Quantifying the Effect of PAN/MS Resolution Ratios on Pansharpening Quality Using IDSSIM

Saadi, Ramin; Imanpour, Fatemeh; Saradjian Maralan, Mohammad Reza

doi:10.22059/eoge.2025.400279.1186

Quantifying the Effect of PAN/MS Resolution Ratios on Pansharpening Quality Using IDSSIM

Document Type : Original Article

Authors

Ramin Saadi

Fatemeh Imanpour

Mohammad Reza Saradjian Maralan

School of Surveying and Geospatial Engineering, University of Tehran, Tehran, Iran.

10.22059/eoge.2025.400279.1186

Abstract

Pansharpening generates high-resolution multispectral (MS) imagery by combining the fine spatial detail of panchromatic (PAN) images with the spectral richness of MS data. While most prior studies emphasize algorithm development, the effect of the PAN-to-MS spatial resolution ratio on fusion quality has been largely overlooked. In this work, we systematically examine ratios ranging from 1:2 to 1:16, using Wald’s protocol to simulate consistent PAN and MS inputs. A benchmark dataset of nine image pairs from GeoEye-1, WorldView-2, WorldView-3, and WorldView-4—covering diverse urban, vegetated, and agricultural scenes—was employed. Fusion was performed with the Gram–Schmidt method, and quality was evaluated using the Image Decomposition-based Structural Similarity (IDSSIM) index. Results show that moderate ratios, particularly around 1:5, consistently yield the most favorable balance between spatial sharpness and spectral preservation, although some variation occurs depending on scene characteristics. These findings demonstrate that resolution ratio selection plays a decisive role in pansharpening performance and provide practical guidance for both operational workflows and future sensor design.

Keywords

Pansharpening

Image Fusion

Gram–Schmidt

IDSSIM

Subjects