A computer vision based approach for 3D building modelling of airborne laser scanner DSM data

Madhavan, B. Babu; Wang, Caihua; Tanahashi, Hideki; Hirayu, Hidekazu; Niwa, Yasumasa; Yamamoto, Kazuhiko; Tachibana, Keihachiro; Sasagawa, Tadashi

doi:10.1016/j.compenvurbsys.2005.01.001

Cited by 13 publications

(5 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A review of most of the methods can be found in Ioannidis et al (2009). With the increasing demand of 3D city models and availability of LiDAR data, 3D building reconstruction has received extensive attention, and many method for building reconstruction have been proposed (Gruen 1998, Haala and Brenner 1999, Maas and Vosselman 1999, Stilla and Jurkiewicz 1999, Stilla et al 2003, Suveg and Vosselman 2004, Brenner 2005, Madhavan et al 2006, Sugihara and Hayashi 2008, Alexander et al 2009, Jang and Jung 2009, Pu and Vosselman 2009, Tang et al 2010. Meanwhile, building damage assessment based on building reconstruction and image classification has become a research topic.…”

Section: Introductionmentioning

confidence: 99%

A framework for automated assessment of post-earthquake building damage using geospatial data

Dong

Guo

2011

International Journal of Remote Sensing

View full text Add to dashboard Cite

Based on the triangulated irregular network (TIN) model, barycentric coordinates and random points, a new method was developed for more accurate characterization of 3D shape signatures of buildings using Light Detection and Ranging (LiDAR) data and Geographic Information System (GIS). The new method was applied to four simulated building models with flat, pent, gable and hip roofs to test the detection of changes in 3D building shapes in a post-earthquake scenario:(1) a three-storey building model becomes two-storey after losing the first floor;(2) a severe damage of a two-storey building model; and (3) a total collapse of a two-storey building model. The new method was then applied to real LiDAR data from four buildings in Harris County, TX, USA. All the changes in 3D shapes of the models and real buildings were successfully detected using 3D shape signatures. Sensitivity analyses were carried out to test the influence of LiDAR point density and new points in TIN triangles on 3D shape signatures. The results suggest that LiDAR point density of 0.5 point m −2 or higher can generate stable 3D shape signatures of buildings, and that the density of the new points in TIN triangles does not affect 3D shape signatures significantly. Built upon the new method and results, a framework was proposed for an automated assessment of post-earthquake building damage using geospatial data. A flowchart was presented to provide more details of the framework, and the advantages and limitations of the framework were also discussed. It is expected that the framework can be effectively applied to post-earthquake building damage assessment and other disaster scenarios that involve major changes in 3D building shapes.

show abstract

Section: Introductionmentioning

confidence: 99%

A framework for automated assessment of post-earthquake building damage using geospatial data

Dong

Guo

2011

International Journal of Remote Sensing

View full text Add to dashboard Cite

show abstract

“…The first consists of studies that demonstrate methods for building height mapping using data for part of a city, at a resolution sufficient to understand intra-urban variations. While there are multiple studies that demonstrate the use of LIDAR for such building height mapping (Bonczak and Kontokosta, 2019; Madhavan et al, 2006; Malpica et al, 2013; Priestnall et al, 2000; Shan and Sampath, 2005), we do not discuss these since at present such methods are expensive and not easily scalable, especially in less wealthy countries where much of the future urbanization is expected to occur. Non-LIDAR based studies typically rely on stereo imagery (e.g., d'Angelo et al, 2010; Poli and Caravaggi, 2013), single-view satellite imagery (e.g., Liu et al, 2020) or high-resolution Synthetic Aperture Radar (SAR) data (e.g., Geiß et al, 2019; Gamba et al, 2000; Marconcini et al, 2014; Sun et al, 2017; Thiele et al, 2006).…”

