Subpixel-Level Edge Feature Matching for SAR and Optical Images Based on Zernike Moments
Huan Qian, Jianwei Yue, Min Chen, Modi Wang, Haiqiang Xin · 2020
A sub-pixel edge feature matching method based on Zernike moment is proposed. By using PPB filtering method to preprocess SAR images, Zernike moments are used to extract sub-pixel edge feature maps of SAR images and optical images. It innovatively proposes the extraction of phase consistency feature points based on low-contrast nonsuppressed SAR-Harris multi-scale space. Based on the feature detection results, a HOG feature descriptor is constructed on the subpixel edge feature map, and finally feature matching is performed according to geometric constraints. Experimental results show that this method has obvious advantages in matching SAR images with optical images compared with SIFT, Schwind, Suri, and SAR-SIFT methods.