Novel Registration and Fusion Algorithm for Multimodal Railway Images with Different Field of Views

Baoqing Guo,Zujun Yu,Liqiang Zhu,Xingfang Zhou,Yingzi Lin

doi:10.1155/2018/7836169

Baoqing Guo, Zujun Yu + Show 3 more

Open Access

https://doi.org/10.1155/2018/7836169

Copy DOI

Abstract

Objects intruding high-speed railway clearance do great threat to running trains. In order to improve accuracy of railway intrusion detection, an automatic multimodal registration and fusion algorithm for infrared and visible images with different field of views is presented. The ratio of the nearest to next nearest distance, geometric, similar triangle, and RANSAC constraints are used to refine the matching SURF feature points successively. Correct matching points are accumulated with multiframe to overcome the insufficient matching points in single image pair. After being registered, an improved Contourlet transform fusion algorithm combined with total variation and local region energy is proposed. Inverse Contourlet transform to low frequency subband coefficient fused with total variation model and high frequency subband coefficients fused with local region energy is used to reconstruct the fused image. The comparison to other 4 popular fusion methods shows that our algorithm has the best comprehensive performance for multimodal railway image fusion.

Highlights

High-speed railway is a competitive transportation all over the world
In order to solve the problem of unclear edge and textures, we proposed an improved Contourlet transform method combined with total variation (TV) model and local region energy
We propose a novel registration and fusion algorithm for multimodal railway images with different field of views

Summary

Introduction

High-speed railway is a competitive transportation all over the world. As of 2017, the length of commercial high-speed railway lines in China had exceeded 25,000 km. Visible images get the detailed information of colors and texture under good illumination. IR and visible image fusion will bring great benefit to railway clearance intrusion detection. The quality of fusion is bounded by the quality of fusion algorithm, and by the outcome of prior registration algorithm Due to this dependency, images are always assumed to be pre-aligned [9]. Visible and IR images acquired separately are not pre-aligned and even have different field of views and depth. These bring great difficulties to image fusion. We need both excellent registration and fusion algorithms in railway applications

Methods

Results

Conclusion