The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2012
This work presents a combined bottom-up and top-down approach to extraction and refinement of bui... more This work presents a combined bottom-up and top-down approach to extraction and refinement of building footprints from airborne LIDAR data. Building footprints are interesting for many applications in urban planning. The cadastral maps, however, may be limited for certain areas or not be updated frequently. Airborne laser scanning data is therefore considered by many people in the last decade as an important alternative data for change detection and update of building footprints. Laser scanning data of city scenes, however, often shows noise and incompleteness because of, e.g., the clutter by vegetation and the reflection of windows/waterlogged depressions on the roof. Results of the bottom-up detection may thus be limited to incomplete or irregular polygons. We employ 3D Hough transform to detect the building points. An improved joint multiple-plane detection scheme is proposed to find and label the laser points on multiple roof facets synchronously. The bottom-up processing provides not only a rough point segmentation but also additional 3D information, e.g., roof heights and horizontal ridges. Using these as priors, a top-down reconstruction is conducted via generative models. We consider the building footprint as an assembly of regular primitives. A statistical search by means of Reversible Jump Markov Chain Monte Carlo and Maximum A Posteriori estimation is implemented to find the optimal configuration of the footprint. By these means a robust and plausible reconstruction is guaranteed. First results on point clouds with various resolutions show the potential of this approach.
In this paper we propose a generative statistical approach for the three dimensional (3D) extract... more In this paper we propose a generative statistical approach for the three dimensional (3D) extraction of the branching structure of unfoliaged deciduous trees from urban image sequences. The trees are generatively modeled in 3D by means of L-systems. A statistical approach, namely Markov Chain Monte Carlo-MCMC is employed together with cross correlation for extraction. Thereby we overcome the complexity and uncertainty of extracting and matching branches in several images due to weak contrast, background clutter, and particularly the varying order of branches when projected into different images. First results show the potential of the approach.
In this paper we propose an approach for the three-dimensional (3D) extraction of the branching s... more In this paper we propose an approach for the three-dimensional (3D) extraction of the branching structure of unfoliaged deciduous trees from urban wide-baseline image sequences. The trees are generatively modeled in 3D by means of L-systems. A statistical approach, namely Markov Chain Monte Carlo (MCMC) is employed together with cross correlation for the extraction of the branches. With this generative statistical approach we avoid the complexity and uncertainty of extracting and matching branches in several images due to weak contrast, background clutter, and particularly the varying order of branches when projected into different images. First results show the potential of the approach. Zusammenfassung: Extraktion der 3D Verzweigungsstruktur unbelaubter Laubbäume aus Bildsequenzen. Dieses Papier stellt einen Ansatz für die Extraktion der drei-dimensionalen (3D) Verzweigungsstruktur unbelaubter Laubbäume aus städtischen Bildsequenzen mit langer Basis vor. Die Bäume werden in 3D mittels L-Systemen modelliert. Markoff Ketten Monte Carlo (MCMC) wird zusammen mit Kreuzkorrelation für die Extraktion der Äste genutzt. Mit diesem generativen statistischen Ansatz wird die Komplexität und Unsicherheit der Extraktion und Zuordnung von Ästen in mehreren Bildern wegen schwachem Kontrast, Störobjekten im Hintergrund und insbesondere der z.T. unterschiedlichen Ordnung der Äste nach Projektion in verschiedene Bilder vermieden. Erste Ergebnisse zeigen das Potential des Ansatzes.
We present an automatic building type (usage) labeling based on the footprint data. The usage inf... more We present an automatic building type (usage) labeling based on the footprint data. The usage information of buildings is of great interest for many applications, e.g., navigation, city planning and emergency management. This attribute, however, is generally not provided in the volunteered data sources like OpenStreetMap and is often incomplete even in the official cadastral maps. In this paper, we propose a method to enhance the maps with the building usage information exclusively using the geometric and topological features in the footprint data. A general category is predefined with four classes: residential, commercial, industrial and public. A novel inference framework is proposed using two new high-level (composite) geometric characteristics for the local description of the individual buildings and the Markov Random Field model to incorporate the contextual constraints of the neighborhood. Experiments are performed on both OpenStreetMap and cadastral data showing the potential of the proposed method.
Annals of Gis / Geographic Information Sciences, 2009
This paper presents a generative statistical approach for the automatic three-dimensional (3D) ex... more This paper presents a generative statistical approach for the automatic three-dimensional (3D) extraction and reconstruction of unfoliaged deciduous trees from terrestrial wide-baseline image sequences. Unfoliaged trees are difficult to reconstruct from images due to partially weak contrast, background clutter, occlusions, and particularly the possibly varying order of branches in images from different viewpoints. This work combines generative modeling by L-systems
In this paper we propose a generative statistical approach for the three dimensional (3D) extract... more In this paper we propose a generative statistical approach for the three dimensional (3D) extraction of the branching structure of unfoliaged deciduous trees from urban image sequences. The trees are generatively modeled in 3D by means of L-systems. A statistical approach, namely Markov Chain Monte Carlo -MCMC is employed together with cross correlation for extraction. Thereby we overcome the complexity and uncertainty of extracting and matching branches in several images due to weak contrast, background clutter, and particularly the varying order of branches when projected into different images. First results show the potential of the approach.
