Study On Lane Detection Based On Machine Vision

Visual information is the main information can be obtained from the traffic environment. Therefore, machine vision technology is an indispensable part of IVS (Intelligent Vehicle System), and provides a huge potential market for application. By using computer vision to extract and recognize the traffic environment, the results of recognition can be used directly for auxiliary pilot or automatic pilot.At the beginning of this paper, it firstly introduce the development of intelligence automobile oversea and domestic. At the same time, some oversea companies and research institutions that are related to the intelligence automobile are listed to show their products of intelligence automobile. Then some applications of machine vision are discussed in the orientation assistance navigation system, the self-adapted cruise control system, the self-moving assistance system, the traffic signal identification system, the assistance change-carriageway, etc. At the end of this part, we describe the development of machine vision in the 1980 years.The paper discussed several kind of commonly image enhancement algorithm with small complex and small numeration to the path picture processing effect. Through the experimental result comparison, discovered gathers the picture from the path recognition system, uses the histogram balanced transformation the way processing picture effect to be better, for the certain degree influence regarding the path recognition under the sleet condition sleet pellet to use the spiced salt noise to carryon the simulation to it, The discovery uses two times of one dimensional center values filters the effect to be better, may obtain with the two-dimensional center value filter close result.The analysis and compared with 6 kind of classical derivative operators to examine the edge operator: Roberts operator, Sobel operator, Prewitt operator, Laplacian operator, Canny operator and LoG operator. Draws the conclusion is, on the computation quantity and computation complex, the LoG operator and the Canny operator which is more complex than that are all compared to the front several first orders derivative operator, and the computation quantity is also bigger. Compared with the path picture edge examination effect which may see the Sobel operator not only the computation quantity slightly but also simple, under the noise condition picture examination effect is all quite satisfying and is quite been small the noise influence, the edge of examination result is quite continual, localization also quite accurate, sentences the spot to be less by mistake.The paper proposed the structure path picture model, uses the path edge search algorithm, carries on to the path picture edge with draws, the structure path examina--tion and the recognition algorithm which uses the performance forecast quadrangle regarding the vehicle route search to examine the region, then used the robustness to search characteristic point the fitting a straight line method. Thus distinguishes the boundary of the road picture .Because the algorithm is based on the structure characteristic and the statistical characteristic, therefore has a stronger robustness in the practical application.