Design and Evaluation of Parking Position Detection with Human Cooperation for Automatic Parking

Abstract

Parking an automobile is a troublesome act as it requires a certain amount of driving skill. In recent years, a parking assist system has been partially implemented. This conventional system mainly recognizes a parking space surrounded by a lane. However, this conventional system is merely a driving support system and supports safe driving based on the driving technique of a driver. An automatic parking system has been studied previously. However, this system depends on the development of infrastructure, such as cameras and sensors or communication technology, in parking lots to detect parking space. This paper suggests a novel automatic parking system in which drivers indicate a parking space from the outside of a vehicle and assist the vehicle in recognizing its environment so that the vehicle can park semi-automatically. This novel system is another approach to park an automobile with human cooperation. Specifically, the driver parks by using an onboard camera and an indication device that indicates a parking space. An onboard camera detects the indication device; therefore, a vehicle detects the parking space. The method to detect the indication device is based on HSV-colored conversion and Hough transformation. The distance between the parking space and the vehicle is calculated based on the onboard camera. Assuming real parking spaces, we evaluated the detection of a parking space by a vehicle by using the proposed method. These results show su ffi cient detection to make parking possible and to implement the proposed system.