Font Size: a A A

Image Recognition And Measurement Algorithm Design And Realization On Nameplate Classification And Quality Detection

Posted on:2017-02-07Degree:MasterType:Thesis
Country:ChinaCandidate:Y K GuoFull Text:PDF
GTID:2308330485478407Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Measurement technology and image recogniton technology based on machine vision has been widely used in industrial applications. In industrial manufacture,circular metal nameplate in the process of stamping will appear deformation caused varying degrees deviation of radian,beside the surface of nameplate coated character could appear plating wrong characters this situation, therefore it needs to distinguish several kinds of characters and meantime measure the nameplate section radius. Eliminating the nameplate which does not conform to the requirements of these situations.The author of this paper has carried on the design and realization of image recognition and measurement algorithm for these cases,and achieve good results. The main research contents are as follows:1. The overall introduction of nameplate classification and quality inspection systemFirstly it introduces the overall framework and work flow of the nameplate classification and quality inspection system,and then according to the functional requirements and performance index of the detection system, choose different light sources and camera, set the corresponding working environment to make sure the effect of measurement and identification best,at the end of chapter introduced the hardware and software of system.2. The design and realization of nameplate character classification algorithmFirst it introduces open source computer vision library OpenCV in VS2010 environment configuration.According to the character classification requirment designed nameplate character classifiction algorithm.The whole process of algorithm realization mainly included image binarization, image denoising, image morphology processing, character positioning, character segmentation, normalization of the characters, character refinement, character classifiction, and introduced some commonly used methods of binarization, denoising method, and the morphological processing in details, and used for image recognition method of template matching. The end of this section is summarized.3. The design and realization of the nameplate section radius measurement algorithm This section first tells the traditional method of circular arc radius measurement principle and the method based on least squares fitting round radius measuring principle. Focus on the method of the least squares fitting circle to carry on algorithm design, the whole process including the nameplate image preprocessing, contour extraction, and the design of arc fitting parts.Finally this chapter is summarized.4.The test result and analysis of nameplate character classification algorithm and nameplate section radius measurement algorithmThis section mainly tells test method and result analysis of nameplate classifiction algorithm and radius measurement algorithm. Finally this chapter is summarized.In the end,the full paper is summarized, the main research findings are illustrated and points out the deficiency and needs to be improved.
Keywords/Search Tags:Machine Vision, least square method, Character classification, Template matching
PDF Full Text Request
Related items