基于PCA和随机Hough变换的调车信号灯检测

针对传统信号灯检测方法中存在的弊端,提出了一种主成分分析(Principal Component Analysis,PCA)和圆拟合与分段随机Hough变换(Randomized Hough Transform,RHT)圆检测相结合的信号灯检测方法。该方法利用主成分分析结合圆拟合来简化图像的复杂背景,然后通过随机Hough变换进行分段圆检测,得到信号灯的准确位置,最后结合线路特征及信号灯的颜色特点,确定本股道有显示的调车信号灯。对调车信号灯图像进行仿真实验,结果表明,该方法检测准确度高,实时性好,并对有颜色信号灯的检测具有良好的检测效果,适合于对复杂环境中调车信号灯的准确检测和定位。

Shunting Signal Lamp Detection Based on PCA and Randomized Hough Transform

In this paper,to overcome the disadvantages of traditional methods of signal lamp detection,a new method of signal lamp detection was proposed,which is based on principal component analysis( PCA) and circle fitting,as well as segmented circle detection of randomized Hough transform( RHT). In this method,the principal component analysis and circle fitting should be carried out to simplify the complex background of image first,and then segmented circle detection should be conducted by the use of randomized Hough transform so as to ascertain the accurate position of signal lamp; finally incombination with line's characteristics and signal lamp's color,the displayed shunting signal lamps on this line can be determined. Moreover,a simulation experiment on the images of shunting signal lamp was carried out in this paper,and the results show that this method is of high accuracy,good real-time performance,good detection effect on the colored signal lamps,therefore suitable for accurate detection and positioning of shunting signal lamps in complex environment.