Thermal Detection of People With Mobility Restrictions for Barrier Reduction at Traffic Lights Controlled Intersections

Rapid advances in deep learning for computer vision have driven the adoption of RGB camera-based adaptive traffic light systems. These systems improve traffic safety and pedestrian comfort. However, they often overlook the needs of people with mobility restrictions. Moreover, RGB camera–based detection is vulnerable to adverse weather and raises privacy concerns. To address these issues, we propose a fully automated traffic light system based on thermal detectors. It dynamically adjusts signal durations for individuals with walking impairments or mobility burden. It also triggers auditory signals for visually impaired individuals. Together, these features help create barrier-free intersections for all users. To this end, we build a thermal dataset for people with mobility restrictions (TD4PWMR). It captures diverse pedestrian scenarios, particularly focusing on individuals with mobility aids or mobility burdens. It includes various environmental conditions, such as different lighting, weather, and crowded urban settings. While thermal imaging offers advantages in terms of privacy and robustness to adverse conditions, it also presents challenges for object detection. Thermal images lack color and fine texture details. They also typically have lower resolution than RGB images. To overcome these limitations, we develop YOLO-Thermal, a novel variant of the YOLO architecture. It integrates advanced feature extraction and attention mechanisms. These improvements enhance detection accuracy and robustness in thermal imaging. Experiments demonstrate that the proposed thermal detector outperforms existing detectors, while the proposed traffic light system effectively enhances barrier-free intersections. The source codes and dataset are available at https://github.com/leon2014dresden/YOLO-THERMAL.