Phase quality map based on local multi-unwrapped results for two-dimensional phase unwrapping.

Area of Science:

• Geosciences
• Remote Sensing
• Computer Science

Background:

• Phase unwrapping is crucial for Digital Elevation Model (DEM) generation from Interferometric Synthetic Aperture Radar (InSAR) and Interferometric Synthetic Aperture Sonar (InSAS) data.
• Algorithm efficiency and result reliability are key challenges in InSAR/InSAS phase unwrapping.

Purpose of the Study:

• To develop and implement a novel phase quality map for improving the efficiency and accuracy of phase unwrapping algorithms.
• To accelerate the phase unwrapping process using a Graphics Processing Unit (GPU) environment.

Main Methods:

• A new phase quality map is designed, calculating quality based on the difference between two integration paths within a local wrapped phase window.
• The quality map computation is implemented on a GPU for parallel processing, uploading wrapped phase data to device memory.
• Kernel functions are utilized on the GPU to compute phase quality in parallel using blocks of threads.

Main Results:

• The GPU implementation significantly accelerates the computation of the phase quality map.
• The proposed quality-guided algorithm demonstrates enhanced correctness and reliability of the unwrapped phase results.
• Tests on simulated and real InSAS data validate the accuracy and efficiency of the method.

Conclusions:

• The developed GPU-accelerated phase quality map effectively addresses the efficiency and reliability challenges in InSAR/InSAS phase unwrapping.
• This method provides a robust solution for accurate DEM reconstruction from interferometric data.
• The parallel processing approach on GPUs offers substantial performance gains for complex remote sensing data processing.