An Adaptive Approach for Segmentation of 3d Laser Point Cloud

Lari, Zahra; Habib, Ayman; Kwak, Eunju

doi:10.5194/isprsarchives-xxxviii-5-w12-103-2011

Cited by 34 publications

(27 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead of using 2D-image segmentation algorithms, such as Mean Shift Segmentation, we identified planar surfaces from the given point cloud (e.g., either the derived sparse point cloud from SfM approach or derived point cloud from initial semi-global dense matching) through a 3D point cloud segmentation. In this paper, the adopted point cloud segmentation is developed by (Lari et al, 2011). Since 3D point clouds include more geometric information regarding the reconstructed objects, segments derived from point cloud segmentation are more accurate.…”

Section: Refinement Using Point Cloud Segmentsmentioning

confidence: 99%

Planar Constraints for an Improved Uav-Image-Based Dense Point Cloud Generation

He¹,

Habib²,

Al‐Rawabdeh³

2015

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

Self Cite

View full text Add to dashboard Cite

ABSTRACT:In this paper, we proposed a new refinement procedure for the semi-global dense image matching. In order to remove outliers and improve the disparity image derived from the semi-global algorithm, both the local smoothness constraint and point cloud segments are utilized. Compared with current refinement technique, which usually assumes the correspondences between planar surfaces and 2D image segments, our proposed approach can effectively deal with object with both planar and curved surfaces. Meanwhile, since 3D point clouds contain more precise geometric information regarding to the reconstructed objects, the planar surfaces identified in our approach can be more accurate. In order to illustrate the feasibility of our approach, several experimental tests are conducted on both Middlebury test and real UAV-image datasets. The results demonstrate that our approach has a good performance on improving the quality of the derived dense image-based point cloud.

show abstract

Section: Refinement Using Point Cloud Segmentsmentioning

confidence: 99%

Planar Constraints for an Improved Uav-Image-Based Dense Point Cloud Generation

He¹,

Habib²,

Al‐Rawabdeh³

2015

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…If BIM is modelled on the basis of previously captured building information, the previous data capture, processing and recognition methods influence the data quality through the deployed technique and the provided LOD related to interoperability issue. (Fai & Rafeiro, 2014;Volk et al 2014 (Lari, 2011. Tang et al, 2010.…”

Section: Shape Knowledge For Geometric Modellingmentioning

confidence: 99%

Historical Buildings Models and Their Handling via 3d Survey: From Points Clouds to User-Oriented Hbim

Chiabrando¹,

Sammartano²,

Spano³

2016

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:This paper retraces some research activities and application of 3D survey techniques and Building Information Modelling (BIM) in the environment of Cultural Heritage. It describes the diffusion of as-built BIM approach in the last years in Heritage Assets management, the socalled Built Heritage Information Modelling/Management (BHIMM or HBIM), that is nowadays an important and sustainable perspective in documentation and administration of historic buildings and structures. The work focuses the documentation derived from 3D survey techniques that can be understood like a significant and unavoidable knowledge base for the BIM conception and modelling, in the perspective of a coherent and complete management and valorisation of CH. It deepens potentialities, offered by 3D integrated survey techniques, to acquire productively and quite easily many 3D information, not only geometrical but also radiometric attributes, helping the recognition, interpretation and characterization of state of conservation and degradation of architectural elements. From these data, they provide more and more high descriptive models corresponding to the geometrical complexity of buildings or aggregates in the well-known 5D (3D + time and cost dimensions).Points clouds derived from 3D survey acquisition (aerial and terrestrial photogrammetry, LiDAR and their integration) are reality-based models that can be use in a semi-automatic way to manage, interpret, and moderately simplify geometrical shapes of historical buildings that are examples, as is well known, of non-regular and complex geometry, instead of modern constructions with simple and regular ones. In the paper, some of these issues are addressed and analyzed through some experiences regarding the creation and the managing of HBIM projects on historical heritage at different scales, using different platforms and various workflow. The paper focuses on LiDAR data handling with the aim to manage and extract geometrical information; on development and optimization of semi-automatic process of segmentation, recognition and modelling of historical shapes of complex structures; on communication of historical heritage by virtual and augmented reality (VR/AR) in a 3D reconstruction of buildings aggregates from a LiDAR and UAV survey. The HBIM model have been implemented and optimized to be managed and browse by mobile devices for not only touristic or informative scopes, but also to ensure that HBIM platforms will become more easy and valuable tools helping all professionals of AEC involved in the documentation and valorisation process, that nowadays more and more distinguish CH policies.

show abstract

“…Kim and Habib (2007) utilize cylinder neighbourhood concept for the segmentation of planar patches using parametric-domain methods successfully. Filin and Pfeifer (2006), Lari et al (2012) and Lari and Habib (2013) improve their planar surface segmentation results by considering the noise level and the physical shape of the associated surface. However, all mentioned papers take advantage of the parametric-domain methods that are computationally not efficient.…”

Section: Related Workmentioning

confidence: 99%

“…Many researchers segment planar surfaces using parametric-domain methods e.g. Lari et al (2012), Filin and Pfeifer (2005), Vosselman and Dijkman (2001), Kim and Habib (2007) and Maas and Vosselman (1999) who segment planes in urban environments. Roggero (2002) segments an urban environment by employing PCA in attribute space.…”

Section: Related Workmentioning

confidence: 99%

Segmentation of planar surfaces in LiDAR point clouds of an electrical substation by exploring the structure of points neighbourhood

Arastounia

Lichti

2014

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:According to the Department of Energy of the USA, today's electrical distribution system is 97.97% reliable. However, power outages and interruptions still impact many people. Many power outages are caused by animals coming into contact with the conductive elements of the electrical substations. This can be prevented by covering the conductive electrical objects with insulating materials. The design of these custom-built insulating covers requires a 3D as-built plan of the substation. This research aims to develop automated methods to create such a 3D as-built plan using terrestrial LiDAR data for which objects first need to be recognized in the LiDAR point clouds. This paper reports on the application of a new algorithm for the segmentation of planar surfaces found at electrical substations. The proposed approach is a region growing method that aggregates points based on their proximity to each other and their neighbourhood dispersion direction. PCA (principal components analysis) is also employed to segment planar surfaces in the electrical substation. In this research two different laser scanners, Leica HDS 6100 and Faro Focus 3D , were utilized to scan an electrical substation in Airdrie, a city located in north of Calgary, Canada. In this research, three subsets incorporating one subset of Leica dataset with approximately 1.7 million points and two subsets of the Faro dataset with 587 and 79 thousand points were utilized. The performance of our proposed method is compared with the performance of PCA by performing check point analysis and investigation of computational speed. Both methods managed to detect a great proportion of planar points (about 70%). However, the proposed method slightly outperformed PCA. 95% of the points that were segmented by both methods as planar points did actually lie on a planar surface. This exhibits the high ability of both methods to identify planar points. The results also indicate that the computational speed of our method is superior to that of PCA by 50%. It is concluded that our proposed method achieves better results with higher computational speed than PCA in the segmentation of planar surfaces.

show abstract

An Adaptive Approach for Segmentation of 3d Laser Point Cloud

Cited by 34 publications

References 6 publications

Planar Constraints for an Improved Uav-Image-Based Dense Point Cloud Generation

Planar Constraints for an Improved Uav-Image-Based Dense Point Cloud Generation

Historical Buildings Models and Their Handling via 3d Survey: From Points Clouds to User-Oriented Hbim

Segmentation of planar surfaces in LiDAR point clouds of an electrical substation by exploring the structure of points neighbourhood

Contact Info

Product

Resources

About