Image-based Airborne LiDAR Point Cloud Encoding for 3D Building Model Retrieval

  • 陳 俋臣

Student thesis: Master's Thesis

Abstract

With the development of Web 2 0 and cyber city modeling an increasing number of 3D models have been made available on web-based model-sharing platforms with many applications such as navigation urban planning and virtual reality A 3D model retrieval system based on the concept of data reuse is proposed to retrieve building models similar to a user-specified query The basic idea is to reuse existing 3D building models instead of reconstructing models from point clouds Models in databases are generally encoded compactly with a shape descriptor to retrieve models efficiently However most of the geometric descriptors in related work are applicable to polygonal models only In this study the input query of the model retrieval system is a point cloud acquired by light detection and ranging (LiDAR) systems because of the efficient scene scanning and spatial information collection Using point clouds with sparse noisy and incomplete sampling as input queries is more difficult than using 3D models Given that the roof of a building is more informative than the other parts in an airborne LiDAR point cloud an image-based approach is proposed to encode both point clouds from input queries and 3D models in databases The main goal of data encoding is to encode the models in the database and input point clouds consistently First top-view depth images of buildings are generated to represent the geometric surface of a building roof Second geometric features are extracted from the depth images according to the height edge and plane of the building Finally descriptors are extracted by establishing spatial histograms and are used in the 3D model retrieval system For data retrieval the models are retrieved by matching the encoding coefficients of point clouds and building models Experiments on a database that includes about 1 000 000 3D models collected from the Internet are conducted to evaluate data retrieval Results show that the proposed method is superior to related methods
Date of Award2016 Sep 8
Original languageEnglish
SupervisorChao-Hung Lin (Supervisor)

Cite this

Image-based Airborne LiDAR Point Cloud Encoding for 3D Building Model Retrieval
俋臣, 陳. (Author). 2016 Sep 8

Student thesis: Master's Thesis