Font Size: a A A

Research On 3D Reconstruction Method Based On Engineering Drawings Image Of The Flange

Posted on:2020-11-25Degree:MasterType:Thesis
Country:ChinaCandidate:H YuFull Text:PDF
GTID:2392330590481575Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
Engineering drawings are important technical materials in the machinery industry,which condenses the wisdom and efforts of researchers.Intelligent identification and 3D reconstruction can effectively discover and protect the information contained in them.Although the 3D reconstruction project has been proposed for more than 40 years,it still has problems such as low efficiency,poor accuracy and ambiguity of the model.Although3 D reconstruction of the subject fruitful but there are some limitations: most of the research is based on dxf and dwg files,however the desired conversion of paper drawings often;Most research requires harsh conditions and requires specialized systems and computers with powerful computing power;New methods and new technologies are not applied much.Aiming at the above problems,this paper proposes a flange part reconstruction method based on the drawing image with the flange parts as an example.Methods to study engineering drawings,scanning,denoising,segmentation and thinning on it,then Harris corner recognition algorithm and Hough circle detection algorithm were used to extract and count the features.BP neural network is used to identify and extract the necessary parameters for the reconstruction of parts,and finally the 3D reconstruction of the flange model is realized.(1)The computer reading process of engineering drawings is introduced in detail,and the gray level transformation and noise reduction methods of raster images obtained by scanning are introduced.The traditional denoising algorithm and NL-means algorithm are described and analyzed.The feasibility of NL-means algorithm for engineering drawings denoising is proved by experiments.(2)This paper introduces the traditional localization algorithm,and describes the principle of LSD line detection algorithm in detail.Based on this,a character localization method of engineering drawings based on LSD line detection is proposed,and its feasibility is proved by experiments.This paper introduces the classical character recognition algorithm and depth belief network,and uses DBN to recognize engineering drawings characters.(3)In this paper,the image is processed based on the additive principle of images,and a more easily segmented image is obtained.This paper describes the basic principle of the Otsu method and the maximum entropy method,and introduces the FOA and BA algorithms in detail.On this basis,a method of engineering drawings segmentation based on swarm intelligence algorithm is proposed,which can segment engineering drawings quickly and accurately.(4)The flange parts are deeply studied,and the features and uses of flange parts are introduced.The 3D modeling software Solid Works is introduced in detail,and the concrete steps of its parametric modeling are described in detail.This paper analyzes the basic principle of Harris corner recognition,and uses it to identify the flange parts drawings.The corner information can be accurately obtained through experimental verification.The basic principle of Hough circle detection is analyzed,and the flange part drawings are identified by Hough circle detection.The circle information can be obtained accurately by experiment.The basic principle of BP neural network and the detailed process and steps of building BP neural network with MATLAB software are described.On this basis,BP neural network is used to identify the engineering drawings of flange parts,and the reconstruction information of flange parts is obtained.Based on this,3D reconstruction of flange parts is carried out by parameterized modeling function of Solid Works.In summary,this paper presents a three-dimensional reconstruction method of flange parts based on advanced technology and achievements such as image processing feature recognition and neural network.To a certain extent,the efficiency and precision of reconstruction are improved;this paper which takes paper drawings as the research object broadens the practical application of 3D reconstruction to a certain extent.This paper Using MATLAB and Solid Works as the platform do not need to design a special system or platform,which is conducive to the realization and promotion of two-dimensional to three-dimensional transformation.Experiments show that this method is effective and rapid in reconstructing the 3D model of flange parts,which provides a new idea and method for the 3D reconstruction of Engineering drawings,and has a certain guiding significance for the follow-up study.
Keywords/Search Tags:3D Reconstruction, Engineering drawings, Image identification, Image segmentation, Threshold method, Feature extraction, Back propagation
PDF Full Text Request
Related items