This paper proposed a method for detecting 3D pose points inside the human body based on skin multi-person linear (SMPL) model and mapped 2D pose points inside the human body from multiple perspectives to 3D pose points in real scenes using a clustering algorithm in order to address issues such as continuous modeling jitter and local distortion of model results caused by the existing methods of constructing 3D human body models based on 2D human body surface pose points. The Kalman filter is introduced to denoise the attitude points of the human body. In the process of constructing a human 3D model based on 3D pose points, this paper corrects the gradient descent regression network based on an automatic variational method and constructs an end-to-end human 3D modeling network SMPL-VAE, which is more in line with the local modeling of human motion structure while maintaining the overall proportion. The test on the open data set Shelf revealed that the attitude points could be correctly matched for various targets, and the mean position error per joint (MPJPE) was improved by 3.88, 7.56, 12.88, respectively, compared with other methods. Additionally, the percentage of correct key points (PCK) was improved by 3.5, 6.91, and 9, respectively, compared with other methods.

Many multi-object pedestrian tracking algorithms have been proposed in computer vision, and great progress has been made in tracking efficiency and accuracy recently. Practical applications are severely hampered by the fact that the majority of tracking techniques now in use are still unable to address the issues of object occlusion and reappearance in camera perspectives. To tackle the above problems in dense crowds under multi-vision, the multi-target pedestrian tracking method is based on fusion feature correlation. The feature pool was updated based on GMM to reduce feature pollution caused by dense people. To ensure the tracking universality, the similarity threshold of target features was calculated dynamically based on K-means. The similarity of fused features is used to associate the pedestrian features, with the homography constraint check to determine the addition and reappearance of pedestrians, which reduces error and miss tracking. The results of experiments using several algorithms on the public dataset Shelf indicate that the suggested method's average accuracy is 16.05% and 7.39% higher than that of other methods, while its average success rate is 16.04% and 4.16% higher. The average error tracking rate under the complete video is 10.11%, which achieves significant results in controlling mistracking and effectively associates with the original ID after the pedestrian’s reappearance.