Zhao, Jiaqi (Beijing Jiaotong University), Jiang, Wei (Beijing Jiaotong University), Hu, Xiao (International Digital Economy Academy), Liao, Canjie (Beijing Jiaotong University), Cai, Baigen (Beijing Jiaotong University), Wang, Jian (Beijing Jiaotong University), Liu, Jiang (Beijing Jiaotong University), Ba, Xiaohui (Beijing JIaotong University), Lu, Debiao (Beijing Jiaotong University)

Depth Perception and Object Detection System of the Plateau Railway Scene Based on Binocular Vision

Scheduled for presentation during the Video Session "On-Demand Video Presentations" (VP-VP), Saturday, November 22, 2025, 08:00−18:00, On-Demand Platform

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on April 2, 2026

Abstract

GNSS/INS integrated navigation and positioning technology is the main monitoring method in extensive railway operations. However, the environment along the Qinghai-Tibet Railway is complex. Affected by high altitude, low air pressure, extreme climate and many mountainous areas, monitoring technology faces many limitations. Tunnels and mountains may block satellite signals. The sudden change in temperature causes a large temperature drift in the IMU and increases the systematic error. To solve this problem, this paper proposes a binocular vision system scheme with depth perception and target detection functions. In terms of hardware, a binocular camera module with large dynamic range characteristics is selected, and a rigid fixation and synchronous data acquisition method is adopted. Field tests show that the gauge measurement error of this system increases nonlinearly with the increase of the detection distance, and the average accuracy (mAP50) of target detection reaches 92%. The influence of train vibration on the external parameters of the camera is within 0.16 pixels, which proves the applicability of the system in high-altitude environments.