Automatic Detection of Building Points From Lidar and Dense Image Matching Point Clouds

Abstract

ABSTRACT:This study aims to detect automatically building points: (a) from LIDAR point cloud using simple techniques of filtering that enhance the geometric properties of each point, and (b) from a point cloud which is extracted applying dense image matching at high resolution colour-infrared (CIR) digital aerial imagery using the stereo method semi-global matching (SGM). At first step, the removal of the vegetation is carried out. At the LIDAR point cloud, two different methods are implemented and evaluated using initially the normals and the roughness values afterwards: (1) the proposed scan line smooth filtering and a thresholding process, and (2) a bilateral filtering and a thresholding process. For the case of the CIR point cloud, a variation of the normalized differential vegetation index (NDVI) is computed for the same purpose. Afterwards, the bare-earth is extracted using a morphological operator and removed from the rest scene so as to maintain the buildings points. The results of the extracted buildings applying each approach at an urban area in northern Greece are evaluated using an existing orthoimage as reference; also, the results are compared with the corresponding classified buildings extracted from two commercial software. Finally, in order to verify the utility and functionality of the extracted buildings points that achieved the best accuracy, the 3D models in terms of Level of Detail 1 (LoD 1) and a 3D building change detection process are indicatively performed on a sub-region of the overall scene.

Method for Creating Digital Orthophotomaps of Increased Informational Value

“…[Sauerbier 2004]. Processing of dense three-dimensional clouds of points can facilitate the extraction of structural peculiarities of objects, the development of relief models, and the performance of other procedures of geographic information modelling, including the design of orthophotomaps [Maltezos, Ioannidis 2015;Karantzalosa et al 2015;Cho, Snavely 2013].…”

“…Approaches to the detection of vegetation structures and buildings using point classification are well developed for the processing of the data, which is obtained from laser scanning. They mainly apply the analysis of local geometrical properties (that is, the local geometry features analysis), such as height, local level of a point's height over the neighbouring points, intensity of reflection, and intensity of the registered brightness (return intensity, and image intensity) [Maltezos et al 2015]. Some authors focus their attention on the methods of complex application of LIDAR data, and multi-spectral photography as well as aerial surveying of high geometric differentiation in RGB format with spectral data of distant probing.…”

Consistent Roof Geometry Encoding for 3D Building Model Retrieval Using Airborne LiDAR Point Clouds

“…Therefore, sharp edges are used as geometric features to represent roof geometries. Inspired by the study [15], Laplacian of Gaussian (LoG) filter is adopted to extract sharp edges while alleviating disturbances from noise. The LoG is composed of Gaussian and Laplacian filters, where the low-pass kernel function in the Gaussian filter is used to suppress noise and the second derivative kernel function in Laplacian filter is utilized to extract sharp edges.…”

Knowledge Based 3d Building Model Recognition Using Convolutional Neural Networks From Lidar and Aerial Imageries

“…In the third group of modelling approaches, the combination of model driven and data driven algorithms is used to have an optimal method for compensating the weakness of each methods. According to our study, current methodologies and algorithms on building detection and extraction problem can be divided into four groups as; plan fitting based methods (Mongus, et al, 2014); filtering and thresholding based methods (Maltezos, et al, 2015;Hermosilla, et al, 2011) such as morphological methods (Yu, et al, 2010); segmentation based methods such as binary space partitioning (Wichmann, et al, 2015), shadow based segmentation (Singh, et al, 2015;Ngo, et al, 2015), and region growing based algorithms (Matikainen, et al, 2010;Awrangjeb, et al, 2013); and finally the latest group, different supervised classification methods (Hermosilla, et al, 2011;Guo, et al, 2011;Karantzalos, et al, 2015;Vakalopoulou, et. al, 2011).…”