Research On Image Acquisition And Detection Of Special-shaped Complex Parts Based On Machine Vision

Under the circumstance that the traditional measurement methods are gradually unable to adapt to the rapid development of modern industry,the detection method is changing from contact measurement to non-contact measurement.In the field of noncontact measurement,image detection technology based on machine vision has excellent performance in many aspects,thus it has been widely concerned and applied.Specialshaped parts have free-form surfaces that cannot be accurately described by analytic geometry,traditional detection methods have low detection efficiency and poor detection effect,and there are some difficulties in detection.In order to solve this problem,this paper studies the image acquisition and detection of special-shaped parts based on machine vision,aiming at improving the detection efficiency and accuracy of specialshaped parts by image detection technology.The main research contents of this paper are as follows:(1)Decide to use linear array CCD as the main image acquisition equipment,analyze the principle of image acquisition of CCD image sensor,establish the mathematical model of linear array camera imaging process based on the principle,and extract the key parameters affecting the imaging accuracy of linear array camera from the model: pixel gray value and pixel column standard.(2)The four coordinate systems of the imaging system are analyzed,the mapping relationship between the real object and the image is established by the transformation matrix between the four coordinate systems,the non-linear factor?geometric distortion existing in the mapping is studied,and a distortion correction method based on neural network is proposed.Experiments show that this method has better correction effect than traditional polynomial fitting method.(3)After analyzing the image acquisition process of linear array camera,it is concluded that the main factor restricting the sampling accuracy of linear array camera is that the scanning mechanism cannot accurately control the sampling coordinates.In this paper,an image acquisition method based on fusion grating ruler is proposed.This method can not only achieve high-precision image acquisition of linear array camera,but also improve the sampling accuracy to sub-pixel level.Based on this method,an image acquisition platform is set up.Through the measurement of the object to be measured,good detection effect has been achieved,and the highest measurement accuracy can reach 1 ?m.(4)The shortcomings of the traditional edge detection algorithm for special-shaped curve are analyzed,that is,the detection accuracy stays at the pixel level,and when the image resolution is low,the measurement error will be large.A sub-pixel edge detection algorithm is proposed.The algorithm can accurately fit the profile curve under pixel and improve the measurement accuracy.From the fitting curve,the analytic function which can approximate the profile curve can be obtained,and the position coordinates of the points on the profile curve can be obtained.