Analysis of Deep Learning-Based Colorization and Super-Resolution Techniques for Lidar Imagery

TL;DR

This paper reviews deep learning methods for enhancing lidar images through colorization and super-resolution, highlighting their potential for improving autonomous system tasks under challenging conditions.

Contribution

It provides a comprehensive review and qualitative analysis of DL-based colorization and super-resolution techniques for lidar imagery, including computational performance assessment.

Findings

Lidar images are robust under low-light and foggy conditions.

DL-based super-resolution improves image resolution and quality.

Colorization enhances interpretability of lidar images.

Abstract

Modern lidar systems can produce not only dense point clouds but also 360 degrees low-resolution images. This advancement facilitates the application of deep learning (DL) techniques initially developed for conventional RGB cameras and simplifies fusion of point cloud data and images without complex processes like lidar-camera calibration. Compared to RGB images from traditional cameras, lidar-generated images show greater robustness under low-light and harsh conditions, such as foggy weather. However, these images typically have lower resolution and often appear overly dark. While various studies have explored DL-based computer vision tasks such as object detection, segmentation, and keypoint detection on lidar imagery, other potentially valuable techniques remain underexplored. This paper provides a comprehensive review and qualitative analysis of DL-based colorization and super-resolution methods applied to lidar imagery. Additionally, we assess the computational performance of these approaches, offering insights into their suitability for downstream robotic and autonomous system applications like odometry and 3D reconstruction.