Building Detection Using Segmentation, Corner Detection andMorphological Operators in Satellite Images

With the accessibility of high resolution satellite images, there has been much interest to extract manmadeobjects such as buildings. The building detection is one of the most difficult problems in high resolution satelliteimages. In this paper an automatic building detection technique has been proposed. Firstly Otsu’s thresholdingmethod with a determined factor is used to segment input image and extract candidate building regions. In thesecond step morphological operations are used to improve the results. Meanwhile the corners of the original imageare achieved by Harris Corner Detector . Then, the true building points are looked up in the domain with acombination of segmentation and corner detection. The filtered regions (by morphological operators) that aresurrounded by corner points are selected as final candidates. Experimental results show that the proposed techniquemade promising results.

Keywords:

building detection , satellite images , corner detection , morphological operators , segmentation