摘要
由于Kinect设备具有快速安全、经济便携的特点,提出了利用Kinect实现扫描、重建以及测量的人体成分分析系统框架.通过非接触式光学技术捕捉人体表面几何特征数据,利用KinFu实时进行三维数据的点云配准与模型的表面重建,并采用类积分法与蒙特卡罗法分别计算模型相关部位的周长与体积.从扫描到模型重建过程耗时约3min,模型光滑精细,腰围等部位的尺寸测量精确度为97.7%,体积测量精确度为96.8%.试验结果证明了系统的效用性.
A body composition analysis system framework is presented to perform scanning,reconstruction and the measurement by using the Kinect device with the advantages of fast,safety,low-cost and portable.The body surface geometry data are captured through the non-contact optical technology,the model surface is reconstructed in real-time with the registration for the 3D point cloud data by KinFu,and then the circumference and volume of the relevant parts of the model are computed by using quasi-integral method and Monte Carlo method respectively.Scanning and 3Dreconstruction cost about 3min all together with a smooth and fine result,the accuracy of the dimension measurement is 97.7%,and volume measurement accuracy is 96.8%.The effectiveness of the proposal has been exhibited by the experimental results.
出处
《东华大学学报（自然科学版）》
CAS
CSCD
北大核心
2015年第1期78-83,共6页
Journal of Donghua University(Natural Science)
基金
国家自然科学基金资助项目(81372057)
关键词
人体成分分析
Kinect扫描
三维重建
周长测量
体积测量
body composition analysis
Kinect scanning
3D reconstruction
circumference measurement
volume measurement
