A No-Reference Multivariate Gaussian–Based Spectral Distortion Index for Pansharpened Images | Advanced Image Quality Assessment #WorldResearchAwards
Introduction Pansharpening plays a crucial role in remote sensing by integrating high-spatial-resolution panchromatic data with lower-resolution multispectral imagery to produce visually rich fused products. Despite its widespread use, pansharpening often introduces spectral distortions that can undermine the reliability of quantitative analyses such as classification, change detection, and biophysical parameter retrieval. Traditional quality assessment approaches, especially full-reference metrics, are limited by the availability of ground truth data, while many no-reference (NR) methods struggle to distinguish spectral distortions from spatial artifacts. This challenge motivates the development of robust NR spectral quality indices tailored specifically for pansharpened imagery. Limitations of Existing No-Reference Quality Metrics Current NR quality assessment methods, including widely used indices such as QNR, primarily focus on global consistency or mixed spatial–spectral pro...