Iterative Feedback–Optimized IHS Pansharpening: Application to IKONOS Imagery

Abstract

Intensity-Hue-Saturation (IHS) is a component-substitution method that employs mathematical transformation between IHS and Red-Green-Blue (RGB) color domains to transfer spatial detail from the panchromatic image to the resampled multispectral bands. In this study, an iterative feedback-optimized approach is proposed to enhance the IHS pansharpening method. This approach starts by applying a low-pass filter to the output of the IHS fusion, resulting in an updated low-resolution multispectral image. This image is then reintroduced into the IHS fusion process, and the steps are repeated until a predefined quality threshold, measured by the Quality with No Reference (QNR) index, is achieved. The mathematical analysis demonstrates that this iterative approach transfers spatial details from the panchromatic image to the multispectral bands gradually over multiple iterations, in contrast to the standard IHS method where this detail injection occurs in a single step. This allows controlled and data-driven enhancement of spatial information. Experiments using IKONOS satellite images show that the proposed approach consistently improves the fusion quality of IHS pansharpening both numerically and visually, compared to the standard IHS algorithm and to two alternative baseline methods: Principal Component Analysis (PCA) and High-Pass Filtering (HPF).