motion track image is enhanced.The contour of the enhanced image is used as the segmentation boundary to determine the foreground area and the background area to realize the segmentation of the motion trajectory image.The experimental results show that the proposed method can segment the motion track image according to the gait and tactile information in the process of human movement,and the segmentation uniformity can be improved to more than 80%.