Research on binocular stereo matching based on infrared imaging

Infrared imaging is less affected by lighting conditions and smoke compared to visible light imaging, however, the lower resolution and lack rich texture details making infrared imaging unsuitable for stereo matching. To enhance the quality of infrared stereo imaging and 3D reconstruction, an advanced stereo matching algorithm was proposed in this paper. Firstly, the checkerboard based on Peltier effect is designed for infrared camera calibration, which provides an accurate calibration tool for infrared camera calibration. Secondly, the disparity map is obtained by these process of cost calculation, cost aggregation, and disparity calculation. And then, the disparity is optimized by belief propagation algorithm and Weighted Least Square method, which improves accuracy of the infrared binocular stereo matching. Finally, compared the experiment results with traditional BM and SGBM algorithm, the proposed algorithm demonstrates excellent performance in stereo matching tasks, and 3D reconstruction demonstrates good visual effects.