A new approach to extract rectangular building from aerial urban images

Proposed a new primitive-structure-based approach for extracting rectangular building from aerial urban images. We obtained all kinds of primitive-structure that compose a rectangle by analysing geometrical primitive structure from the straight-line-graph that has been produced by edge-detection and contour-tracking and straight-line-splitting, then merge these primitive structures to rectangles through the merging arithmetic. The experimental results show that the approach is many times computationally faster than previous methods. The results of performance assessment also support the robustness and precision of the approach developed.

[1]  Ramakant Nevatia,et al.  Detection of buildings using perceptual grouping and shadows , 1994, 1994 Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[2]  Alessandro Mecocci,et al.  Circular shape detection in remote sensing multispectral images , 1992 .

[3]  Ramakant Nevatia,et al.  Detecting runways in complex airport scenes , 1990, Comput. Vis. Graph. Image Process..

[4]  John F. Canny,et al.  A Computational Approach to Edge Detection , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5]  Taejung Kim,et al.  Development of a graph-based approach for building detection , 1999, Image Vis. Comput..