Track line detection method and system

Abstract

The invention discloses a track line detection method and a track line detection system. The detection method comprises the following steps: acquiring an image of a to-be-detected track line; determining a starting point position of a to-be-detected track line by utilizing an SSIM similarity algorithm according to the image; adopting a windowing detection mode to obtain edge data of a track line in each window in the image; performing Hough transform on the edge data of the track line in each window to obtain a track line corresponding to each window; and performing polynomial fitting on the track lines corresponding to all the windows to obtain a complete track line. The starting point position of the track line can be accurately detected, detection of the track line is achieved, and meanwhile the accuracy of track line detection is improved.

Description

technical field [0001] The invention relates to the field of track detection, in particular to a track line detection method and system. Background technique [0002] Lane line detection is mainly to detect white / yellow solid lines, double yellow lines, etc. on black asphalt or gray cement roads, while track line detection is mainly to detect railway track lines. [0003] The general method for determining the starting point of the traditional lane line is as follows: [0004] Most of the existing lane detection methods use a single visual clue to extract the features of the lane. The direction feature of the lane and the color of the lane show a dark-bright-dark color feature relative to the color of the road surface, and some are directly based on white and yellow. Color filtering is used to determine the lane line and the starting point of the lane line. This method requires a significant change in the color gamut, or requires trace lines of a specific color to be disti...

AI Technical Summary

Problems solved by technology

This method is extremely susceptible to interference from the background environment, and tracking errors are not suitable for track line detection

[0007] Therefore, because the track line background is more complex than the track line, usually there are many disturbing objects such as sleepers, stones, power supply lines and communication equipment around the track line, the traditional lane line detection methods are not suitable for track line detection. The lane line processing method cannot accurately find the starting point of the track line

Images