Visual vehicle localization system for smart parking application

Abstract

With a vision of proposing a fully automated parking management solution for smart parking, in which all the operations in the process of parking will be automated. And as a first step we will focus on vehicle localization inside parking based on image processing theory. Video based localization algorithms present an important interest in the field of intelligent video surveillance, the integration of such functionality in the surveillance system will revolt their classic roles. Navigation tool and other amazing systems can easily build based on such feature. This paper describes an implementation of a FPGA based real-time visual system for vehicle localization. Vehicle in the video frames are extracted after the application of the background subtraction method on the input image using a background reference image. The dynamic threshold used is computed by the Otsu method. Finally, the object mask resulting from the segmentation process is used to compute the relative distance to the camera based on the relation between the ratio of the size of a vehicle on the camera sensor and its size in real life which is a function of the camera focal length and distance between the vehicle and the camera. The experimental results show that the proposed system is sufficiently satisfying the real time constraint (under the 100 MHz frequency a 32 frames per second is achieved for the 1440 * 1080 resolution, and under 50 MHz frequency a 41 frames per second is achieved for the 640 * 480 resolution) with an accuracy error around the centimeter level.