A Coarse-to-Fine Change Detection Method for Remote Sensing Sparse Cultivated Land

Remote sensing (RS) images contain rich geographic information. For specific application scenarios like cultivated land, it is necessary to select areas of interest to reduce data scale and focus on detailed features. In this article, an innovative coarse-to-fine change detection method (CFCD) for sparse cultivated land is proposed to address these problems. Coarse screening module (CSM) first removes irrelevant low-difference image pairs, and then fine detection module (FDM) accurately locate change areas in remaining images. Experimental results show that two coarse screening methods can take out many disturbed images, and provide strong support for subsequent fine detection methods to achieve performance improvement.