TexLiDAR: Automated Text Understanding for Panoramic LiDAR Data

TL;DR

This paper introduces TexLiDAR, a novel method that leverages panoramic 2D images from advanced LiDAR sensors and large pre-trained models to improve text understanding and object detection in LiDAR data, bypassing the limitations of 3D point cloud processing.

Contribution

The paper proposes using 2D panoramic images from LiDAR sensors with large models like Florence 2 for zero-shot captioning and detection, offering a new approach to LiDAR-text integration.

Findings

Florence 2 produces more informative captions than existing methods.

The approach achieves superior object detection performance.

It enables real-time, high-accuracy detection in challenging scenarios.

Abstract

Efforts to connect LiDAR data with text, such as LidarCLIP, have primarily focused on embedding 3D point clouds into CLIP text-image space. However, these approaches rely on 3D point clouds, which present challenges in encoding efficiency and neural network processing. With the advent of advanced LiDAR sensors like Ouster OS1, which, in addition to 3D point clouds, produce fixed resolution depth, signal, and ambient panoramic 2D images, new opportunities emerge for LiDAR based tasks. In this work, we propose an alternative approach to connect LiDAR data with text by leveraging 2D imagery generated by the OS1 sensor instead of 3D point clouds. Using the Florence 2 large model in a zero-shot setting, we perform image captioning and object detection. Our experiments demonstrate that Florence 2 generates more informative captions and achieves superior performance in object detection tasks compared to existing methods like CLIP. By combining advanced LiDAR sensor data with a large pre-trained model, our approach provides a robust and accurate solution for challenging detection scenarios, including real-time applications requiring high accuracy and robustness.