Page 1 of 1
Sigma 0 for Urban Change Detection
Dr. François Cavayas
Geography Department, University of Montreal, Canada,
Tel: +1-514 343 8016,
E-mail: Francois.cavayas@umontreal.ca
Dr. Yves Baudouin
eography Department, University of Québec in Montreal, Canada,
Tel: +1-514 987-3000 7706#,
E-mail: Baudouin.yves@uqam.ca
Claire Gosselin
Synetix Inc., Canada, Tel: +1-514-598-1991,
E-mail: claire@gosselin.com
Dr. Rejean Simard
Synetix inc., Malaysia, Tel : +60-12-235-8397,
E-mail: rejean@simard.com
ABSTRACT:
Many public-sector agencies require current, reliable data on urbanized areas. Updating
geospatial databases at scales ranging from 1:10 000 to 1:250 000 stands out as one of the major
concerns of agencies involved in land management. What they are looking for is an efficient and
economical system for detecting urban change and facilitating map updating. SIGMA 0 , a new
geomatics system aimed at updating land-use maps in urban areas with RADARSAT-1 or
Envisat imagery, could be just what they need.
Until now, there has been no automated, rapid, and effective method for detecting change. Part
of the problem lies with optical satellite images usually applied for change detection. In fact
urban area heterogeneity and object size make using very high resolution images (less than 5 m)
necessary. These images, however, contain an excessive amount of spatial and spectral detail
that makes tasks as basic as determining by automatic means a building's perimeter complicated
and very often inaccurate.
RADARSAT and Envisat offer a solution with images that have a resolution between 10 and 25
m that are not affected by weather conditions. Radar images are sensitive to the presence of
elevated objects, such as buildings. The appearance of buildings on areas previously known as
vacant is an undeniable indicator of urban change . Tests carried out with SIGMA 0 show that it
automatically detects nearly all new buildings erected on the vacant lots indicated on the land-use
map to be updated. The percentage of error is usually inferior to 5%.
SIGMA 0 was developed by the Laboratory of Remote Sensing, University of Montreal, and the
GIS Laboratory, University of Quebec in Montreal. It is presently being commercialized and
adapted for new map updating applications with the collaboration of SYNETIX and under the
financial support of the Canadian Space Agency and the Quebec Ministry of Economic
Development, and Regional and Research.
Page 1 of 1
