基于图像的人体三维重建-控制科学与工程专业论文.docxVIP

Classified Index: TP391.41 U.D.C.: 621 Dissertation for the Master Degree in Engineering IMAGE BASED METHOD FOR 3D HUMAN BODY RECONSTRUCTION Candidate： Xihong Hu Supervisor： Prof. Zexiang Li Co-Supervisor: Prof. Xiaojun Wu Academic Degree Applied for： Master of Engineering Speciality： Control Science and Engineering Affiliation： Shenzhen Graduate School Date of Defence： December, 2008 Degree-Conferring-Institution： Harbin Institute of Technology 哈尔滨工业大学工学硕士学位论文 哈尔滨工业大学工学硕士学位论文 I I 摘 要 物体几何模型可以通过用坐标测量器和三维扫描仪获得的点云数据来获 得。这种通过用点云数据来获取模型的方法的精度通常是比较高的，然而， 它的速度往往比较慢。在服装行业中，有时一种快速，低耗的人体几何模型 的获取是必要的。 本文提出了一种基于两幅正交人体图像的可应用于服装行业的人体几何 模型的获得方法。 首先，我们采用了几种分割方法对正侧面的两幅正交视图进行分割，分 别得到它们的边缘轮廓，并且轮廓上的特征点也能得到。然后分析了这几种 方法得出结果的差异，选出了一种满足本文需求的方法。其次，对三维几何 模型进行切片，得到模型的切片轮廓，并且对图像轮廓构造不同的高度层， 根据图像的高度层数构造模型的切片厚度，最后根据图像中的数据点运用一 种简单有效的变形方法对模型的切片轮廓进行变形得到新的轮廓数据，通过 三维重建这些新的数据点即获得新的三维模型。这种方法简单、经济，并且 安全，结果显示，得出的新三维模型与被测对象具有比较高的相似性，也即 验证了此方法的可行性。 关键词 边缘轮廓；模型切片；变形方法；三维重建 II II Abstract The geometric model of an object is generated from a cloud of points, which can be obtained by coordinate measuring machines (CMM) and 3D laser scanners. The accuracy of the models generated from cloudy points is always high. However, their speed is always slow. In the garment industry, sometimes, a rapid and low-cost 3D virtual human reconstruction approach is required. The research presented in this thesis is to develop a technique of virtual human modeling for the garment industry from two photographs of a human body in two orthogonal views. Firstly, in our approach, an efficient image segmentation method is applied to two photographs (front-view and side-view) to obtain the contours of the human body separately. Then, the features on the human contours are extracted. Secondly, construct different height contours using image segmentation data having been obtained to get the slices of the images. Then the model is stratified to get slices of the model according to the number of the slices o