Linear Features for Semi-Automatic Registration and Change Detection of Multi-Source Imagery

摘要

Minimizing geometric and radiometric differences within images are key issues for change detection. To compensate for geometric differences, an accurate image registration method is suggested using the Modified Iterated Hough Transform (MIHT) as a matching strategy. Once geometric differences have been compensated, radiometric differences will be circumvented by using extracted edges, which are invariant to changes in the illumination conditions. Iterated Edge Filling (IEF) method is applied as a new change detection method. Several templates based on geometric shapes of artificial features are used to fill gaps between edges in each image. After minimizing geometric and radiometric differences, difference images are generated to analyze changed area. Experimental results using real data proved the feasibility of the suggested approach for deriving a quantitative estimate of changes among the registered temporal images.