Vision-based Approach in Finding Multitype Parking Stall Entrance

This paper proposes a novel method of finding the entrance of the parking stall for PAS to cope with multitype parking stalls while drying in the parking aisle. For this purpose, we propose to categorize parking stalls into two types and three shapes for each type. Utilizing the information along the parking lines surrounding the parking stall, the type and the location information of the parking stall can be obtained to guide the vehicle. Experimental results show that the proposed method can successfully detect various types of parking stalls enough to park a vehicle via autonomous valet parking.