摘要
由于光照不均和被测物表面污染等原因,使视觉测量系统所采集图像可能携带块状和带状噪声,而通常的边缘检测算法无法有效地去除这类噪声,由此造成假边缘。以此为切入点,提出了当噪声所造成假边缘几何尺寸小于被测物体本身的几何尺寸时,可以去除假性边缘的"直线模板去噪法"。所论方法在检测阶跃状边缘时提出,同时可将此法推广到屋脊状边缘处理。实验证明该算法有较好的去噪能力。
In the AOI system,the zonal or cluster noise may cause the extraction of false edges.So the paper puts forward a new method called ＂the method of de-noise based on linear template＂ to eliminate the false edges,when the size of the object＇s real edge is larger than the size of false edges.Firstly the method is used to detect the step edge,and then the method is extended to detect the ridgelike edge.To compare the effect of this new method against some of the usual methods,a real image is processed by using these different methods,and the result show that the new method can eliminate the noise effectively.
出处
《上海应用技术学院学报（自然科学版）》
2010年第1期18-21,共4页
Journal of Shanghai Institute of Technology: Natural Science
关键词
去噪
直线模板
视觉检测
边缘检测
de-noise
linear template
vision inspection
edge detection
