Label Correction for Road Segmentation Using Road-side Cameras

TL;DR

This paper presents a semi-automatic annotation method utilizing roadside cameras to improve road segmentation models across diverse weather conditions, reducing manual labeling effort and enhancing robustness.

Contribution

It introduces a novel semi-automatic annotation technique that transfers labels across frames using frequency domain registration, validated on extensive roadside camera data.

Findings

Segmentation performance improved with semi-automatic labels

Method effective across different weather conditions

Enhanced robustness demonstrated on multiple datasets

Abstract

Reliable road segmentation in all weather conditions is critical for intelligent transportation applications, autonomous vehicles and advanced driver's assistance systems. For robust performance, all weather conditions should be included in the training data of deep learning-based perception models. However, collecting and annotating such a dataset requires extensive resources. In this paper, existing roadside camera infrastructure is utilized for collecting road data in varying weather conditions automatically. Additionally, a novel semi-automatic annotation method for roadside cameras is proposed. For each camera, only one frame is labeled manually and then the label is transferred to other frames of that camera feed. The small camera movements between frames are compensated using frequency domain image registration. The proposed method is validated with roadside camera data collected from 927 cameras across Finland over 4 month time period during winter. Training on the semi-automatically labeled data boosted the segmentation performance of several deep learning segmentation models. Testing was carried out on two different datasets to evaluate the robustness of the resulting models. These datasets were an in-domain roadside camera dataset and out-of-domain dataset captured with a vehicle on-board camera.