Robust locally weighted regression for ground surface extraction in mobile laser scanning 3D data
A new robust way for ground surface extraction from mobile laser scanning 3D point cloud data is proposed in this paper. Fitting polynomials along 2D/3D points is one of the well-known methods for filtering ground points, but it is evident that unorganized point clouds consist of multiple complex st...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Published: |
ISPRS
2013
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/20.500.11937/26232 |
| _version_ | 1848751927432577024 |
|---|---|
| author | Nurunnabi, Abdul West, Geoff Belton, David |
| author2 | M. Scaioni |
| author_facet | M. Scaioni Nurunnabi, Abdul West, Geoff Belton, David |
| author_sort | Nurunnabi, Abdul |
| building | Curtin Institutional Repository |
| collection | Online Access |
| description | A new robust way for ground surface extraction from mobile laser scanning 3D point cloud data is proposed in this paper. Fitting polynomials along 2D/3D points is one of the well-known methods for filtering ground points, but it is evident that unorganized point clouds consist of multiple complex structures by nature so it is not suitable for fitting a parametric global model. The aim of this research is to develop and implement an algorithm to classify ground and non-ground points based on statistically robust locally weighted regression which fits a regression surface (line in 2D) by fitting without any predefined global functional relation among the variables of interest. Afterwards, the z (elevation)-values are robustly down weighted based on the residuals for the fitted points. The new set of down weighted z-values along with x (or y) values are used to get a new fit of the (lower) surface (line). The process of fitting and down-weighting continues until the difference between two consecutive fits is insignificant. Then the final fit represents the ground level of the given point cloud and the ground surface points can be extracted. The performance of the new method has been demonstrated through vehicle based mobile laser scanning 3D point cloud data from urban areas which include different problematic objects such as short walls, large buildings, electric poles, sign posts and cars. The method has potential in areas like building/construction footprint determination, 3D city modelling, corridor mapping and asset management. |
| first_indexed | 2025-11-14T08:00:30Z |
| format | Conference Paper |
| id | curtin-20.500.11937-26232 |
| institution | Curtin University Malaysia |
| institution_category | Local University |
| last_indexed | 2025-11-14T08:00:30Z |
| publishDate | 2013 |
| publisher | ISPRS |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | curtin-20.500.11937-262322023-02-08T05:20:33Z Robust locally weighted regression for ground surface extraction in mobile laser scanning 3D data Nurunnabi, Abdul West, Geoff Belton, David M. Scaioni R. C. Lindenbergh S. Oude Elberink D. Schneider F. Pirotti Terrain Classification DTM Point Cloud Filtering 3D Modelling Segmentation Feature Extraction Outlier A new robust way for ground surface extraction from mobile laser scanning 3D point cloud data is proposed in this paper. Fitting polynomials along 2D/3D points is one of the well-known methods for filtering ground points, but it is evident that unorganized point clouds consist of multiple complex structures by nature so it is not suitable for fitting a parametric global model. The aim of this research is to develop and implement an algorithm to classify ground and non-ground points based on statistically robust locally weighted regression which fits a regression surface (line in 2D) by fitting without any predefined global functional relation among the variables of interest. Afterwards, the z (elevation)-values are robustly down weighted based on the residuals for the fitted points. The new set of down weighted z-values along with x (or y) values are used to get a new fit of the (lower) surface (line). The process of fitting and down-weighting continues until the difference between two consecutive fits is insignificant. Then the final fit represents the ground level of the given point cloud and the ground surface points can be extracted. The performance of the new method has been demonstrated through vehicle based mobile laser scanning 3D point cloud data from urban areas which include different problematic objects such as short walls, large buildings, electric poles, sign posts and cars. The method has potential in areas like building/construction footprint determination, 3D city modelling, corridor mapping and asset management. 2013 Conference Paper http://hdl.handle.net/20.500.11937/26232 10.5194/isprsannals-II-5-W2-217-2013 ISPRS unknown |
| spellingShingle | Terrain Classification DTM Point Cloud Filtering 3D Modelling Segmentation Feature Extraction Outlier Nurunnabi, Abdul West, Geoff Belton, David Robust locally weighted regression for ground surface extraction in mobile laser scanning 3D data |
| title | Robust locally weighted regression for ground surface extraction in mobile laser scanning 3D data |
| title_full | Robust locally weighted regression for ground surface extraction in mobile laser scanning 3D data |
| title_fullStr | Robust locally weighted regression for ground surface extraction in mobile laser scanning 3D data |
| title_full_unstemmed | Robust locally weighted regression for ground surface extraction in mobile laser scanning 3D data |
| title_short | Robust locally weighted regression for ground surface extraction in mobile laser scanning 3D data |
| title_sort | robust locally weighted regression for ground surface extraction in mobile laser scanning 3d data |
| topic | Terrain Classification DTM Point Cloud Filtering 3D Modelling Segmentation Feature Extraction Outlier |
| url | http://hdl.handle.net/20.500.11937/26232 |