Self-adaptive human body 3D point cloud generation method based on 4D millimeter wave radar

The invention discloses a self-adaptive human body 3D point cloud generation method based on a 4D millimeter wave radar. The method specifically comprises the following steps: acquiring original point cloud data of different movement postures of a human body by using the millimeter wave radar; preprocessing the radar original point cloud data to obtain preliminarily detected target information; according to dynamic and static information in the preliminary detection information, in combination with human respiration and other characteristics, accurately positioning each human body position, removing most multipath interference, and obtaining roughly estimated human body data; performing multi-frame dynamic accumulation on similar targets to improve the point cloud density of the targets; and finally, further removing multipath noise through a designed parameter adaptive DBSCAN algorithm, and obtaining accurate human body 3D point cloud data. According to the human body posture point cloud generation method, t