Section: Introductionmentioning

confidence: 99%

“…Both LIDAR and SAR are active remote sensing systems in which laser and radio/microwaves, respectively, are transmitted and reflections are collected by a sensor (NOAA, n.d.; NASA, n.d.). In recent years, several studies have used LIDAR datasets to derive high accuracy building height maps (Bonczak and Kontokosta, 2019; Madhavan et al, 2006; Malpica et al, 2013; Priestnall et al, 2000; Shan and Sampath, 2005). However, in comparison to satellite stereo imagery, LIDAR surveys are extremely expensive to carry out, especially at the scale of megacities in the developing world where there are severe resource constraints.…”

Section: Introductionmentioning

confidence: 99%

Generating megacity-scale building height maps without DGNSS surveyed GCPs: An open-source approach

Tripathy

Balakrishnan

Franchis

et al. 2022

Environment and Planning B: Urban Analytics and City Science

View full text Add to dashboard Cite

This paper describes an open-source method for generating megacity-scale building height maps without proprietary software or Differential Global Navigation Satellite System surveyed ground control points (GCPs). We use the open-source Satellite Stereo Pipeline (S2P) software along with four scenes of 2.5m resolution Cartosat-1 data for Bengaluru to demonstrate this. Digital Surface Models (DSMs) of 5 m resolution are generated using S2P, and terrain removal is achieved using 30m SRTM data resampled to 5m. The resulting normalized DSM is calibrated and validated using 1270 GCPs. These were generated by counting the number of occupiable floors of buildings and marking zero elevation ground points. The final building height map of Bengaluru covers a total area of about 1420 km2, and across the four scenes, has RMSE values ranging from 2.8m to 3.9m—an error of approximately one floor. Furthermore, we implemented this workflow using stereo imagery for Mumbai, and the RMSE values obtained were comparable to those for Bengaluru. Hence the method we describe is a very cost-effective way of generating megacity-scale building height maps. The height maps generated using this method can be used to better understand numerous urban characteristics including land use intensity and population distribution and can play a crucial role in urban planning and policy making.

show abstract

“…However, recent research is developing methods to derive models in an automated manner. Concerning 3D city modelling, recent work focuses on the modelling of buildings using ALS data (Brenner, 2005; Madhavan et al, 2006; Oude Elberink and Vosselman, 2009) and TLS and MLS data for detailed façade reconstruction (Früh et al, 2005; Pu and Vosselman, 2009; Tang et al, 2010). So far only a little attention has been paid to the detection and modelling of vegetation from MLS for this purpose.…”

Section: Introductionmentioning

confidence: 99%

Tree modelling from mobile laser scanning data‐sets

Rutzinger

Pratihast

Elberink

et al. 2011

The Photogrammetric Record

121

View full text Add to dashboard Cite

In recent times mobile laser scanning (MLS) has been used to acquire massive 3D point clouds in urban areas and along road corridors for the collection of detailed data for 3D city modelling, building façade reconstruction and capture of vegetation and road features for inventories. The objectives of this paper are the extraction of tree features from such data‐sets and the modelling of trees for the purpose of visualisation in 3D city models. After the detection of high vegetation the point cloud is reduced using a 3D alpha shape approach. Then the required model parameters such as crown and stem height, crown and stem diameter, and crown shape are derived and the trees are modelled individually in a realistic manner. The tree model so generated correctly represents the overall appearance of the tree. However, the inner structure such as the branching of the tree crown is parameterised. The workflow reduces the point cloud by means of a step‐by‐step process, which eases the handling of the massive MLS data‐sets. The thinning using 3D alpha shapes reduces the amount of data to be processed by about 95%. It is shown that the model parameters are not influenced by the thinning procedure employed. This proves the robustness of the data reduction method and the tree modelling approach.

show abstract

A computer vision based approach for 3D building modelling of airborne laser scanner DSM data

Cited by 13 publications

References 18 publications

A framework for automated assessment of post-earthquake building damage using geospatial data

A framework for automated assessment of post-earthquake building damage using geospatial data

Generating megacity-scale building height maps without DGNSS surveyed GCPs: An open-source approach

Tree modelling from mobile laser scanning data‐sets

Contact Info

Product

Resources

About