BUILDING EXTRACTION FROM REMOTE SENSING IMAGERIES BY DATA FUSION TECHNIQUES

Ezhili .G, Akshaya .V.S

Abstract: This paper presents a data fusion approach for manmade objects extraction from high-resolution IKONOS satellite images. Buildings can have various complex forms and roofs of various compositional materials. Their automatic extraction from imagery is a very difficult problem. Applying normal image processing methods could not achieve satisfied performance, especially for high-resolution satellite images. It is based on edge maps derived from IKONOS data. Local changes or variations of the intensity of the imagery (such as edges and corners) are important information for image processing and pattern recognition. K-MEANS clustering is one of the most popular techniques that can be used to classify satellite images. This technique coupled with canny edge detection, which has double threshold technique is less fooled by noise, forms a very good tool in detection of man-made features. The above mentioned techniques are applied to one meter IKONOS imagery of the highly urbanized Singapore city, to detect building edges within scene.

Keywords: Canny edge detection, RGB color matrices, Gaussian filter, K-means clustering, non-maximum suppression.

DOI: https://doi.org/10.15623/ijret.2013.0203021