Research on collision prevention early warning technology of parking lot vehicles

The collision prevention early warning technology is critical in vehicle safety field because early warning technology has a significant influence on the vehicle reliability. However, collision prevention early warning is a challenging problem due to different special scenes. For the scene of parking lot, this study proposes the early warning method based on vision. First, the working condition of the parking lot is introduced, mainly including the perceivable working condition and imperceivable working condition. Next, the improved YOLOv5 algorithm is used to detect vehicles in the perceivable working condition through binocular camera, while an imperceivable collision avoidance safety distance model is designed to avoid collision in advance in the imperceivable working condition. The results of the experiment and simulation have a satisfying performance.