Research And Implementation Of Parking Assist Technology Based On Digital Image

With the number of motor vehicles increasing, it is more difficult to find unoccupied parking space, especially in those large cities, this question distresses the majority of drivers. In cities, drivers cannot find unoccupied parking space easily, and when they backing their cars into the parking spaces they usually cannot avoid crashing other cars, especially for those beginners. So, intelligent parking assist systems are very necessary for cars. After years of development, parking assist system has changed for several generations. Now, the universal parking assist systems consist of machine vision components and ultrasonic components, and it can be predicted that this combination will continue to be the main trend of industry in the next period of time.The system introduced in this article also consists of machine vision components and ultrasonic components, and the article mainly completes the following several works. Firstly, it preprocesses images by using regular steps such as graying of the image, image denoising, edge detecting and binary etc. After preprocessing images, it gets binary images, then it can use a custom template to filter binary images in order to clear away most background noise. Then, it can use a modified contour extraction algorithm to extract parking contour. Secondly, using modified Hough transform algorithm detects parking lines, then drawing straight lines that parking lines belong to, based on these straight lines, it can find the parking entry points. Thirdly, for several usual parking space models, it can formulate corresponding backing lines in order to make the system be suitable for most parking spaces. Finally, for the case of idle parking space being obstructed, it can use machine vision technology and ultrasonic technology to detect idle parking space.The experimental results show that the algorithms which related to the intelligent parking assist system introduced in this article can better detect parking lines and can better extract parking entry points, it also can draw backing lines reasonably by using these algorithms. Then, using machine vision technology and ultrasonic technology can also detect the parking spaces which are obscured. So, the algorithms not only have better robustness and efficiency, but also have better adaptability.