文档介绍：基于动态阈值的边沿检测法在赛道识别中的应用
摘要:在智能车比赛中,在智能车比赛中,使用摄像头传感器来采集路面信息,可以获取大量的图像信息,这样需要好的算法滤除画面中的噪点,并正确的提取引导线。对图像信息处理结果的好坏直接影响到赛车控制。采用固定阈值二值化方法进行中心线寻找,缺点是寻找到的黑线长度会明显变小,前瞻性降低。本文提出了防串线识别方法可以正确提取引导线起始行中心,并且与动态阈值的边沿检测方法结合提取出黑色引导线。实验证明该方法的可靠性比固定阈值二值化方法高,具有一定的使用价值。
关键词:动态阈值,边沿检测,智能车寻迹
Edge detection based on Dynamic threshold to the Smart car track identification
Abstract:
In the smart petition, using CMOS sensors to collect the road information, you can access a large number of image information. There must be a good algorithm to filter out the noise and extract the correct guide line. The results of image processing will directly impact on the smart car control. The drawback of the simple fixed threshold binarization method to find the centerline ,the disadvantages is the length of the centerline will obviously be smaller, and also reduce the forward-looking. In this paper, the anti-string line identification method can accurately extract the starting line center of the guide line, and with the dynamic threshold method of edge detection to extract the black guide lines. Experiments prove thatthe reliability of this method is high than the fixed threshold Binarization method, it has a certain Practical value.
Key words: dynamic threshold, edge detection, smart car tracing.
1、引言
由于摄像头架设高度比较低,约为29厘米。摄像头采集到的图像在最近处黑色引导线和白色基板的灰度值相差比较大(一般大于15),而画面较远端处的白点和黑点的灰度值相差比较小(一般小于10),几乎很难分辨,这给黑线提取的前瞻性带来