Nonlinear intensity measurement for multi-source images based on structural similarity

Feature-based algorithms are widely used in automatic matching of multi-source images (e.g., LiDAR, optical, infrared, map, and SAR images). However, it remains a challenging task to find sufficient correct correspondences for image pairs in the presence of significant noise and nonlinear intensity differences. To solve this problem, this paper proposes a novel feature descriptor named the histogram of maximum phase congruency (HMPC), which is based on the structural properties of images. Then, a novel distance formula is designed by normalizing the phase orientation and histogram value to calculate the similarity. Furthermore, the precise bilateral matching principle and consistency-checking algorithm based on the phase orientation are used to perform matching between the corresponding point sets. Benefiting from combinatorial features, the proposed method can effectively capture the structural information of images and present robust matching performance for complex texture structures and noise images compared to that of the sole feature, and it has been tested with a variety of SAR, LiDAR, optical,and map datas. The results demonstrate that the proposed HMPC achieves a more robust and accurate matching performance than many state-of-the-art methods.

Automated summary

This paper proposes a novel feature descriptor HMPC based on the structural properties of images, with a new distance formula designed for calculating similarity. The method shows robust and accurate matching performance compared to state-of-the-art methods when applied to multi-source image matching tasks.