Human body behavior recognition method based on depth map and skeleton points

The invention provides a human body behavior recognition method based on a depth map and skeleton points, and the method comprises the steps: carrying out the different-scale segmentation of a behavior sequence through employing a time pyramid, and reserving the time sequence information in a behavior; only relevant data of important parts which contribute much to behaviors is used for feature extraction, similar data in different behaviors is removed, and the feature'purity 'is high; the distribution condition of the motion trails of the important parts of the human body in the space is accurately expressed in a space sub-grid dividing mode. Practical application shows that the features extracted by the method have good discrimination for human body behavior recognition. 本发明提供了一种基于深度图和骨骼点的人体行为识别方法，使用时间金字塔对行为序列进行不同尺度的分割，保留了行为内部的时序信息；只使用对行为贡献大的重要部位的相关数据进行特征提取，去除了不同行为中较为相似的数据，特征"纯度"高；通过划分空间子格的方式，较为精确的表达了人体重要部位的运动轨迹在空间中的分布情况。实际应用情况表明，该方法提取的特征，对于人体行为识别具有较好的区分度。