摘要
为了能够快速有效地检测精密样板的廓形,基于Canny理论提出了一种边缘提取的方法,该方法在保持Canny算法优点的同时,通过修正高斯边缘滤波函数提高了边缘的定位精度,通过使用Sobel算子计算梯度消除了局部噪声的影响,通过使用一维线性插值计算梯度方向上点的梯度幅值,使用动态阀值进行图像分割,减少了Canny算法的空间复杂度和时间复杂度。实验证明该方法能够高效地提取单像素宽度的二维轮廓,如果使用高分辨率的工业摄像机拍照,就可以代替三坐标测量仪检测精密模板。
In order to quickly and effectively detect the templet profile, an edge extraction method based on Canny theory is put forward.While keeping the advantage of Canny operator,the method improves the positioning accuracy by amending Gaussian filter edge functions ,and effectively eliminates the influ- ence of local noise by the use of the Sobel operator to calculate gradient,and effectively reduces the space and time complexity by using of the one-dimensional linear interpolation to calculate the amplitude of the points in the gradient direction and by using of the dynamic threshold to make image segmentation^Experi- merits prove this method can be effectively used to extract the single-pixel-width two-dimension outline. If images are taken with the industry video camera of high resolution power,it can detect the precision templet instead of three-coordinate measuring machine.
出处
《机械设计与制造》
北大核心
2012年第12期177-179,共3页
Machinery Design & Manufacture
基金
辽宁省教育厅科研项目(L2010216)
