Robust Registration of Multimodal Remote Sensing Images Based on Structural Similarity

Abstract

Automatic registration of multimodal remote sensing data (e.g., optical,\nLiDAR, SAR) is a challenging task due to the significant non-linear radiometric\ndifferences between these data. To address this problem, this paper proposes a\nnovel feature descriptor named the Histogram of Orientated Phase Congruency\n(HOPC), which is based on the structural properties of images. Furthermore, a\nsimilarity metric named HOPCncc is defined, which uses the normalized\ncorrelation coefficient (NCC) of the HOPC descriptors for multimodal\nregistration. In the definition of the proposed similarity metric, we first\nextend the phase congruency model to generate its orientation representation,\nand use the extended model to build HOPCncc. Then a fast template matching\nscheme for this metric is designed to detect the control points between images.\nThe proposed HOPCncc aims to capture the structural similarity between images,\nand has been tested with a variety of optical, LiDAR, SAR and map data. The\nresults show that HOPCncc is robust against complex non-linear radiometric\ndifferences and outperforms the state-of-the-art similarities metrics (i.e.,\nNCC and mutual information) in matching performance. Moreover, a robust\nregistration method is also proposed in this paper based on HOPCncc, which is\nevaluated using six pairs of multimodal remote sensing images. The experimental\nresults demonstrate the effectiveness of the proposed method for multimodal\nimage registration.\n