A Novel Method for Suppressing Shadow and Judging Pose Based on Projection Histogram

摘要

For outdoor visual surveillance scenes, a novel method based on projection histogram is proposed to suppress shadow and judge pose of human. Foreground regions are firstly extracted by background subtraction method. According to the different projection features of real object and shadow, the shadow is detected and suppressed. In order to alleviate disturbances from different human's motion poses, the upper half of human is projected to the horizon axis. Utilizing the upper projection histogram, the major axis is located. Walking poses of human are also judged according to the upper projection histogram by introducing a weighting kernel function. Experiments show that our method can suppress shadow successfully and can improve the accuracy of locating human's major axis, also can judge human's walking pose by 94% accuracy. The algorithm can be implemented with completely real-time speed.