Raw Light Detection and Ranging Data exists as a point cloud  a collection of irregular spaced 3D points. One can then triangulate those points formulating a triangulated irregular network or TIN where a single triangle uniquely defines space existent in the point cloud. Several advantages exist for triangulating the irregular LiDAR data: (i) With each triangle uniquely defining a given space within the point cloud, it is then possible to execute spatial relational queries on the TIN. (ii) It is possible to interpolate the raw LiDAR data to exist at regular intervals using the LiDAR TIN (iii) Furthermore, it is possible to derive several features for triangles existing in LiDAR TIN: (a) Raw LiDAR points serving as vertexes for a given triangle in the LiDAR TIN (b) The planar coefficients of a triangle in the LiDAR TIN (c) The triangles adjacent to a given triangle in the LiDAR TIN (d) The normal vector of a given triangle in the LiDAR TIN (e) The angle between normal vectors between adjacent triangles in the LiDAR TIN In 3D Reconstruction of Irregular Spaced LIDAR, Shorter and Kasparis propose a filtering scheme for removing noise from the LiDAR data. With the noise removed, this then enables one to use Fuzzy SART to cluster the normal vectors of the LiDAR TIN to automatically detect roof top planar structures within the LiDAR data. Fore more on this research effort please refer to the following conference paper and journal publication: 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 2123 (pp. 1924).  Conference 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).  Journal Paper 
