| The Applications of the image processing technique on vehicle driver assistance systems can make driving safer. It is the main function that vehicle driver assistance system can detect obstacles. In this thesis, various methods of image processing were made use of and put forward. This thesis mainly deals with background withdrawing and obstacles detecting, mathematic way of obtaining the distance and relative velocity of the obstacles.Firstly, this thesis applies one method of median filter algorithm to denoise the noise in the precondition. Secondly, in this thesis, a method of obtaining partial average value to withdraw background was put forward. The method applies a general knowledge that there exists certain background area in front of the vehicle. After the background withdrawing, the minimum diplo-multiply method was applied to match line and get the current lane which the driving vehicle is in.For the purpose of detecting obstacles in the image, this author puts forward a method of obtaining foreground pixel average value of an object in a partial window. If the obtained average value above a threshold, the object can be identified as an obstacle.In the part of measuring distance of obstacles, a existed method is used to reconstruct driveway line, the method was put forward by Zhou Xin, Huang Xi-Yue and Li Yu .The aim of realizing highway obstacle detection system is providing assistance for driving, the main function is detecting obstacles and alarming as soon as obstacles approach the dangerous distance. The two parameter based on to give alarm is the distance and the relative velocity of obstacles.For obtaining the relative velocity of an obstacle, we must track it. In this thesis, amethod of making use of tracking window was put forward and the method be used to track obstacles.Finally, the function of detecting obstacles was realized. The result of experimentation shows that the system that was designed has the properties of real time and robust.
|
PDF Full Text Request