Aiming at the problem of serious distortion in the body deformation of 3-D garment simulation, an algorithm of left-right section structure of human body was provided. The body was divided into front and rear parts and deformed, respectively. First of all, the processed point cloud body was used to identify the feature of the body, and the points on the shoulder line and side seam were selected and designated as the feature points. The 3-D coordinates of these points were marked and obtained. Secondly, the space quadric equation was introduced and constructed as left-right section model by analyzing coordinate data from these feature points. Thirdly, based on the surface reconstruction of reverse engineering, the 3-D coordinates of specified feature points were substituted into the quadric equation,the system of non-homogeneous linear equations was constituted, and the parameters of the system were solved and the quadric equation coefficients were obtained. In the end, the quadric surface equation constituting the left-right section was applied to the simulation of human body, and a left-right section containing the feature points was built. The algorithm accurately constructs the left and right section of the body which divides the body into two parts. It is shown a good effect when the section is applied to the body deformation of the garment simulation.%针对三维服装仿真人体变形严重失真的情况,提出一种构造人体左右剖面的算法,将人体分成前后2个部分分别进行变形.首先,对处理过的扫描人体进行特征识别,选取肩线和体侧线上的点,指定为特征点,标识并获取其三维坐标信息;其次,对这些特征点的坐标数据进行分析,构造空间二次曲面方程作为左右剖面模型;然后,利用逆向工程的曲面重构,将获取的指定特征点的三维坐标数据代入二次曲面方程,构成非齐次线性方程组,通过反求方程组的参数,获得二次曲面方程的系数;最后,将构成左右剖面的二次曲面方程应用到仿真人体上,形成包含特征点的冠状面.研究结果表明,该方法精准地构造了人体冠状面,将仿真人体分成前后2个部分,并应用到服装仿真人体的变形,效果良好.
展开▼
