Development of Component Modules for an Integrated Video Surveillance System for EV Automatic Charging

This paper presents an integrated video surveillance system (IVSS) that enhances the safety, efficiency, and automation of electric vehicle (EV) charging processes. Utilizing advanced deep learning-based vision technology, the IVSS monitors and tracks EVs as they enter and park in designated charging slots, facilitating seamless interaction with an automatic charging system (ACS). Vehicle model classification and license plate recognition functions are developed to identify electric vehicles (EVs) in the IVSS. By utilizing pose estimation technology based on a perspective matrix, the IVSS estimates the pose of a parked EV and then detects its charging port. A collision risk prediction function between pedestrians and an automatic charging robot is developed using multi-object detection (MOD) and tracking (MOT) based on AI vision technology. The system functionality and performance to be developed are verified based on actual electric vehicle automatic charging environments. Experiments and analyses are conducted based on individual datasets for each function. These experiments demonstrate the IVSS's effectiveness in addressing key challenges in EV charging automation, showing potential for practical application while considering implementation costs and system adaptability.