Research Of Rotary Parts Non-contact Measurement Algorithms

With the rapid development of modern industry, the traditional means of detection hasbeen difficult to cope with the requirements of the rotary type work increasingly highmeasurement accuracy. Because of the Laser detection system has a lot of advantages, such asfast speed, a high degree of automation, lower cost and so on, there have importantapplications in the product processing and testing industrial and aviation, spaceflight.underthe conditions determined by the hardware system, To study algorithms to improvemeasurement accuracy from the perspective of signal processing has become a hot topic oftoday’s measurement.According to the characteristics of the original detection signal of the part,Thedissertation uses the multi-scale filtering algorithm combined with time-frequency topretreat,firstly the original sampled signal is filtered in the time domain to eliminate thecareless errors,and then the Kalman prediction filter is used to extract detection signalpreliminary,lastly the signal is decomposed to the scale and frequency band by wavelet packet.multi-scale threshold filters the signal in the frequency domain to extract the useful signalaccurately. Through the cubic b-spline fitting method calculating the value of outer diameterof axially symmetrical parts, and the fitting precision can reach microns.Based on artificialimmune algorithm, respectively, the least squares circle method (by LSCM), the minimumzone circle method (MZCM), the smallest circum circle (MCCM) and the largest inscribedcircle method (MICM)is used to evaluate the roundness error, at the same time ,the MCCMMICM MZCM error and concentricity error evaluation model based on artificial immune, theanalysis shows that: artificial immune algorithm can improve the accuracy and efficiency ofthe shape and position errors . Finally, the system error analysis is completed.