摘要
为实现对接触网定位器坡度的动态测量,提出基于计算机视觉技术的定位器坡度的动态检测算法。采用中值滤波算法和维纳滤波算法对拍摄到的图像序列进行图像预处理;对序列图像中的接触网定位器进行提取识别以及角点检测;选用NCC角点匹配算法对检测序列图像中检测到的角点进行匹配;在匹配角点的基础上,利用基于仿射不变性的直线匹配算法,对各幅图像中的直线进行匹配,从而可以自动获取图像序列中各图像直线的夹角,实现定位器坡度的自动动态测量并且减小了测量结果的不确定性。
To realize dynamic detection of slope gradient of catenary locator, the dynamic detection algorithm of slope gradient of locator based on computer vision technology is proposed in this article. The image preprocessing for image sequences is implemented by the median filtering algorithm and Wiener filtering algorithm; The catenary locator in the image sequence is extracted and recognized; The NCC corner matching algorithm is used for the detected corners matching in image testing sequences; On the basis of corner matching, the straight lines in the images are matched by affine invariant straight match algorithm. Then the straight line angle in the images can be automatically obtained, which can realize the locator slope＇s automatic dynamic detection and can reduce the uncertainty of the detection results.
出处
《铁道标准设计》
北大核心
2013年第1期105-108,126,共5页
Railway Standard Design
关键词
接触网
定位器
坡度
计算机视觉
动态测量
算法
catenary
locator
slope gradient
computer vision
dynamic detection
algorithm
