Registration between SAR and optical images based on improved phase consistency algorithm

The traditional Scale-Invariant Feature Transform (SIFT) algorithm relies heavily on grayscale information, which poses challenges when dealing with the grayscale differences between optical and Synthetic Aperture Radar (SAR) images. It may even lead to reverse differences, failing to capture the consistency between the two image types. Therefore, we propose an enhanced SIFT algorithm that adopts a uniform feature point extraction strategy and focuses on maintaining gradient consistency to obtain highly consistent feature point information between optical and SAR images. By employing this improved strategy, we can better capture the common features between optical and SAR images, resulting in more accurate and robust image registration. This approach significantly enhances the accuracy and reliability of registration, especially when dealing with discrepancies between optical and SAR images. Experimental results demonstrate its superior matching performance with respect to the state-of-the-art methods.