运用DLT法缩小三维解析数据误差的研究

Reducing data error in three-dimensional digital analysis by DLT Method

在人体运动的分析中,提取运动信息是必须的。对于人体运动的三维分析,图像系统常常是利用二维图像作三维重构,然后根据重构数据进行运动学、动力学分析。在重构技术中DLT算法由于其众多优点而得到了广泛应用。但是,由于人的运动是复杂而精细的,三维重构误差对运动分析误差的影响成为三维分析研究的关键点。针对三维重构所引起的误差,讨论了如何通过有效的方法来减小误差,以提高三维解析数据的精度,并提出了相关的方法和措施,为编制三维解析系统时提高精度提供可靠的理论依据。

It is necessary to obtain movement information in biomechanical analysis. For three-dimensional digital analysis, image systems usually rebuild two-dimensional image into three-dimensional, and then process the data for further kinematics or dynamics analysis, In rebuilding technique, DLT method is widely applied for its many advantages. However, human movement is complex and subtle. Therefore, the influences of three-dimensional rebuilding error to movement analysis become a key point of three-dimensional digital analysis. This paper discusses methods to reduce the error and provided several ways and means in order to improve data precision in three-dimensional digital analysis.