Research On Robust Filtering Algorithm In Surface Roughness Measurement

In modern society,not only put forward higher requirements on the fine degree of industry on surface micro morphology characteristics of the surface,in the daily life of people on their own use of consumer goods processing precision has certain requirements.The roughness and waviness shape,composed of surface feature,surface roughness described is the micro appearance of parts of the surface,with short wavelength(high frequency)characteristics,the accurate measurement and evaluation of far-reaching significance,has very important significance for machining parts.The evaluation method of surface topography used in this study is based on the Gauss filtering method.The experimental data are measured by stylus surface roughness meter.Aiming at the limitations of the traditional Gauss filter in use,there is a special feature of gross error and data values in the surface and lead to misleading benchmark by Gauss filter normal distribution bound caused by the deformity,introduces the robust theory to improve it.After the robust filtering algorithm needs to be achieved,robustness metrics and classical robust estimation methods are analyzed,and the most commonly used M estimation is selected as robust estimation method.Then the M estimation is analyzed in detail.The implementation of robust Gauss filter algorithm uses a modified weight function method,the specific implementation of the iterative algorithm.According to the characteristics of different data,the corresponding robust estimation weight function is selected to achieve the best processing effect.Algorithm of the tool used is the Matlab2015 Edition,first preprocess data and graphics data,according to the characteristics of the QC function for processing,and compared before and after the introduction of the robust filtering effect,prove that the introduction of appropriate robust algorithm can influence caused by abnormal characteristics of good data on inhibition filter reference,can getmore accurate surface roughness filtering line,meet the requests.