基于Kinect深度图像信息的人体运动检测

Human motion detection based on the depth image of Kinect

人体运动检测是计算机视觉人体运动分析的关键环节。根据Kinect深度图像的特点,引入并改进Vi Be算法处理深度图像进行人体运动检测。考虑到深度图像中地面像素值连续性造成的地面附近运动检测困难,提出了一种自适应的图像分层处理和不同邻域模式的建模方式,增加了去除"鬼影"现象的参考模型。像素分类时增加了前景点检验步骤,通过当前像素与参考模型的比较消除"鬼影"。在模型更新方面增加了基于前景点的背景模型更新策略,解决了"黑影"现象问题。采用阈值法对分类结果进行了误检点消噪处理。实验结果表明所提出的改进Vi Be算法能够比较准确地检测出人体运动。

Human motion detection is a key step of human motion analysis based on computer vision.According to the characteristics of Kinect depth image,improved ViBe algorithm is used to detect human motion in depth image.Considering the motion near the ground is difficult to detect due to the continuity of pixel values of depth image,a modeling method based on adaptive hierarchical image processing and different neighborhood patterns is proposed.A reference model for removing the ＂ghost ＂ phenomenon is presented.A foreground detection process is added in pixel classification,which can remove the ＂ghost＂phenomenon through comparing the current pixel with the reference model.In terms of model update,a foreground pixels based background model update strategy is adopted to solve the ＂shadow＂phenomenon problem.The classification result is denoised using the threshold method to remove incorrect detected pixels.Experimental results show that the improved ViBe algorithm can detect the human motion accurately.