Data Processing Of Vertical Scanning White Light Interferometry System And Its Application In Grinding Wheel Morphology Survey

With the rapid development of production and processing, three-dimensionalmeasurement has become the focus of the relative study. Vertical scanning white lightinterferometer based on surface scanning has advantages of high efficiency, highprecision, large range and no phase ambiguity. This technology has been applied inthe area of micro surface measurement, especially microelectronic devices whichcontain grooves or steps. This paper introduces the principle of vertical scanningwhite light interferometer, analyses several data processing methods as well as itsapplication in resin diamond grinding wheel topography measurement.The main contents and innovations are as follows:(1)Starting from the basic theory of light interference, vertical scanning whitelight interferometer which combines with phase-shifting interferometry and whitelight interferometry technology was described in detail. Complete optical imagingsystem based on the principle was designed, which can get and save the white lightinterference fringes. The related parameters of the optical system were also analyzed.(2)Several kinds of data processing of vertical scanning white-lightinterferometer were introduced and analyzed. Besides, a new method based onGaussian curve using particle swarm optimization algorithm was proposed. Twogroups of simulation data, a micro channel and two standard samples with differentkinds of reticles were sampled and processed respectively.(3)Vertical scanning white-light interferometer was used in resin diamondgrinding wheel surface measurement. According to the difference reflectioncharacteristics of the resin material and diamond material, a theory of grainsidentification based on the difference between of maximum and minimum ofinterference intensity was proposed. Then two methods were provided. One of themis based on the threshold of interference intensity difference; the threshold selectionwas also discussed. Another method was based on data fusion technology, which usedeigenvalues of boundary process to detect and recognize the grains. Different sampling intervals were tested in order to find their influences of these two methods.