The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Download
Citation
Articles | Volume XL-1/W5
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XL-1/W5, 387–393, 2015
https://doi.org/10.5194/isprsarchives-XL-1-W5-387-2015
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XL-1/W5, 387–393, 2015
https://doi.org/10.5194/isprsarchives-XL-1-W5-387-2015

  11 Dec 2015

11 Dec 2015

AUTOMATIC URBAN ILLEGAL BUILDING DETECTION USING MULTI-TEMPORAL SATELLITE IMAGES AND GEOSPATIAL INFORMATION SYSTEMS

N. Khalili Moghadam1, M. R. Delavar2, and P. Hanachee3 N. Khalili Moghadam et al.
  • 1GIS Dept., School of Surveying and Geospatial Eng., College of Eng., University of Tehran, Tehran, Iran
  • 2Centre of Excellence in Geomatic Eng. in Disaster Management, School of Surveying and Geospatial Eng., College of Eng., University of Tehran, Tehran, Iran
  • 3School of Urban Planning, College of Fine Arts, University of Tehran, Tehran, Iran

Keywords: Automatic illegal building detection, building infractions, multi-temporal satellite images, K-means, Geospatial Information System

Abstract. With the unprecedented growth of urban population and urban development, we are faced with the growing trend of illegal building (IB) construction. Field visit, as the currently used method of IB detection, is time and man power consuming, in addition to its high cost. Therefore, an automatic IB detection is required. Acquiring multi-temporal satellite images and using image processing techniques for automatic change detection is one of the optimum methods which can be used in IB monitoring. In this research an automatic method of IB detection has been proposed. Two-temporal panchromatic satellite images of IRS-P5 of the study area in a part of Tehran, the city map and an updated spatial database of existing buildings were used to detect the suspected IBs. In the pre-processing step, the images were geometrically and radiometrically corrected. In the next step, the changed pixels were detected using K-means clustering technique because of its quickness and less user’s intervention required. Then, all the changed pixels of each building were identified and the change percentage of each building with the standard threshold of changes was compared to detect the buildings which are under construction. Finally, the IBs were detected by checking the municipality database. The unmatched constructed buildings with municipal database will be field checked to identify the IBs. The results show that out of 343 buildings appeared in the images; only 19 buildings were detected as under construction and three of them as unlicensed buildings. Furthermore, the overall accuracies of 83%, 79% and 75% were obtained for K-means change detection, detection of under construction buildings and IBs detection, respectively.