This technology is an aerial image fusion method that extracts high-frequency components from multispectral (MS) and panchromatic (PAN) images using bilateral filtering. It creates a detail image by applying a dual-scale guided filter to the intensity high-frequency components—approximated via adaptive coefficients—and the PAN high-frequency components, which is then fused with the MS image.
Satellite sensors cannot provide high spatial and high spectral resolution simultaneously, necessitating pansharpening to fuse MS and PAN images. Existing fusion methods have limitations, often causing significant spectral distortion or failing to sufficiently restore spatial details missing from MS images.
This technology uses an optimization function to determine band-specific adaptive coefficients, approximating intensity components to match PAN images. It extracts multi-scale details from the difference between two-stage guided filter outputs, reducing spectral distortion while enhancing spatial resolution. Applicable to cartography, agricultural disaster detection, and smart city satellite map visualization, it provides clear, high-resolution multispectral images that retain original color fidelity.
N/A
