Rooftop Detection Based on 3D Line Data Using Fast Graph Search

This paper highlights 3D building rooftop detection procedure based on 3D line data. Our approach begins with area-based stereo matching method to generate digital elevation map (DEM). Subpixel interpolation, normalized cross correlation and multi-resolution scheme are employed to generate an elaborate DEM. 3D lines are evaluated by using line fitting of DEM on 2D line coordinates of ortho-image. In order to use elevation in line fitting, the reliability of elevation is tested using self-consistency. 3D rooftop hypothesis is established by using an undirected graph, in which close cycles represent complete rooftops hypotheses. We test the proposed method with the synthetic aerial images. The experiment result shows that the extracted 3D rooftop realistically reflects real urban site model.

[1]  Ramin Zabih,et al.  A non-parametric approach to visual correspondence , 1996 .

[2]  Ramakant Nevatia,et al.  Detection and Modeling of Buildings from Multiple Aerial Images , 2001, IEEE Trans. Pattern Anal. Mach. Intell..

[3]  Edward M. Riseman,et al.  Token-based extraction of straight lines , 1989, IEEE Trans. Syst. Man Cybern..

[4]  Mohammed Bennamoun,et al.  Non-parametric techniques for fast and robust stereo matching , 1997, TENCON '97 Brisbane - Australia. Proceedings of IEEE TENCON '97. IEEE Region 10 Annual Conference. Speech and Image Technologies for Computing and Telecommunications (Cat. No.97CH36162).

[5]  Pascal Fua,et al.  Self-Consistency: A Novel Approach to Characterizing the Accuracy and Reliability of Point Correspon , 1998, ICCV 1999.

[6]  M. J. Hannah.  A system for digital stereo image matching , 1989 .

[7]  Robert T. Collins,et al.  Task driven perceptual organization for extraction of rooftop polygons , 1994, Proceedings of 1994 IEEE Workshop on Applications of Computer Vision.