Research And Realization Of The Micro Displacement Detection System Based On Machine Vision

Micro displacement technology is one of the key technologies in the field of precision positioning,precision instrument manufacturing,precision measurement and so on.It plays a very important role in the field of industrial production and scientific research.With the development of science and technology,the micro and nano technology is rising rapidly,so that the precision requirement of the micro displacement technology is increasing,and so is the corresponding micro displacement detection technology.Nowadays,micro displacement detection methods are mainly as follows: the capacitive micro displacement detection,laser interference micro displacement detection,laser triangulation micro displacement detection and visual displacement detection.Visual displacement detection means that when the measuring object to be measured,using a video camera to replace the visual organ,the image as the carrier of information distance,machine vision technology is used to deal with the image of the object to be measured,withdrawing the characteristic information and the actual displacement of the object to be measured.Displacement detection based on machine vision technology,whose platform is simple,non-contact online detecting,easy to integrate and manage,high precision,is widely used in engineering fields.This paper does research on micro-displacement detection system based on machine vision to detect the micro-displacement of the piezoelectric actuator.According to the characteristics of the piezoelectric actuator,this paper designs a micro-displacement detection scheme suitable for the piezoelectric actuator.The scheme is to locate the objective feature borderline after segmenting the piezoelectric actuator in the denoised micro-displacement image to obtain the image displacement after moving to calculate the actual actuator's micro-displacement.For denoising,the four common denoising methods: mean filtering,median filtering,Gaussian filtering,bilateral filtering are made comparative experiments with the actuator images.The bilateral filtering is chosen since it has the best random noise filtering effect.In addition,for reflection,this paper proposes the Ostu segment method combined the truncation threshold to be applicant in reflection suppression to segment the reflection area,and does experiment with several actuator images under different illumination,verifying its feasibility,restraining the non-uniform illumination of the images,to a certain extent.For micro displacement image,image segmentation using threshold and edge detection are experimented to segment the actuator.The experimental results show that because of the influence of illumination,the pixel distribution in image and other factors,the piezoelectric actuator is not divided completely correctly using the main several threshold segmentation methods;in the situation of unsatisfactory illumination,some severely affected "false" edges are prone to appear using edge detection.In this paper,a image segmentation method combined the image segmentation using the basic global threshold with the edge detection using Canny operator is used to extract the contour edges of the piezoelectric actuators in the micro-displacement images.The preliminary location of the objective feature borderline of the moving side of the actuator is detected by Hough linear transform.The least square method is compared and chosen to find the accurate position of the objective feature borderline.Finally,the least square method is improved to suppress the influence of the line noises on the basis of the actual situation,increasing the location accuracy.This paper calibrates the camera,locates the objective feature borderline of the actuator in micro-displacement image,by machine vision technology,realizes the micro-displacement detection of the piezoelectric actuator.When the pixel equivalent is to 17.764 microns / pixel,the uncertainty of the borderline location is 0.3 pixel.The uncertainty of micro-displacement detected results is up to 7.9 micron,meeting the design requirements at the present stage.