Land Cover Change Mapping at the Subpixel Scale With Different Spatial-Resolution Remotely Sensed Imagery

Abstract

Extracting land cover change (LCC) information at the subpixel scale is important when coarse-resolution remotely sensed images are used for change detection. Although fraction images derived from soft-classification technologies can be used for subpixel LCC detection, the spatial distribution of changed subpixels within each coarse-resolution pixel cannot be provided. This letter presents a subpixel LCC mapping (SLCCM) algorithm, aiming to predict the spatial pattern of LCC at the subpixel scale between bitemporal images through comparing the former high-resolution land cover map and the latter fraction images derived from the coarse-resolution image. The resulting subpixel LCC map is determined by the spatial dependence principle and an LCC rule in each mixed pixel. The proposed algorithm was evaluated with simulated and real images, and the results showed the effectiveness of the proposed method for SLCCM.