The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Download
Citation
Articles | Volume XXXVIII-4/W19
https://doi.org/10.5194/isprsarchives-XXXVIII-4-W19-233-2011
https://doi.org/10.5194/isprsarchives-XXXVIII-4-W19-233-2011
07 Sep 2012
 | 07 Sep 2012

FITTING OF PARAMETRIC BUILDING MODELS TO OBLIQUE AERIAL IMAGES

U. S. Panday and M. Gerke

Keywords: 3D reconstruction, parametric building model, fitting, plane sweeping, overhang estimation, oblique aerial images, laser scannin

Abstract. In literature and in photogrammetric workstations many approaches and systems to automatically reconstruct buildings from remote sensing data are described and available. Those building models are being used for instance in city modeling or in cadastre context. If a roof overhang is present, the building walls cannot be estimated correctly from nadir-view aerial images or airborne laser scanning (ALS) data. This leads to inconsistent building outlines, which has a negative influence on visual impression, but more seriously also represents a wrong legal boundary in the cadaster. Oblique aerial images as opposed to nadir-view images reveal greater detail, enabling to see different views of an object taken from different directions. Building walls are visible from oblique images directly and those images are used for automated roof overhang estimation in this research. A fitting algorithm is employed to find roof parameters of simple buildings. It uses a least squares algorithm to fit projected wire frames to their corresponding edge lines extracted from the images. Self-occlusion is detected based on intersection result of viewing ray and the planes formed by the building whereas occlusion from other objects is detected using an ALS point cloud. Overhang and ground height are obtained by sweeping vertical and horizontal planes respectively.
Experimental results are verified with high resolution ortho-images, field survey, and ALS data. Planimetric accuracy of 1cm mean and 5cm standard deviation was obtained, while buildings' orientation were accurate to mean of 0.23° and standard deviation of 0.96° with ortho-image. Overhang parameters were aligned to approximately 10cm with field survey. The ground and roof heights were accurate to mean of – 9cm and 8cm with standard deviations of 16cm and 8cm with ALS respectively. The developed approach reconstructs 3D building models well in cases of sufficient texture. More images should be acquired for completeness of overhang results and automatic accuracy check of roof parameters.