Study Of Gear Measurement Technology Based On Machine Vision

For the purpose of preventing low accuracy, heavy workload and other defects of thetraditional contact measurement method, in this paper, a measurement means for sizedeviation of middle and small module spur is proposed based on machine vision and digitalimage processing. Taking middle and small module spur gear as a measurement object, usingthis measurement method to calculate the pitch deviation and the tooth profile deviation inorder to realize the real-time, non-contact measurement of gear deviation. Using the techniqueposed in this paper, the taken gear images are pre-processed correspondingly by suitabledigital image processing. Then the clear gear profile image is gotten. Using softwareprogramming to calculate the coordinates of the center of the gear, finally the mathematicsmodel of image measurement for circular pitch error and the profile error are designed.The methods of getting gear tooth profile and pre-processing technology are firstlydescribed by this paper. Several means of typical graphics filtering and graphics brinkdetection operator are presented in detail. According to the principle analysis and a largenumber of experiments, the bilateral filter is chosen for image preprocessing. The classicalcanny digital image edge detection operator is studied specifically. This paper does severalcorresponding improvements on the parts of solving for the gradient magnitude, selecting highand low thresholds, connecting edge, and then poses an improved canny digital graphics brinkgauging arithmetic of adaptation. Taking the standard gauge block and the gear as theexamples to do the tests on the proposed algorithm and the traditional algorithm to see theirdifference, the comparison results fully prove that the posed arithmetic is effective and reliable.Then introduces several classical sub-pixel subdivision algorithms and fitting algorithms forsub-pixel edge used commonly in digital image processing. Using the bezier surfaceinterpolation arithmetic for sub-pixel subdivision and fitting arithmetic based on Gauss surfacefor sub-pixel edge in digital image. This measurement system is demarcated, that is, to obtainthe ratio between the actual dimension and the pixel dimension. Calculating the centercoordinates of the gear shaft hole by using least square method, that is, the center coordinatesof the gear. Then spur gear circular pitch error and the profile error mathematic module aredevised based on GB, devising the measurement course of individual circular pitch error,cumulative pitch error, total cumulative pitch error and profile error Taking the standard involute spur gear as an instance using the universal tool microscope to gauge the mentioneddeviation of the gear. Then making use of the designed arithmetic to compute each errorfinally, then outcomes are contrasted. By comparison the proposed algorithm is much betterthan the traditional methods in accuracy and efficiency, the posed method is effective andpracticable. Finally, some feasible measures are given to heighten the measuring accuracy ofthe system after analyzing the error in the system.Experiments point that, this proposed way can achieve real-time, high-precision,non-contact measurement for pitch error and profile error of middle and small modulus spurgear, Which has a very inclusive application prospect.