Detection of buildings from different satellite images by using image enhancement techniques

Abstract

The province of istanbul and its immediate surroundings have been facing serious environmental problems within the past forty years due to migration from the rural areas, industrial and tourism-oriented investment. The city has about 20% of the total population of the country while it makes up only 0.74% of the total surface area of Turkey. The conventional methods of urban data collection and analysis cannot deliver necessary information in a timely and cost effective fashion. To meet the demand for current and accurate data remotely sensed images are increasingly used as a source of data. As the resolution of remotely sensed data increases, more detailed information can be obtained. In this study, the Tuzla region, located in the east coastline of Istanbul, was examined. At the beginning this region developed as a small village having flat, productive agricultural lands and had summer residential area characteristics, then it developed rapidly because of regional investments. As for the satellite images in the study, SPOT satellite images dated 1999, IRS1C dated 2000, and LANDSAT 7 TM dated 2001 were used. For detection of buildings from satellite images, many image fusion algorithms and a high pass filtering technique have been used. At the end of the study, a comparison of these methods has been made.