Co-registration of photogrammetric and LiDAR data in urban environments.

摘要

To date, most building data extraction and reconstruction for mapping and 3D modeling is done using either LiDAR data or optical data. The complementary characteristics of photogrammetric data and LiDAR data can be utilized to generate new algorithms and improve the accuracy of current methods for processing building data in a wide range of applications, such as city modeling, change detection, and object recognition. However, the integration of data from both LiDAR and photogrammetric domains is hard to achieve directly because of their heterogeneity (e.g., different data types, times of data collection, reference systems). Therefore, the development of automatic and robust data alignment techniques is necessary for the integration of both LiDAR and photogrammetric data.;In this thesis, the co-registration steps between LiDAR and photogrammetric DSM data are analyzed and solutions are proposed and implemented. For a robust 3D geometric transformation both planes and points are used. Initially planes are chosen as the co-registration primitives. A region growing algorithm based on a Triangulated Irregular Network (TIN) is implemented to extract planes from both datasets. Point clouds have also been used as another registration primitive to complement the plane-based registration. Next, an automatic and iterative process for identifying and matching corresponding planes from the two datasets has been developed and implemented. The extracted planes are associated as plane pairs, initially by a building matching process which is then followed by the plane matching algorithm. Then three different geometric registration algorithms are used to obtain accurate transformation parameters between the two datasets. The 3D conformal transformation method and the attitude quaternion are the two methods applied to obtain the transformation parameters using the corresponding plane pairs. Following the mapping of one dataset onto the coordinate system of the other, the Iterative Closest Point (ICP) algorithm is then applied, using the corresponding building point clouds to further refine the transformation solution. Experimental results together with their assessments are presented and discussed to demonstrate the applicability of the proposed approach.