基于立体视觉和Kinect相机的人体三维建模算法研究

打开文本图片集

中图分类号：TP391.41 文献标识码：A 文章编号：1674-0033（2025）04-0047-05

Abstract：In view of the limitedaccuracy of humanbody 3Dmodeling technology，such as the lack of accurate projection to human body image contour and pose depth information，a human body 3D modeling algorithm is proposed based on stereo visionand Kinect camera.By matching the point cloud data set collected by Kinect camera with the initial human body model established under stereo vision technology， the model parameters are optimized constantly to realize the 3D modeling of human body based on a single image.The experimental results show that the root-mean-square eror of the proposed model is reduced by 31% on average compared with the Hmr algorithm，and the proposed model can be more accurately projected into the human body contour in the color image，and a more accurate human body 3D model can be reconstructed.

Key words：human body 3D modeling;computer vision;Kinect

随着计算机视觉和图形处理技术的迅速发展，基于图像的三维人体建模技术在诸多领域有着越来越广泛的应用。（剩余5984字）