复杂背景下高层建筑钢结构板孔心影像坐标量测

Measurement of image coordinates of hole center in steel structure plate under complex background

针对复杂背景下钻有圆孔的钢结构板图像,为实现圆孔中心的精确定位,根据近景摄影测量中二维直接线性变换的方法确定各圆孔对应的感兴趣区域,并分别对其进行二值化、形态学等处理,之后采用Canny算子检测圆孔像素级边缘,同时根据边缘轮廓所占面积作为条件,剔除错误边缘;为提高边缘点的定位精度,在像素级边缘点的基础上使用Zernike矩获取圆孔的亚像素级边缘点,最后对边缘点集用最小二乘的方法进行多次椭圆拟合,从而确定圆孔中心在图像中的精确位置.试验结果表明,该方法可以获得较高的孔心影像坐标量测精度,点位"均方差"达到0.182mm.

This paper concerns about a method to precisely position the center of circular holes on steel plates structure images which were taken under complex background.In this algorithm,the method of 2dimension direct linear transformation is adopted to locate the region of interests（ROI）of related circular holes.After handling the ROI of image with binaryzation and morphological processing,we detect the edge of the circular holes with Canny edge detector on pixel-level and eliminate wrong edges according to the area it takes up.In order to improve the accuracy of positioning,Zernike moment is applied to obtain subpixel level edge based on the pixel level edge.Finally,the image coordinates of each hole center are localized precisely by ellipse best fitting.The test results show that the algorithm proposed can achieve high accuracy of measuring center of circular holes,with mean square error of 0.182 mm.