本文關鍵詞： 三維激光掃描技術 點云數據 點云去噪 點云簡化 曲面重建　出處：《東華理工大學》2013年碩士論文　論文類型：學位論文

【摘要】：三維激光掃描技術是二十世紀九十年代新興的一門測量技術，它采用非接觸式高速激光測量，以獲取研究目標的三維坐標和數碼照片的方式，快速高效的得到目標的三維立體信息。因此該技術被廣泛應用于逆向工程、醫(yī)學、考古、建筑業(yè)、船舶、電力、市政建設等眾多領域。但是三維激光掃描獲取的數據存在各種誤差和噪聲，而且一般得到的原始掃描數據密度較大，在利用這些數據進行后續(xù)的處理、曲面重建時，為了保證精度和速度，需要對原始數據進行預處理。另一方面，進行曲面重建時，涉及幾何計算、拓撲關系建立以及三維模型的繪制問題。因此對三維激光掃描數據處理與曲面重建方法進行研究具有重大理論意義與現實意義。針對以上問題，本文從以下三個方面進行了深入研究。 1.點云去噪。在分析已有去噪算法的基礎上，針對散亂點云，，給出了基于最小二乘法的分步去噪方法。通過對噪聲點進行分類，先利用求取包圍盒最大連通域的方法去除離群點，再用最小二乘法擬合出K鄰域內點的最佳逼近平面，通過判定鄰域各點到該平面的距離與設定的閾值的大小來去除振幅較小噪聲，達到了預定效果。 2.點云簡化。針對基于曲率簡化的方法能較好的保留重建曲面的細節(jié)特征，但由于需要計算每個點的曲率并進行比較，效率較低，本文給出了基于輪廓點提取的簡化方法，通過先提取輪廓點作為簡化點云需要保留的特征點，然后再以輪廓點中曲率最小值為分界，把點云數據按曲率大小分為兩個部分。對于大于等于最小值的部分，進行輪廓點保留，其他點刪除處理，對于小于最小值的部分，根據曲率精簡原則進行簡化處理。該方法既繼承了基于曲率簡化的方法能較好的保留重建曲面的細節(jié)特征的優(yōu)點，又在一定程度上減少了計算量，提高了效率。 3.曲面重建。介紹了常用的曲面重建方法：參數曲面重建、隱式曲面重建、分片線性曲面重建、細分曲面重建與變形曲面重建，并分析了各自優(yōu)缺點。詳細介紹了利用Geomagic Studio進行曲面重建的過程，最后用該軟件對三個點云數據進行了曲面重建。
[Abstract]:In 1990s, 3D laser scanning technology is a new measurement technology. It uses non-contact high-speed laser measurement to obtain 3D coordinates and digital photos of the research object. Therefore, this technology is widely used in reverse engineering, medicine, archaeology, construction, ship, electricity. Municipal construction and many other fields. But 3D laser scanning data obtained by a variety of errors and noise, and generally the original scanning data density is high, in the use of these data for subsequent processing. In order to ensure the accuracy and speed of surface reconstruction, it is necessary to preprocess the original data. On the other hand, geometric calculation is involved in surface reconstruction. Therefore, it is of great theoretical and practical significance to study the methods of 3D laser scanning data processing and surface reconstruction. This article has carried on the thorough research from the following three aspects. 1. Point cloud denoising. Based on the analysis of existing de-noising algorithms, a step de-noising method based on least square method is presented for scattered point clouds. The noise points are classified. The outlier is removed by the method of finding the largest connected domain of the bounding box, and then the best approximation plane of the point in the K-neighborhood is fitted by the least square method. By determining the distance from each point to the plane and the value of the threshold value, the amplitude of the noise is reduced, and the predetermined effect is achieved. 2. Point cloud simplification. The method based on curvature simplification can better preserve the detailed features of surface reconstruction, but because of the need to calculate and compare the curvature of each point, the efficiency is low. In this paper, a simplified method based on contour point extraction is presented. Firstly, the contour point is extracted as the feature point to be retained by the simplified point cloud, and then the minimum curvature in the contour point is taken as the boundary. The point cloud data is divided into two parts according to the curvature. For the parts greater than or equal to the minimum, the contour points are reserved, the other points are deleted, and the parts less than the minimum value are processed. According to the principle of curvature reduction, the method not only inherits the advantages of curvature simplification method, but also reduces the computational complexity to a certain extent. Improved efficiency. 3. Surface reconstruction: parametric surface reconstruction, implicit surface reconstruction, piecewise linear surface reconstruction, subdivision surface reconstruction and deformation surface reconstruction. The process of surface reconstruction using Geomagic Studio is introduced in detail. Finally, three point cloud data are reconstructed with the software.
【學位授予單位】：東華理工大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：P225.2

【參考文獻】

相關期刊論文 前10條

1 宋宏;;地面三維激光掃描測量技術及其應用分析[J];測繪技術裝備;2008年02期

2 劉大峰;廖文和;戴寧;程筱勝;;散亂點云去噪算法的研究與實現[J];東南大學學報(自然科學版);2007年06期

3 錢勇先,高智勇,林家瑞;MRI圖像重建網格化算法的研究進展[J];國外醫(yī)學.生物醫(yī)學工程分冊;2001年04期

4 吳湘,趙萬生,魏莉;三維幾何表示法[J];航天制造技術;2002年04期

5 周克勤;趙煦;丁延輝;;基于激光點云的3維可視化方法[J];測繪科學技術學報;2006年01期

6 張舜德,朱東波,盧秉恒;反求工程中三維幾何形狀測量及數據預處理[J];機電工程技術;2001年01期

7 倪小軍;姜曉峰;葛亮;;特征保留的點云數據自適應精簡算法[J];計算機應用與軟件;2011年08期

8 馬曉泉;;地面三維激光掃描技術及其在國內的應用現狀[J];科技信息;2012年29期

9 羅先波,鐘約先,李仁舉;三維掃描系統(tǒng)中的數據配準技術[J];清華大學學報(自然科學版);2004年08期

10 向日華,王潤生;一種基于高斯混合模型的距離圖像分割算法[J];軟件學報;2003年07期

本文編號：1447427