A New Roadway Eventual Obstacle Detection System Based on Computer Vision

Area of Science:

• Computer Vision
• Road Safety Engineering
• Artificial Intelligence

Background:

• Road accidents due to unexpected obstacles, particularly wildlife, have increased significantly.
• Existing detection systems may be costly or lack robustness in varied conditions.
• There is a need for advanced, accessible systems for real-time obstacle detection.

Purpose of the Study:

• To develop and evaluate a low-cost computer vision system for detecting and classifying moving road obstacles.
• To enhance road safety by providing early warnings for potential hazards.
• To assess the system's performance under diverse environmental conditions.

Main Methods:

• Utilized low-cost hardware and open-source software for system development.
• Employed infra-red and colour video imaging for data input.
• Implemented computer vision algorithms for detecting and classifying moving elements.
• Conducted experimental evaluations under various weather and illumination conditions.

Main Results:

• The system successfully detected and classified different types of moving road obstacles.
• Demonstrated robust performance across a range of weather and lighting scenarios.
• Validated the effectiveness of the proposed computer vision approach.

Conclusions:

• The developed system offers a viable and cost-effective solution for roadway obstacle detection.
• This technology has the potential to significantly improve road safety and reduce accidents.
• The system's robustness makes it suitable for real-world deployment in diverse environments.