节点文献

基于散乱数据点的物体表面重建方法的研究

Research on Object Surface Reconstruction Based on Scattered Points

【作者】 王付刚;

【导师】 范剑英;

【作者基本信息】 哈尔滨理工大学 , 光学工程, 2007, 硕士

【摘要】 随着激光测距扫描等三维数据获取硬件技术的日趋完善,利用物体表面三维数据来建立真实的物体数字模型,成为近年来国际图形学界的一种发展趋势。基于散乱数据点的曲面重建作为一个重要研究课题也得到了广泛的探讨和研究,成为国际上的研究热点之一,在逆向工程、可视化等领域有着广泛的应用。本文采用的数据是由多线结构光三维测量系统获得的,数据点集只包含点的三维坐标信息。由于遮挡特性及物体表面反射率等原因,导致数据缺损。并且由于该系统获取的数据量很大,致使曲面重构速度缓慢。文中着重研究了一种分区域插值的方法,实现了条纹上插值和条纹间插值,弥补了数据,提高了插值精度和插值速度;采用了一种由二维曲线Splitting算法拓展到三维空间的自适应门限3D Splitting算法对散乱数据进行压缩,取得了较高的压缩比,又保持了物体的细节信息。本文重点研究了多线结构光获取的物体表面散乱数据点集空间直接三角剖分算法,用三角面片来逼近物体的表面。根据散乱数据的特点,采用了一种相邻轮廓线同步前进的空间直接剖分算法,即选取两条相邻轮廓线上具有最大z坐标值的一对点连接构成起始边。按照本文采用的点搜索方法寻找第三个点并构造出三角形,将该点与另一轮廓上的第一个点连接组成下一三角剖分的初始边,直到所有轮廓上的点用完为止,则相邻两条轮廓线之间剖分完毕。当所有轮廓线都剖分完毕后则完成了三角剖分曲面重建的全过程。本文还对剖分后的三角形进行了优化,优化后的三角形更加逼近物体表面。实验验证表明:分区域插值方法弥补了缺失的数据;3D Splitting数据压缩算法取得了较高的压缩比。相邻轮廓同步前进三角剖分算法可以快速地重建物体表面,优化后的结果更加接近于原始物体表面。

【Abstract】 With the improvement of hardware technique of 3D dataset capturing, such as laser-ranging-scanning technique and so on, establishing the real object’s digital model with 3D data on the surface has become the trend of development in international graph field in recent years. As an important research subject, surface reconstruction based on scattered points is also widely discussed and studied, which has been one of the most popular research and has been widely applied in the field such as Reverse Engineering and Visualization.All the data in this article is obtained from the measuring system of the multi-line structured light, and the data sets only contains the three dimensional coordinate information. Occlusion and reflectivity of object’s surface lead to data defective. And excessive data obtained makes the speed of surface reconstruction lower. One kind of sub-region interpolation method mainly studied in this paper is to realize interpolation on and between stripes, remedy data and enhance the precision and speed of interpolation. One sort of adaptive threshold 3D Splitting method from splitting method of 2D to 3D, which can compress scattered data with high compression ratio and keep the detail information of object.The spatial triangulation method of scattered data points on the surface of object obtained by multi-line structured light is the key point of this paper, which approaches surface of object with triangle slice. According the features of scattered data, a kind of method of synchronous march between adjacent contours is adopted, namely forming initial boundary by choosing a pair of dots with maximal z coordinate on two adjacent contour lines connected. Data research method adopted in this article is to search the third dot and make up a triangle, then to constitute the initial side of next triangle by connecting this dot and another dot on the other contour, with all the dots on the contours are connected, the triangulation is finished. Until all contour lines are used, the whole process of surface reconstruction of triangulation is completed. Optimization of triangulation is also introduced in this paper, to make it closer to the surface of object.The result of experiment shows that the method of sub-region interpolation can remedy lacked data, 3D Splitting compression arithmetic can get higher compression ratio. Synchronous march between contiguous contours is able to fast reconstruct the surface of object, and the optimized result closer to the original object.

  • 【分类号】TP391.41
  • 【被引频次】7
  • 【下载频次】329
节点文献中: