Intelligent Research Based on Deep Learning Recognition Method in Vehicle-Road Cooperative Information Interaction...

Area of Science:

• Intelligent transportation systems
• Computer vision
• Machine learning

Background:

• Vehicle-road collaborative information interaction systems facilitate data sharing for enhanced urban transport.
• These systems are crucial for improving urban transportation infrastructure and economic growth.
• Intelligent research is needed to optimize recognition methods within these systems.

Purpose of the Study:

• To investigate deep learning recognition methods for vehicle-road collaborative information systems.
• To analyze the components, functions, and applications of these systems.
• To compare the performance of deep learning algorithms against traditional methods.

Main Methods:

• Exploration of vehicle-road collaborative information system concepts and structures.
• Implementation and analysis of three deep learning recognition methods: background extraction, YOLOv2, and DeepSORT.
• Comparative simulation experiments evaluating deep learning versus traditional algorithms.

Main Results:

• Deep learning methods achieved an 8.66% higher intersection ratio for vehicle target detection.
• Recall rates for vehicle target detection were 7% higher with deep learning.
• Vehicle flow recognition accuracy improved by 1.8%, and emergency decision-making showed reduced early warning times.

Conclusions:

• Deep learning algorithms demonstrate significant superiority and feasibility in vehicle-road collaborative information systems.
• These methods offer enhanced performance in vehicle detection, flow identification, and emergency response.
• The findings support the adoption of deep learning for advanced intelligent transportation solutions.