3D Reconstruction PDF Print E-mail
Written by Nicholas Shorter   
Monday, 04 August 2008 05:07

Light Detection and Ranging (LiDAR) systems sample a given terrain into a finite collection of irregularly spaced data points. A triangulation algorithm can then adaptively connect the data points to represent the depicted terrain with a series of interconnected, non overlapping triangles or a triangulated irregular network (TIN). The TIN can then be analyzed and coplanar triangles can be clustered into groups corresponding to the depicted buildings’ roof planes. Then extracting these roof planes and merging them can lead to reconstructing the sampled, original, irregularly spaced, data points into three dimensional complete models. This is all to be done autonomously, that is, with no user intervention.

Several military and commercial applications exist for 3D reconstruction from LiDAR data. For the military, the analysis of LiDAR data can be used for target recognition applications. Work in training volume correlation filters (VCFs) to recognize tanks and other military vehicles within LiDAR data has recently been developed. There is investigation underway for mounting LiDAR sensors on unmanned aerial vehicles (UAVs). This would enable aerial surveying of terrains in which military forces were denied access too. Scenes surveyed by an UAV or high flying plane with a LiDAR sensor could then be reconstructed into 3D models. The analysis of 3D models of given terrains has a variety of commercial applications: urban planning; network planning for mobile communication, spatial analysis of air pollution and noise nuisances, geographical information systems, and security services.

Preprocessing techniques are proposed for the development of a 3D Reconstruction algorithm designed for autonomously reconstructing three dimensional models from urban and residential buildings depicted in raw LIDAR data. First, a greedy insertion triangulation algorithm, modified with a novel noise filtering technique, triangulates the raw LIDAR data. Second, the normal vectors of the triangulated raw LIDAR data are then passed to an unsupervised clustering algorithm – Fuzzy Simplified Adaptive Resonance Theory (Fuzzy SART). Fuzzy SART returns a rough grouping of the coplanar triangles. Then, a proposed multiple regression algorithm further refines the coplanar grouping by further removing outliers and deriving an improved planar segmentation of the raw LIDAR data. Finally, further refinement is achieved by calculating the intersection of the best fit roof planes and moving nearby points close to that intersection to exist at the intersection, resulting in straight, straight roof ridges.

The above described paradigm was tested on the following three buildings:

 

Aerial Image Buildings

 

The reconstruction algorithm  produced the following results:

 

Reconstructed Buildings

 


Fore more on this research effort please refer to the following conference paper and presentation:

Shorter, N.S.; Kasparis, T. (2008). Clustering Irregular Spaced LiDAR TINs for 3D Reconstruction. The 5th International Conference On Cybernetics and Information Technologies, Systems and Applications, Orlando, Florida, June 29 - July 2 (published)

Conference Paper

Shorter, N. S.; Kasparis, T. (2006). 3D Reconstruction of Irregular Spaced LIDAR, Proceedings of the 6th WSEAS International Conference on Systems Theory and Scientific Computation, Elounda, Greece, August 21-23 (pp. 19-24).

Journal Paper

Shorter, N. S.; Kasparis, T. (2006). Fuzzy SART Clustering for 3D Reconstruction from Irregular LIDAR Data, WSEAS Transactions on Signal Processing, Vol. 2, No. 8 (pp. 1122 to 1129).

Conference Paper

 

Last Updated on Thursday, 28 August 2008 22:27