Research And Implementation Of Parking Slot Markings Recognition System Based On Vision

With china’s car market is gradually mature, more and more people will ascend to the security as a preferred factor in vehicle purchases. So equipped with parking assistance system has become the development trend of automobile market, And in parking assistance system, parking guidance function more and more get the welcome of the market, so, it has the very strong application value. Parking detection is the key technology of automatic parking, there has two fashion parking technologies in market currently, one kind is based on ultrasonic parking detection method, and one kind is based on visual parking methods. Compared with the ultrasonic parking detection method, the visual parking detection method is more reliable, safer and cheaper costs,The most important is based on visual parking detection system can detect more complex scenes. Therefore parking detection system based on visual becomes a research hotspot in recent years.This paper research for parking detection system based on the rear fisheye camera. First, we research a series of method which may used on the system, at the same time we verifies the effectiveness of these methods, and finally achieve a Parking slot markings recognition system based on version. Based on the system construction sequence, this paper can divided into image pre-processing module, edge detection module, Hough transformation and line extraction module, the parking line detection module, parking slot generation module, parking slot tracking module. For each module we find the algorithms for the problems need to be solved and we explained the detail of implementation process.Finally, this paper gives the system test in various complex scenes, including the rain and snow conditions, Results show that in these complex scenes, the technology shows in this paper can effectively detect the parking slot. Therefore, the technology shows in this paper is effective.