Lane Line Detection Method Based on Neighborhood Gray Scale Characteristics

In order to solve the problem that lane line detection algorithm is difficult to achieve a balance between the robustness and timeliness in the complex road traffic environment, put forward a kind of road image segmentation method based on neighborhood gray scale characteristics, to deep mining lane line contour information. On the basis of this, puts forward a lane line detection optimization algorithm based on sampling row bidirectional scan and imaging model constraint candidate feature points. Through the offline test for detection algorithm in data set that have many scenarios, the average precision is 96.02%, accuracy is 95.33%, average detection time about 31.07 milliseconds per frame, so it can basically meet the requirements of system for real-time. The experimental results show that the method can realize the accurate detection for lane line in a variety of road scene, has good robustness and real-time performance.