Study On The Algorithm Of Lane Mark Identifying And Tracking In Intense Illumination

Safety is always one of the most important topics in vehicle engineering. Inrecent years with the rapid development of transportation, traffic accidents ofhighway, especially the malignant traffic accident, are greatly increasing. Underthis circumstance, Safety Driving Assist technologies, as a part of IntelligentVehicle technologies, are paid more and more attention. Safety Driving AssistSystem uses the information from Intelligent Apperceiving System and makesdecision and programming. Then the system gives advice to the driver or replacesthe driver to control the vehicle partially. Some developed countries, such as theU.S., U.K., Japan, Germany etc., have made their research programs to developSafety Driving Assist technologies and have already achieved great progress.Lane Tracking System is one part of Safety Driving Assist System. Somecountries have developed successfully some Lane Identifying and Tracking Systembased on machine vision. The purpose is to supply basic to self-guiding vehicle,vehicle road off and so on in the freeway or other Structural Road. Lane TrackingSystem works out of the door, so it is often influenced by variable illumination.Most of the vehicles' running time is in daytime, and most of the weather is fine.So the lane mark identifying and tracking in intense illumination is a problem thatmust be solved.The title of this paper is "Study on the Algorithm of Lane Mark Identifyingand Tracking in Intense Illumination". The paper includes six parts: 1. Study on thealgorithm of image contrast enhancement in intense illumination, 2. Study on thealgorithm of edge enhancement of lane image, 3. the algorithm of imagesegmentation based on 2-D gray level histogram, 4. Lane mark identifying basedon Hough Transportation, 5. Lane mark tracking based on the fusion of AOI andKalman filter algorithm, 6. The experiment of lane mark tracking in intenseillumination.The gray level is high and the contrast is low in intense illumination image. Sothe paper uses the algorithm of histogram cone-shaped extending to enhance imagecontrast before image segmentation. The paper also attempts the algorithm oflinearization of gray-level histogram. Through comparing the two algorithms, thefirst algorithm has better effect, and it cost less time. So the paper chooses theformer algorithm. Experiments show that the image contrast enhancement inevidence after using the algorithm processing the image. And it offers a good basicto follow image procession.In order to identify accurately, the paper enhances the edge for image. It alsocan avoid large white area tampering with identifying to enhance edge. The lanemarks are two biases in the image, so the paper uses 45 orientation PrewittOperator processing the right half image, and uses 135 orientation PrewittOperator processing the left half image. The advantage of this algorithm is that theeffect is very good, and it cost a short time.The paper uses the algorithm of 2-D gray histogram to binarization image.Firstly, the paper makes the threshold to segment the obvious background andtarget partly. Then the paper makes anther threshold to segment the other pelses.Experiments show that the algorithm can segment the image exactly.This paper establishes lane restriction before identifying the lane mark. And ituses Hough transformation to identify the lane mark. Most of the lane marks arediscontinuous, and sometimes they may be smudgy, blurry, or sheltered by othervehicles. The paper adopts Hough transformation to identify the lane mark, becauseit has ability to avoid noise and error. In order to improve the veracity and thereliability of identifying, the paper identifies the edge of the lane mark, and then itgets the midline of the two edges as the identifying result.When tracking the lane mark, this paper uses the algorithm which fuses AIOand Kalman filter. In order to exert the advantages of AIO and Kalman filter, thepaper fuses the two algorithms, which improves the real time ability and reliability.Experiments show that the algorithm has good real time ability and reliability.The software application system is developed using Visual C++6.0. Thearithmetic methodologies have been tested and reached the prospective targets, andthe availability has been probed.