Annotation-Free Detection of Drivable Areas and Curbs Leveraging LiDAR Point Cloud Maps

2026-03-29

AI Summary

This work addresses the high cost and vulnerability to occlusion and sparsity inherent in existing deep neural network–based approaches for drivable area and curb detection, which rely heavily on manual annotations. To overcome these limitations, the authors propose a Map-based Automatic Data Labeling (MADL) framework that leverages LiDAR mapping and localization to generate high-quality training labels without human intervention. A data curation agent is integrated to filter out low-quality samples, enabling an end-to-end, annotation-free training pipeline. By exploiting multi-frame fused LiDAR maps, MADL mitigates the constraints of single-frame point clouds for the first time. Extensive experiments on KITTI, KITTI-CARLA, and 3D-Curb benchmarks demonstrate that the proposed method significantly outperforms both existing self-supervised and manually annotated approaches, achieving superior accuracy and robustness.

Abstract

Drivable areas and curbs are critical traffic elements for autonomous driving, forming essential components of the vehicle visual perception system and ensuring driving safety. Deep neural networks (DNNs) have significantly improved perception performance for drivable area and curb detection, but most DNN-based methods rely on large manually labeled datasets, which are costly, time-consuming, and expert-dependent, limiting their real-world application. Thus, we developed an automated training data generation module. Our previous work generated training labels using single-frame LiDAR and RGB data, suffering from occlusion and distant point cloud sparsity. In this paper, we propose a novel map-based automatic data labeler (MADL) module, combining LiDAR mapping/localization with curb detection to automatically generate training data for both tasks. MADL avoids occlusion and point cloud sparsity issues via LiDAR mapping, creating accurate large-scale datasets for DNN training. In addition, we construct a data review agent to filter the data generated by the MADL module, eliminating low-quality samples. Experiments on the KITTI, KITTI-CARLA and 3D-Curb datasets show that MADL achieves impressive performance compared to manual labeling, and outperforms traditional and state-of-the-art self-supervised methods in robustness and accuracy.

Problem

• drivable area detection
• curb detection
• annotation-free
• LiDAR point cloud
• autonomous driving

Innovation

• annotation-free
• LiDAR mapping
• automatic data labeling
• drivable area detection
• curb detection