Baca A
Department of Biomechanics, IfS, University of Vienna, Wien, Austria.
J Biomech. 1996 Apr;29(4):563-7. doi: 10.1016/0021-9290(95)00033-x.
A method has been developed for the precise determination of anthropometric dimensions from the video images of four different body configurations. High precision is achieved by incorporating techniques for finding the location of object boundaries with sub-pixel accuracy, the implementation of calibration algorithms, and by taking into account the varying distances of the body segments from the recording camera. The system allows automatic segment boundary identification from the video image, if the boundaries are marked on the subject by black ribbons. In connection with the mathematical finite-mass-element segment model of Hatze, body segment parameters (volumes, masses, the three principal moments of inertia, the three local coordinates of the segmental mass centers etc.) can be computed by using the anthropometric data determined videometrically as input data. Compared to other, recently published video-based systems for the estimation of the inertial properties of body segments, the present algorithms reduce errors originating from optical distortions, inaccurate edge-detection procedures, and user-specified upper and lower segment boundaries or threshold levels for the edge-detection. The video-based estimation of human body segment parameters is especially useful in situations where ease of application and rapid availability of comparatively precise parameter values are of importance.
已开发出一种方法，可从四种不同身体构型的视频图像中精确测定人体测量尺寸。通过采用具有亚像素精度的物体边界定位技术、实施校准算法以及考虑身体各部分与记录相机的不同距离，实现了高精度。如果用黑色丝带在受试者身上标记边界，该系统可从视频图像中自动识别节段边界。结合哈策的数学有限质量元节段模型，可将通过视频测量确定的人体测量数据作为输入数据，计算身体节段参数（体积、质量、三个主惯性矩、节段质心的三个局部坐标等）。与最近发表的其他用于估计身体节段惯性特性的基于视频的系统相比，本算法减少了因光学畸变、不准确的边缘检测程序以及用户指定的节段上下边界或边缘检测阈值水平而产生的误差。基于视频的人体节段参数估计在应用简便且能快速获得相对精确参数值的情况下尤其有用。
