Integrated Registration Method with Enhanced Position Awareness for Optical and SAR Images
Image registration is the basis for optical and SAR image information fusion.Most of the existing typical reg-istration methods rely on feature-point detection and matching.However,because of their poor applicability to different scene regions,these methods are prone to problems such as excessive mismatched points and insufficiently effective matched points,resulting in invalid registration.Therefore,this study investigated an integrated registration method with enhanced position awareness for optical and SAR images.This method utilizes a deep neural network to directly re-gress the geometric transformation relationship between images.The proposed method achieves end-to-end high-precision registration without relying on feature-point detection.First,a feature-extraction module that integrates coordi-nate attention is used in the backbone network to extract position-sensitive fine-grained features from the input image pairs.Second,the multiscale features of the backbone network output are fused,taking into account the positional infor-mation of low-level features and semantic information of high-level features.Finally,a loss function that combines the position deviation and image similarity is used to optimize the registration results.Experimental results based on a pub-licly available high-resolution optical and SAR dataset(OS-Dataset)demonstrated that compared with four existing typi-cal algorithms(OS-SIFT,RIFT2,DHN,and DLKFM),the proposed method had good robustness for different scene areas such as urban,farmland,river,repetitive texture,and weak texture scenes,and outperformed the existing algo-rithms in terms of visual effects and a quantitative precision metric.The percentage of average corner errors of fewer than 3 pixels was more than 25%better than that of DLKFM,which had the highest precision among the four algo-rithms.The registration speed was comparable to that of DHN,which was the fastest of the four algorithms.The pro-posed method could achieve high-precision and high-efficiency optical and SAR image registration.
NETL
NSTL
万方数据
