Related Experiment Video
Updated: Jun 27, 2026
11:53
The Modular Design and Production of an Intelligent Robot Based on a Closed-Loop Control Strategy
Published on: October 14, 2017
Rail-BEV: A LiDAR-Centric and Sensor-Aware BEV Perception Framework for Long-Range Railway Obstacle Detection
Jinghan Huang1, Wentao Hu2, Zifeng He1
1School of Civil Engineering and Architecture, East China Jiaotong University, Nanchang 330013, China.
Sensors (Basel, Switzerland)
|June 26, 2026
Summary
Rail-BEV enhances railway safety by improving long-range obstacle detection using LiDAR and camera data. This sensor fusion approach achieves superior performance in identifying obstacles on train tracks.
Area of Science:
- Robotics and Autonomous Systems
- Computer Vision
- Transportation Engineering
Background:
- Reliable long-range onboard perception is critical for advanced railway safety systems.
- Challenges include long braking distances, sparse sensor returns, and constrained rail geometry.
- Existing systems require robust obstacle recognition for safe operation.
Purpose of the Study:
- To present Rail-BEV, a reproducible baseline for LiDAR-centric railway obstacle perception.
- To integrate geometric and visual sensor data for enhanced situational awareness.
- To establish a unified bird's-eye-view (BEV) framework for railway environments.
Main Methods:
- Utilized LiDAR as the primary geometric sensor and an RGB camera for auxiliary visual data.
- Developed a railway-oriented BEV backend with geometry-aware fusion and rail-geometry prediction.
- Incorporated lightweight inference-time structural refinement and ROI-based rail-corridor analysis.
Main Results:
- Rail-BEV achieved the highest overall mean Average Precision (mAP) of 0.6669, outperforming CenterPoint and BEVFusion baselines.
- Demonstrated significant improvements in long-range pedestrian detection.
- Ablation studies confirmed the benefits of visual assistance and rail-corridor refinement.
Conclusions:
- LiDAR-centered sensing, visual assistance, and rail-aware reasoning effectively support long-range railway obstacle perception.
- The Rail-BEV framework offers a reproducible baseline for future research.
- Identified limitations include rail-geometry quality, calibration, sensor degradation, and localization accuracy.
Related Concept Videos
01:30Electronic Distance Measuring Instruments
Electronic Distance Measuring Instruments (EDMs) are essential tools in modern surveying, offering precise distance measurements by emitting electromagnetic signals and calculating the time required for these signals to travel to a target and return. Two primary types of signals are used in EDMs — light waves and microwaves — each suited to specific environmental and distance requirements. Light-wave-based EDMs utilize either infrared or laser light, providing high accuracy over short distances...
01:29Sight Distance in a Vertical Curve
Sight distance on vertical curves is critical in roadway design. It ensures drivers can see far enough ahead to identify and respond to hazards effectively. This directly impacts safety, driver comfort, and the overall efficiency of the transportation network.Vertical curves are classified into crest and sag curves based on their geometry. For crest curves, sight distance is determined by the line of sight between a driver's eye and a small object on the road's surface. Design parameters for...
01:15Depth Perception and Spatial Vision
Depth perception is the ability to perceive objects three-dimensionally. It relies on two types of cues: binocular and monocular. Binocular cues depend on the combination of images from both eyes and how the eyes work together. Since the eyes are in slightly different positions, each eye captures a slightly different image. This disparity between images, known as binocular disparity, helps the brain interpret depth. When the brain compares these images, it determines the distance to an object.