Fast Recognition and Parameters Acquisition of Ellipse Target in Visual Assembly
XU Yi-heb · Jisuanji gongcheng · 2011
For the complex background information of image in the process of visual assembly,the recognition of ellipse(hole) target exits certain difficulty,this paper puts forward a method,using the advantages of randomized Hough transform and the least square method,to fast recognize ellipse target and accurately acquire its parameters.It extracts the image edge points after the preprocessing and uses the method of marking connected domain to validly reduce invalidly sampling and retain noise robustness of the Randomized Ellipse Detection(RED) as well.It makes use of the least square algorithm to fit parameters of the detected ellipse again,and acquires accurate parameters of the ellipse.Experimental results show that the method can obviously reduce invalid sample,quickly and accurately recognize ellipse target under the condition of complicated background information.