An automatic method of area change detection based on histogram matching and morphological operation in high spatial remote sensed imagery

摘要

Remote sensing images of high spatial resolution are widely available that can be obtained or booked in a few days at intervals. A variety of infrastructures such as buildings, roads, squares in suburb and urban region assume area shapes in high spatial imagery. This paper proposes an automatic method for detection of changed patches based on histogram-matching and morphological operation. A pair of two images acquired at different time covering the same region is firstly bilateral matched in spectral range. Then the difference image is derived from the minus or ratio between them. The two kinds of operators designed, implement erosion and dilation on the difference image. Results show the method can effectively separate area changes while wiping out small area due to spectral variation under different illumination conditions.