Visual Parking Occupancy Detection Using Extended Contextual Image Information via a Multi-Branch Output ConvNeXt...

Area of Science:

• Computer Vision
• Artificial Intelligence
• Urban Planning

Background:

• Increasing urbanization and vehicle numbers create significant challenges in finding parking in metropolitan areas.
• Inefficient parking contributes to increased accidents, carbon footprint, and driver stress.
• Technological solutions for real-time parking management are crucial for urban mobility.

Purpose of the Study:

• To propose a novel computer-vision-based system for detecting vacant parking spaces.
• To enhance the accuracy and robustness of parking detection in challenging environmental conditions.
• To improve the efficiency of the urban parking process through advanced technology.

Main Methods:

• Development of a new deep-learning algorithm based on a multi-branch output neural network.
• Utilizing color imagery for parking space occupancy inference.
• Maximizing contextual image information by considering the entire input image for each parking slot.

Main Results:

• The proposed system demonstrates high robustness to varying illumination, camera perspectives, and occlusions.
• The multi-branch neural network effectively infers parking space occupancy.
• Extensive evaluations on public datasets show superior performance compared to existing approaches.

Conclusions:

• The novel deep-learning system offers a robust and accurate solution for real-time vacant parking detection.
• This technology can significantly improve urban parking management and driver experience.
• The approach advances the field of computer vision for intelligent transportation systems.