Detecting changes of buildings using stereo matching from aerial images taken at different positions and times

Abstract

AbstractGIS base maps are updated by using parallax to compare two aerial images taken at different times to interpret structural changes. Since an aerial image generally photographs a wide range, this interpretation work requires a great deal of time. Therefore, in this paper, the authors propose a method of automatically extracting structural changes with high precision to reduce the load of this interpretation work. Not only do the times differ at which the aerial images used for comparison were taken, but the shooting positions also frequently differ. This causes differences in shadow shapes and differences in structure viewing methods, and when aerial images are compared during automatic interpretation, it will cause the detection precision of structural changes to decrease. To deal with this problem, the authors eliminated the effects of differences in structure viewing methods by applying stereo matching and letting noncorresponding regions be structural change candidates. In addition, to correctly detect structural changes from those candidates, the authors used the outdoor light source model, which had been proposed by previous authors, to eliminate the effects of differences in shadow shapes. The results that were obtained by the proposed method showed that the detection precision of structural changes was approximately four times that of the conventional method. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(6): 55–65, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20084