Font Size: a A A

Automatic Building Extraction From The Terrestrial Laser Scanning Data Based On Spherical Projection And Fuzzy C-Mean Clustering Analysis

Posted on:2013-01-22Degree:MasterType:Thesis
Country:ChinaCandidate:L Y WangFull Text:PDF
GTID:2212330374461418Subject:Geodesy and Survey Engineering
Abstract/Summary:PDF Full Text Request
Laser scanners, with its higher measurement accuracy and measurement efficiency, are widely used in surveying. How extract the characteristic of the object from terrestrial laser scanning point cloud data is becoming a popular topic in today's scan data processing and also is the key to the subsequent application of the scan data. This paper take full use of the existing algorithms and the theoretical of research results, with the spherical projection structure network and the FCM clustering method processing the point clouds by choosing test data.The discussion and research in this paper as follows:1. Cylindrical projection and the design of CYLVIEW browserFor a good browsing about the original point clouds and classified data, we use a cylindrical projection method to browse our data, and the noise can be observed, which provides convenience for the next step.2. Fuzzy C-Mean(FCM), the fuzzy clustering method for point clouds classificationFuzzy C-the Mean algorithm method is mainly used for remote sensing image classification interpretation. The method estimate the center of the cluster by the introduction of the membership function that belongs to a different analogy and divide the data, and repeating adjustment the cluster center to make the distance of the data for each dot minimum or meet the termination condition.3. The method of the triangular strutted based on the spherical projectionAccording to the coordinate's correction, the sphere that was regard as a vector space scattered point cloud was projected onto the plane, and get the corresponding two-dimensional coordinates in the plane, then we use the Delaunay algorithm construct plane triangulation network, finally, the flat triangular mesh is mapped to the space to complete the three-dimensional modeling.4. Software developmentThe paper used IDL algorithms, which dissolved into LasaStudio system, therefor the algorithms can build the network of single-station ground-based laser scanning data, and automatic extract building data.
Keywords/Search Tags:3D laser scanning, 3D modeling, Automatic classification of extraction, Spherical projection, FCM clustering, Cylindrical browse
PDF Full Text Request
Related items