图像处理Image processing for machine vision measurement of hot formed parts

发布时间：2013-08-08 01:31:05

Journal of Materials Processing Technology174(2006)1–6Image processing for machine vision measurement of hot formed partsS.B.Dworkin,T.J.Nye∗Mechanical Engineering Department,McMaster University,1280Main Street West,Hamilton,Ont.,Canada L8S4L7Received12September2002;accepted18October2004AbstractDimensional measurement of hot parts,such as forgings,is desirable to permit real-time process control,but typically is inconvenient because of the difficulty in working with very hot workpieces.Machine vision technology is employed in many industries for rapid,in-process non-contact gauging,but in this particular application there are implementation issues.These include workplace illumination,image processing,pattern analysis and extracting important features from the processed images.This paper describes experiments conducted on acquiring and processing images of hot forgings,specifically large open die forgings,and how the use of inexpensive CCD cameras with infrared-passfilters can greatly simplify the image processing and pattern analysis tasks for a machine vision measurement system.©2006Published by Elsevier B.V.Keywords:Open die forging;Image processing;Computer vision;Measurements;Quality control;Infrared sensor1.IntroductionThere are a number of hot-working processes where in-process dimensional measurements are desirable,but are inconvenient to obtain due to the difficulty in handling these hot parts.Manual,contact measurement equipment may be uncomfortable for operators to use,and indeed may present a safety hazard by exposing the operator to such hot parts. Non-contact gauging systems such as machine vision imag-ing systems have the potential to reduce this problem by separating the hot part from the measurement device through an optical medium.A typical digital imaging system is shown in Fig.1.A CCD camera and lens acquires an image of the scene con-taining the part to be measured.This image is digitized through the use of a frame grabber card that stores the image as a two-dimensional array.Each value in the array contains an integer representing the light intensity of the cor-responding pixel(i.e.,picture element)of the camera CCD. Monochrome images are stored as eight-bit integer values in each array position while colour images are typically stored using24bits per pixel,with eight bits representing ∗Corresponding author.Tel.:+19055259140;fax:+19055727944.E-mail address:nyet@mcmaster.ca(T.J.Nye).the intensity of each of red,green and blue light compo-nents.Once an image is acquired,it is normally processed(e.g.,filtered,enhanced,etc.)and pattern analysis(e.g.,edge detec-tion,feature extraction,etc.)is performed.Fig.2shows a typical scene of an open die forging operation.Along with the workpiece(the white object in the upper right quad-rant of the image),the image contains parts of the forging press,dies,miscellaneous equipment and unrelated back-ground details.Before measurements can be taken of the workpiece,it must be identified reliably within the image and the extraneous information must be removed.In general,this can be a very difficult computational task[1,2],so image-processing strategies are commonly employed to simplify the image before pattern analysis takes place.This paper will describe a particular processing technique,image threshold-ing,as it applies to hot parts such as forgings.This technique is described and results of the image processing experiments are outlined.2.Literature reviewTraditionally,open die forgings are measured while hot, using relatively simple measurement tools such as calipers.0924-0136/$–see front matter©2006Published by Elsevier B.V. doi:10.1016/j.jmatprotec.2004.10.019

图像处理Image processing for machine vision measurement of hot formed parts

推荐内容

相关推荐