Research And Implementation Of Lane Recognition Technology Based On Morphology And Distribution Feature

Car is a product with the development of human society, and promotes the progress of human society, and becomes an irreplaceable life transport. With the continuous development of the automobile industry and society progresses, the car in people’s lives has been a wide range of popular and plays an increasingly important role, but the safety of automobile travel is increasingly prominent. Faced with increasingly severe traffic safety situation, to improve vehicle safety and reduce traffic accidents has become the focus of attention of people, in which the development of prevention as the core of the intelligent driver assistance systems has become pressing needs.Lane is an important indicator to the motor cars, so the lane recognition has become an important function in the intelligent driver assistance systems. This paper researched the lane recognition technology by analyzing the lane characteristics of morphological and distribution. The main process of the system was as follows. First, to provide the necessary conditions for the subsequent processing and improve the accuracy of subsequent recognition, processed the image with a series of pretreatment, included grayscale, filtering, binarization, etc. Second, this paper took use of the natural boundary between road region and non-road region, to extract lane feature region by scanning the image from bottom to up in dichotomy. This method greatly improved the real-time of the system. Then, based on the specific characteristics of morphological and distribution, processed the feature region with filter, combined fitting and prediction, to improve the completeness of the algorithm and the accuracy of the identification. Finally, recognized the lane in the later image based on the information from the priori image and this method could reduce the scanning range and reduce interference information to improve the accuracy and the real-time of the tracking recognition.Experimental results show that the lane recognition algorithm based on the characteristics of morphological and distribution can recognize the lane well and can also process the video which has good robustness and high efficiency.