Scaling-up Transformation of Multisensor Images with Multiple Resolutions

Area of Science:

• Remote Sensing
• Image Processing
• Geospatial Analysis

Background:

• Pansharpening low resolution multispectral images (LRMIs) with high resolution panchromatic images (HRPI) is crucial for detailed analysis.
• Traditional intensity-hue-saturation (IHS) methods offer spatial enhancement but often cause spectral distortion in fused high resolution multispectral images (HRMIs).

Purpose of the Study:

• To present a generalized intensity modulation (GIM) technique for improved LRMI pansharpening.
• To extend the IHS transform to handle an arbitrary number of LRMIs while preserving spectral fidelity.
• To minimize spectral distortion during the fusion of multispectral and panchromatic imagery.

Main Methods:

• Developed a generalized intensity modulation (GIM) by extending the IHS transform.
• Incorporated spectral response functions (SRFs) of sensors into the GIM.
• Enhanced generalized intensity by injecting details from HRPI using empirical mode decomposition before modulation.

Main Results:

• The GIM method successfully fused LRMIs with HRPI, achieving superior spatial and spectral quality compared to existing methods.
• Empirical mode decomposition effectively injected HRPI details, improving the generalized intensity.
• Visual analysis and Wald's protocol confirmed the proposed method's effectiveness on Quickbird imagery.

Conclusions:

• The proposed GIM method offers a significant advancement in pansharpening, providing spectrally and spatially enhanced HRMIs.
• GIM is a more robust and accurate approach for fusing LRMIs with HRPI, overcoming limitations of standard IHS techniques.
• The method demonstrates considerable potential for applications requiring high-fidelity fused multispectral imagery.