十字线激光图像的快速边缘检测方法

A fast edge detection method for cross laser image

针对工程应用中十字线激光图像存在较大干扰噪声和要求图像处理速度快的特点,提出了一个快速准确的边缘检测方法。在图像处理时先找出图像中的有用区域,采用基于局域亮度最大的方法在有用区域内找出像素级的边缘点,对这些像素级的边缘点进行亚像素处理得到更高角度的边缘点,用最小二乘法计算出两条直线,最后计算出交点坐标和水平倾角。用640×480像素摄像机(光靶分辨率为0.0627mm/像素)在4.3m摄取50幅图像进行处理,标准差x为0.0119mm,y为0.0235mm,水平倾角为1.89′。方法应用于导轨直线度和扭曲度检测中,取得了很好的效果。

In view of strong noise in cross laser images and requirement for fast processing in industrial application, a fast edge de- tection method for cross laser image is presented. During the image processing, a serviceable area is firstly located, pixel edge points are then detected in the serviceable area by maximum luminance in a local region, these pixel edge points are processed to get sub- pixel edge points which are calculated to get the two lines of the cross by least square method, so the coordinate of the intersection point and the angle. The standard deviation is 0.0119mm for x, 0.0235mm for y and 1.89＇ for angle, with a camera （640x480 pixels and 0.0627mm/pixel as resolution of the target） at distance of 4.3m and 50 samples. This method was used in straightness and twist inspection of rail guides and got satisfied results.