Converting Depth Images and Point Clouds for Feature-based Pose Estimation
This addresses the problem of bridging depth data with classical computer vision for robotic systems, though it appears incremental as an enhancement over existing conversion methods like Bearing Angle images.
The paper tackles the challenge of using raw depth data for sensor registration by converting depth images/point clouds into Flexion images that visualize spatial details hidden in traditional representations. The method yields brighter, higher-contrast images with more visible contours, improving feature-based pose estimation in visual odometry and RGB-D SLAM tasks for all tested features (AKAZE, ORB, SIFT, SURF) compared to Bearing Angle images.
In recent years, depth sensors have become more and more affordable and have found their way into a growing amount of robotic systems. However, mono- or multi-modal sensor registration, often a necessary step for further processing, faces many challenges on raw depth images or point clouds. This paper presents a method of converting depth data into images capable of visualizing spatial details that are basically hidden in traditional depth images. After noise removal, a neighborhood of points forms two normal vectors whose difference is encoded into this new conversion. Compared to Bearing Angle images, our method yields brighter, higher-contrast images with more visible contours and more details. We tested feature-based pose estimation of both conversions in a visual odometry task and RGB-D SLAM. For all tested features, AKAZE, ORB, SIFT, and SURF, our new Flexion images yield better results than Bearing Angle images and show great potential to bridge the gap between depth data and classical computer vision. Source code is available here: https://rlsch.github.io/depth-flexion-conversion.