基于Kinect的人体单关节点修复算法研究

Research on the Algorithm of Human Single Joint Point Repair Based on Kinect

人体姿势识别是人机交互发展中的关键技术,基于Kinect的人体姿势识别技术已成为研究热点。但人体在运动的过程中,经常会出现关节点被身体其他部位遮挡的情况,导致识别准确率下降。本文通过对人体单个被遮挡点进行研究,根据人体运动连续性和关节点自由度约束,提出了一种人体单关节点修复算法。经实验验证,本算法可有效修复单个被遮挡关节点,提高人体姿势识别准确率。

Human posture recognition is the key technology in the development of human computer interaction, and the human body posture recognition based on Kinect has become a hot research topic. However, in the process of the human body movement, the joints are often covered by other parts of the body, resulting in a decline in the recognition accuracy rate. Based on the research of human single occluded points, according to the continuity of human movement and joint freedom constraint, proposes a joint point repair algorithm for human body. The experimental results show that this method can effectively repair the occluded joint points and improve the accuracy of human posture recognition.