Accurate pedestrian detection using RGB-D camera
Zheng Zhang, Xiaobo Li · 2017
In this paper, we present an integrated method for Pedestrian Detection using RGB-D camera. It is clearly that depth channel of the image can provide an extra info. In this paper, depth image is used to construct background model in order to produce an image ROI, and color image makes a texture partition. We combined the 2 steps to generate a novel framework. Using this framework, we deal with detection by removing shadow and occlusion problems. Also an accurate contour is generated for classification. Experimental results demonstrate the effectiveness of our method.