New Method For Lane Detection And Vehicle Offset Distance Calculation

The fast and accurate lane detection is an important component in the intelligent transportation system;it is not only widely used in commercial driving system but also introduced to military training.There are some drawbacks in current lane detection,for example,low efficiency,long time consuming and easily affected by light etc.In order to overcome these defects,this paper carries out abundant research and makes improvement in lane detection field.First of all,this paper studies current technology deeply and then improves detection procedures and algorithms.At the same time,it comes up with one calculation method of vehicle offset distance.In the end,it applies lane detection to military affairs and develops the training and assessment system of parade vehicle based on lane detection.The main research contents of this paper are briefly described as follows:(1)In the process of image pretreatment,this paper comes up with one new algorithm after researching the classic binarization algorithms and calls it mixed normal distribution method.It obtains segmentation threshold by searching the intersection of normal distribution density curves between the target and background.The segmentation threshold is self-adapting and can be used to segment real-time changing images.The results of the experiments show that we can receive excellent binarization effects and the new algorithm takes 27%less time approximately than the traditional algorithm does.(2)In the process of edge detection,this paper analyzes the classic algorithms and improves Canny edge detection algorithm.Based on Canny algorithm,the improved Canny edge detection algorithm adopts Gaussian distribution to get high threshold and low threshold.Meanwhile,it increases the times of Gaussian smooth to eliminate more background interference.The results of the experiments indicate that the new algorithm has better detection effects and it costs the same time as the traditional algorithm does.(3)This paper seeks out the offset distance calculation formula in the ordinary collection distance through analysis,and the offset distance represents the degree of vehicle deviating from the lane.The results of the experiments prove that the offset distance can be calculated accurately at different collection distances.(4)This paper develops the training and assessment system of parade vehicle based on lane detection.This system mainly includes three parts:detection,transmission and assessment;among them,detection is the core of the system.The system is performed field experiments at a military base,the results indicate that the system is stable and well-functional.The comprehensive experiments indicate that the new method can detect lanes quickly and accurately in the complicated external environment.The recognition rate is as high as 91.14 percent,the average time consuming of single-frame image is 34.5 milliseconds.In addition,the training and assessment system of parade vehicle based on lane detection improves the military training results and shortens the military training time,and is widely used in military parade.