| High surface precision and strong reflective properties lead to various surface features submerged in strong light,making it difficult to detect surface defects of the highlighted metal.The existing machine vision surface defect detection technology cannot effectively detect the defects and extract the features of metal surface under high-light pollution.This article explores a new method of weak image acquisition under the strong noise background,through the enhanced intensity of the noise of the input image enhancement,as the radiation energy of the target imaging(ghost)has also been enhanced,replace main image with high signal-to-noise ratio of the ghost,eliminating high due to the metal surface reflection characteristics of light pollution.Then,this method is applied to detect the problems of missing,leaking and error of fin evaporator ridge groove.Combined with the image processing algorithm,a defect detection system is built.Through experiments,the accuracy and reliability of the method are verified.On the basis of eliminating the influence of high-light noise on the surface of the high-reflective metal,the characteristics of the groove on the ridge of the evaporator can be identified effectively.In this paper,we firstly analyze the principle of stray light and ghost imaging,and then simply calculate the formation position of low-order ghost image,and analyze the influence of incidence Angle,optical medium and radius of curvature on ghost image.The simulation analysis is mainly aimed at the drawback that ghost image is not clear enough for traditional industrial CCD camera.The camera lens is optimized from the aspects of lens structure,incident Angle,radius of curvature and mirror material by using Zemax software,and the optimized illuminance diagram and energy diagram are calculated,so as to analyze and optimize the results intuitively.Secondly,the construction of the system,the system is divided into software part and hardware part,the hardware part according to the test system requirements and performance index requirements,design hardware function structure,select the system required hardware and its model.The software system is based on Microsoft Visual Studio 2017 platform Qt plug-in development,development language for C++.Finally,a concise and efficient algorithm is selected for ghost image quality,including image enhancement and feature recognition algorithm.Retinex image enhancement algorithm based on local features is used to enhance the image quality of pre-processed images,reduce image noise,and improve the speed and accuracy of subsequent processing flow.In the feature recognition part,firstly,the image segmentation algorithum is used to separate the feature area,and then the pixel value is used to quantify the area of the defect area.In view of the difference of the pixel value of intact and defect features,the probability and statistics method is used to set the partition threshold,so as to facilitate the machine identification and decision making and the identification of defective products.Experiments show that this algorithm can effectively detect the defects on the ridge surface of the high-gloss finned evaporator.Figure[55]table[5]reference[81]... |
PDF Full Text Request