Deterministic Guided LiDAR Depth Map Completion

TL;DR

A non-deep learning-based approach to densify a sparse LiDAR-based depth map using a guidance RGB image and the pinhole camera model is used for the interpolation process, which shows that it outperforms the state-of-the-art non- Deep Learning-based methods, in addition to several deep learning- based methods.

Abstract

Accurate dense depth estimation is crucial for autonomous vehicles to analyze their environment. This paper presents a non-deep learning-based approach to densify a sparse LiDAR-based depth map using a guidance RGB image. To achieve this goal the RGB image is at first cleared from most of the camera-LiDAR misalignment artifacts. Afterward, it is over segmented and a plane for each superpixel is approximated. In the case a superpixel is not well represented by a plane, a plane is approximated for a convex hull of the most inlier. Finally, the pinhole camera model is used for the interpolation process and the remaining areas are interpolated. The evaluation of this work is executed using the KITTI depth completion benchmark, which validates the proposed work and shows that it outperforms the state-of-the-art non-deep learning-based methods, in addition to several deep learning-based methods.

Figures (8)

• Figure 1: a) Fusion of RGB-image and sparse depth map to obtain a denser depth map; b) Colorized sparse pointcloud; c) Colorized pointcloud from a dense depth map
• Figure 2: (Top) Color fringing; (Bottom) Misalignment artifact
• Figure 3: Interpolation method: Seeking the intersection of ray and plane
• Figure 4: Interpolation loss vs orthogonal loss
• Figure 5: Interpolation error function. The more parallel the plane is, the larger the resulting error is.