Font Size: a A A

Research And Implementation Of High Light Removal Technology For Feature Extraction Of Metal Parts

Posted on:2020-09-26Degree:MasterType:Thesis
Country:ChinaCandidate:Y X LuFull Text:PDF
GTID:2382330572469353Subject:Mechanical design and theory
Abstract/Summary:PDF Full Text Request
The accuracy of geometric features and parameters of mechanical parts is the key to ensure the assembly quality of products.In every link of industry,it is necessary to test and analyze the important characteristic parameters of parts to verify whether they meet the use standards.Due to the advantages of high efficiency and convenience,machine vision technology is widely used in target feature detection.However,due to the Fresnel reflection characteristics of metal parts,the surface of metal parts in high-light environment is prone to mirror reflection.At the same time,the illumination and reflection conditions in industrial production and assembly environment are complex,so the metal parts in the scene are taken as images.The problem of high-light reflection is often encountered in gathering time,which will make the detection result of characteristic parameters of target parts inaccurate.Traditional image feature detection methods often require standard illumination chromaticity and uniform distribution of the image,and the reflected light of the collected object needs diffuse reflection to ensure the integrity of image information.Therefore,when using traditional image feature detection methods to process metal parts images with high light,the problem of inaccurate feature detection will be encountered.In this paper,the problem of low accuracy of geometric feature detection of metal parts in high-light environment and vulnerability of image processing to light source changes is studied in depth.Based on the research results,a feature extraction system of metal parts in high-light environment is developed and verified by an example.The main contents of this paper include:In the first chapter,the research status of color balance method,highlight removal and restoration method,target feature enhancement and extraction method at home and abroad are summarized.This paper analyses and summarizes the progress of the above-mentioned technology research and the existing problems,expounds the research content and the significance of the research,and finally introduces the text structure of this paper.In the second chapter,based on the theory of two-color reflection model of image processing,the method of fitting the chromaticity function in image chromaticity space and the method of obtaining the chromaticity of illumination(mirror component)of image are improved.Combining with diagonal matrix model,image color balance is achieved,and the validity of the proposed method is verified by comparative experiments.In Chapter 3,aiming at the task of removing and repairing high-light components in metal parts images,a method of removing and repairing high-light components based on saturated and unsaturated high-light distinction is proposed.An evaluation model of classification results of pixels based on chroma is constructed and the priority calculation formula of sample block repairing method is improved.The specular components in high-light images are removed and repaired.The validity of the proposed method is verified by comparative experiments.In the fourth chapter,the method of feature enhancement in low illumination region of metal parts in high-light image is studied,and the logarithmic domain mapping method of hue mapping in high dynamic range of image is improved.The scheme of identifying and extracting the external thread parameters and circular features of metal parts is realized,and the feasibility of this scheme is verified by experiments.In the fifth chapter,based on the method and software tool library described in this paper,a feature extraction system for metal parts highlight image is developed,which has the functions of illumination color balance,image highlight removal and restoration,thread feature and circular feature extraction of metal parts,and is applied to the geometric feature extraction of threaded joints and verified.The sixth chapter summarizes the research results of this paper and the existing deficiencies.The prospect of future research content and direction are proposed based on the results provided.
Keywords/Search Tags:Image Processing, High Light Environment, Metal Parts, Geometric Feature Extraction, Detection
PDF Full Text Request
Related items