摘要
使用CCD的影像测量系统进行精加工零件尺寸检测时,零件轮廓边缘的定位精度和处理速度是优先考虑的2个问题。该文介绍了一种基于Facet模型的综合梯度边缘检测算法,并针对其边缘定位精度低和执行效率低的问题,给出一种提高边缘定位精度的算法和加速计算Facet模型的算法。该改进算法使用动态阈值定位边缘以提高边缘定位精度,同时将Facet模型的二维卷积分解为2个一维卷积核,使算法的复杂度从()Oi2降低到O(2i)。通过对理想影像和实际工件影像的边缘检测,说明该算法在影像测量中可以有效抑制工件影像的噪声,获得较好的边界特征,且执行效率较高。
The localization of a part＇s edges and the speed of the image are important factors in detecting the size of the part with a CCD image measurement system. An integrated gradient edge detecting algorithm based on the Facet model is studied in this paper. For the low efficiency and low localization, a dynamic thresholding method is used for enhancing the precision of localization, and an accelerated method is provided for computing 2D Facet model. It reduces the computation complexity from O（i^2） to O（2i）by decomposing the 2D convolution mask to two 1D masks. The principle of the algorithm is analyzed, and the experiments for detecting the edges of some images are finished. The images include an ideal image without noise and a practical image of a circuit board. With experimental results, it is known that the algorithm in this paper can reduce the effects of image noise availably and preferable edges are detected, and it is able to improve the speed of the computation.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第2期187-189,193,共4页
Computer Engineering
作者简介
隋连升（1972--），男，副教授、博士，主研方向：计算机视觉，图形图像处理；E-mail： Wanghui_cba@ 163.com
王慧，硕士研究生
