Research On Positioning System Of Photosensitive Elements And Optical Focusing Lens Based On Visual Detection

In recent years,with the development of industry,the demand for optical instruments has increased dramatically.As an important part of optical instruments,the positioning accuracy between optical focusing lens and photosensitive elements directly determines the quality of optical products.Meanwhile,machine vision-based detection technology has been widely used in automation industry.This technology improves productivity by replacing manual work.This paper studies the application of vision detection technology in the positioning of optical focusing lens components.The main contents include the following parts:(1)Aiming at the focus error of optical focusing lens caused by manufacturing accuracy,the overall scheme of optical focusing lens assembly positioning system based on visual detection is designed,and the selection of light source,camera and lens is carried out according to the requirements of positioning accuracy.(2)Aiming at the uncertainties of the focus position of the optical focusing lens and the irregularities of the spot area of the focus image,a set of focus center detection algorithm is designed.Firstly,the illumination intensity of the spot area in the image is determined according to the spot gradient,region filling and morphological corrosion.Then,the spatial position of the lens focus is determined by comparing with the threshold range.Finally,the focus center is detected by the gravity center method.(3)In view of the complex background circuit,large interference information and small relative size of the photosensitive elements,a set of photosensitive element central detection algorithms is designed.Based on the combination of template matching and contour fitting,this set of algorithms innovatively put forward a rough-first-refined photosensitive element location strategy.In the stage of rough positioning,the template matching algorithms such as MAD,NCC and NCC based on pyramid search are analyzed firstly,and their advantages and disadvantages are compared through experiments.Finally,the NCC algorithm based on pyramid search with low matching time and high accuracy is selected.In the stage of edge extraction,the advantages and disadvantages of edge detection algorithms such as Prewitt,Canny and Qtsu based adaptive Canny are compared and analyzed.Finally,the anti-collision algorithm is selected.An adaptive Canny edge detection algorithm based on Qtsu with strong illumination interference is proposed.In the stage of target extraction,the connected region labeling method is selected to eliminate the interference of non-target factors,so as to prepare for the next step of precise positioning.In the stage of target positioning,the least square circle fitting algorithm is used to achieve accurate target positioning.(4)The experimental platform is built,and the calibration of focus space,industrial camera and optical focusing lens assembly are completed.According to the experimental results,the average positioning accuracy is 0.018 mm,the relative positioning accuracy is 1.87%.Then the error analysis is carried out from the aspects of vibration factor,image processing algorithm,calibration parameters,lighting system and mechanical error,the data show that the error is far less than the requirement of positioning accuracy.The alignment system designed in this paper improves the positioning.In terms of accuracy,it has important reference value for optical instrument industry.