The Liner Structured Light Measurement Technology Of Disk Face Runout

Traditional methods for disk face runout measurement all belong to contacting measurement.When measure disks and electronic devices, etc by this way, the measured disk is pressed by theprobe in the process, so the area of the measured disk will produce elastic deformation, thus causemeasu-rement error. At the same time, because of the changing of the vibration state of rotatingdisk itself caused by the contacting force of the probe,it slao can lead to measurement failure. Sothe non-contacting methods to finish the measurement of the rotating disk face runout is needed.The visual measurement using computer vision technology has completed the partsnon-contacting detection. It adopts the liner structured light sensor measuring system based oncomputer vision measuring technology to measure the face runout of the rotating disk and thespecific research content is as follows:First of all, based on Zhang’s camera model, the measurement model based on the linerstructured light vision sensor was established and calibrated the parameters of the liner structuredlight visual sensor by using a method based on free moving plane target. In the process ofcalibration, It built the partial world coordinate system in the free moving plane target andobtained the partial world coordinates of the feature point on the light plane by using the methodof cross ratio invariability. Then got the cameracoordinates by calibrating the external parametersof the camera. Use the method above to obtain the cameracoordinates of more other feature pointsby moving the plane target to three positions and by means of least square method to fit theequation of the light plane under the camera coordinate system. By calibrating the inter parametersof the camera to get the coordinate undre the camera coordinate system of the subpoint in theimage coordinate system of any point in the light plane. Then got a straight-line equation throughthis point and the optical center of the camera. At last,by solving the Intersection of thisstraight-line equation and the equation of the light plane gained can get the three-dimensionalcoordinates of any point.Because that when the light plane intersected with the the surface measured, it formed acertain pixel width of structured light stripe. Therefore,in order to improve the accuracy of themeasurement model, it needed to pick up thecenter line of the structured light stripe at first. But inthis process, due to the influence of the uneven illumination and inherent characteristic on thesurface of the object to be measured, the extraction of the structured light stripe centerline becamemuch more difficult. To improve the accuracy of the centerline extraction and improve the efficiency, first need to preprocessing, including image cut, image smoothing, imagenormalization and image segmentation. For the method of structured light stripe centerlineextraction, it focused on the comparision of the precision of three methods, including gray-gravitymethod, direction template method and gaussian fitting method and found that about gaussianfitting method, its extraction accuracy can reach sub-pixel level and can be much faster. It is verysuitable for the real-time measurement of the face runout of the rotating disk and it was verified byexperiments.Finally, it used the method above to measure the face runout of disk, and compared alsoanalysised with the measured result of the linear displacement sensor. The result showed thatabout the disc face runout measurement method in this paper compared to the linear displacementsensor, the standard error of mean is less than0.08mm, and its standard deviation is less than0.03mm.The research work of this paper give a major push to the development of the onlinemeasurement for the face runout of the disk parts.