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2012
This work presents a combined bottom-up and top-down approach to extraction and refinement of bui... more This work presents a combined bottom-up and top-down approach to extraction and refinement of building footprints from airborne LIDAR data. Building footprints are interesting for many applications in urban planning. The cadastral maps, however, may be limited for certain areas or not be updated frequently. Airborne laser scanning data is therefore considered by many people in the last decade as an important alternative data for change detection and update of building footprints. Laser scanning data of city scenes, however, often shows noise and incompleteness because of, e.g., the clutter by vegetation and the reflection of windows/waterlogged depressions on the roof. Results of the bottom-up detection may thus be limited to incomplete or irregular polygons. We employ 3D Hough transform to detect the building points. An improved joint multiple-plane detection scheme is proposed to find and label the laser points on multiple roof facets synchronously. The bottom-up processing provides not only a rough point segmentation but also additional 3D information, e.g., roof heights and horizontal ridges. Using these as priors, a top-down reconstruction is conducted via generative models. We consider the building footprint as an assembly of regular primitives. A statistical search by means of Reversible Jump Markov Chain Monte Carlo and Maximum A Posteriori estimation is implemented to find the optimal configuration of the footprint. By these means a robust and plausible reconstruction is guaranteed. First results on point clouds with various resolutions show the potential of this approach.
In this paper we propose a generative statistical approach for the three dimensional (3D) extract... more In this paper we propose a generative statistical approach for the three dimensional (3D) extraction of the branching structure of unfoliaged deciduous trees from urban image sequences. The trees are generatively modeled in 3D by means of L-systems. A statistical approach, namely Markov Chain Monte Carlo-MCMC is employed together with cross correlation for extraction. Thereby we overcome the complexity and uncertainty of extracting and matching branches in several images due to weak contrast, background clutter, and particularly the varying order of branches when projected into different images. First results show the potential of the approach.
In this paper we propose an approach for the three-dimensional (3D) extraction of the branching s... more In this paper we propose an approach for the three-dimensional (3D) extraction of the branching structure of unfoliaged deciduous trees from urban wide-baseline image sequences. The trees are generatively modeled in 3D by means of L-systems. A statistical approach, namely Markov Chain Monte Carlo (MCMC) is employed together with cross correlation for the extraction of the branches. With this generative statistical approach we avoid the complexity and uncertainty of extracting and matching branches in several images due to weak contrast, background clutter, and particularly the varying order of branches when projected into different images. First results show the potential of the approach. Zusammenfassung: Extraktion der 3D Verzweigungsstruktur unbelaubter Laubbäume aus Bildsequenzen. Dieses Papier stellt einen Ansatz für die Extraktion der drei-dimensionalen (3D) Verzweigungsstruktur unbelaubter Laubbäume aus städtischen Bildsequenzen mit langer Basis vor. Die Bäume werden in 3D mittels L-Systemen modelliert. Markoff Ketten Monte Carlo (MCMC) wird zusammen mit Kreuzkorrelation für die Extraktion der Äste genutzt. Mit diesem generativen statistischen Ansatz wird die Komplexität und Unsicherheit der Extraktion und Zuordnung von Ästen in mehreren Bildern wegen schwachem Kontrast, Störobjekten im Hintergrund und insbesondere der z.T. unterschiedlichen Ordnung der Äste nach Projektion in verschiedene Bilder vermieden. Erste Ergebnisse zeigen das Potential des Ansatzes.
We present an automatic building type (usage) labeling based on the footprint data. The usage inf... more We present an automatic building type (usage) labeling based on the footprint data. The usage information of buildings is of great interest for many applications, e.g., navigation, city planning and emergency management. This attribute, however, is generally not provided in the volunteered data sources like OpenStreetMap and is often incomplete even in the official cadastral maps. In this paper, we propose a method to enhance the maps with the building usage information exclusively using the geometric and topological features in the footprint data. A general category is predefined with four classes: residential, commercial, industrial and public. A novel inference framework is proposed using two new high-level (composite) geometric characteristics for the local description of the individual buildings and the Markov Random Field model to incorporate the contextual constraints of the neighborhood. Experiments are performed on both OpenStreetMap and cadastral data showing the potential of the proposed method.
Annals of Gis / Geographic Information Sciences, 2009
This paper presents a generative statistical approach for the automatic three-dimensional (3D) ex... more This paper presents a generative statistical approach for the automatic three-dimensional (3D) extraction and reconstruction of unfoliaged deciduous trees from terrestrial wide-baseline image sequences. Unfoliaged trees are difficult to reconstruct from images due to partially weak contrast, background clutter, occlusions, and particularly the possibly varying order of branches in images from different viewpoints. This work combines generative modeling by L-systems
In this paper we propose a generative statistical approach for the three dimensional (3D) extract... more In this paper we propose a generative statistical approach for the three dimensional (3D) extraction of the branching structure of unfoliaged deciduous trees from urban image sequences. The trees are generatively modeled in 3D by means of L-systems. A statistical approach, namely Markov Chain Monte Carlo -MCMC is employed together with cross correlation for extraction. Thereby we overcome the complexity and uncertainty of extracting and matching branches in several images due to weak contrast, background clutter, and particularly the varying order of branches when projected into different images. First results show the potential of the approach.
Uploads
Papers by Hai Huang