[The meteorological satellite spectral image registration based on Fourier-Mellin transform]

Area of Science:

• Meteorological Science
• Environmental Remote Sensing

Context:

• Meteorological satellite spectral images are crucial for weather and environmental research.
• Accurate image registration is fundamental for analyzing satellite data.
• Existing methods may require complex feature extraction.

Purpose:

• To develop a simplified and accurate image registration method for meteorological satellite spectral images.
• To establish a robust registration process using Fourier-Mellin transform.
• To enable precise alignment of satellite imagery with reference templates.

Summary:

• A new image registration method utilizes Fourier-Mellin transform for meteorological satellite spectral images.
• A landmark template is created from global coastline data.
• Infrared sub-images are processed using edge extraction (Sobel operator) and Fourier-Mellin transform for affine parameter determination.
• The method relies on curve matching, eliminating the need for feature point extraction.

Impact:

• The proposed method achieves high-speed and high-accuracy image registration.
• It simplifies the registration process, making it more accessible.
• Demonstrated robustness with FY-2D meteorological satellite infrared spectral data.