The ways of streamlining digital image processing algorithms used for detection of lines in transport scenes video recording

摘要

The paper is focused on describing the use of digital image processing methods, with emphasis on application in road transport, which warns the driver in case of unintentional lane departure. The problem of detecting lines is extensive and consists of a series of problems which partial results were published in other articles. The authors analyse the possible solutions to find the horizontal traffic signs using segmentation methods designed for finding lines in each video frame of the captured traffic scene on the highway. To speed up the process and reduce the volume of processed data it was necessary to find ways of improving the effectiveness of the line detection algorithm. These methods are described in the present paper. In the section describing the practical implementation we have shown the way of finding road markings using Sobel edge detector algorithm and modified Hough transform in the image data recorded in .AVI format file and from the position of the vehicle and lane found position to detect the crossing of lane.