Research Of 3D Measurement Technology For Inner Surface Of Deep Holes

Parts with deep holes,including bearings,air cylinders and gun barrels,are widely used in automotive industry and defense industry.Deep hole machining is one of the most challenging problems in manufacturing.It is very difficult to ensure the diameter,roundness and straightness tolerance because of the tools' bending and vibration during machining.At the same time,deep holes are usually involved in shaft-hole matching,and in this case,the inner surface of deep holes work in harsh environment with abrasion,vibration and thermal deformation,which usually lead to inner surface defects.These defects will further degrade the parts' sealing and motion performance.Therefore,it is very important to measure the dimension errors,form errors and defects of the deep holes' inner surface.However,it is very difficult to measure the inner surface of a deep hole because of its features,which makes it the research hotspot in the field of instrument science.Firstly,the inner surfaces of deep holes have small diameter and limited space.Therefore,the inner surface measurement system is required to be compact and have a small diameter,which makes it impossible to apply some mature surface measurement technologies such as fringe projection.Secondly,in the context of shaft-hole matching,any kind of errors and defects at any location will lead to part failure.In other words,the inner surface measurement system is require to have full-field measurement capacity,which means that all kinds of contacting measurement,such as coordinate measuring machine,are not applicable.At last,so many parameters,including form errors,dimension errors and surface defects,together influent the working condition of a deep hole.And the best way to obtain all these parameters at one time is to measure a high resolution 3D point cloud of the inner surface,which contains all kinds of errors.So it would be better if the inner surface measurement system has 3D measurement capacity.In order to meet the requirement of compact structure,full-filed capacity and 3D capacity,this dissertation proposed a circle-structured light measurement system for inner surface detection based on laser triangulation and digital photogrammetry.The working principle of this system is:A laser beam is reflected by a conical reflector and then create the circle-structured light,whose travel path is conic or plane;this circle-structured light intersects with the inner surface and create a light stripe,which is captured by a CCD camera;using the parameters of the structured light and CCD camera,the 3D coordinates of inner surface points are calculated basing on laser triangulation algorithms;the 3D point cloud is obtained after scanning the whole inner surface;the form errors,dimension errors and defect information are extracted from the 3D point cloud basing on proposed algorithms.The first thing to realize the measurement is to establish the mathematical model of the proposed measurement system.This dissertation first studied the imaging model and calibration method of CCD cameras,then Zhang's camera calibration algorithm is improved to better work in the proposed system.The improvements mainly include using calibration boards with circle marks and distortion models with consideration of tangential distortion.This dissertation then studied the systematic errors of the circle-structured light generator,and established the expressions of the structured light with consideration of multiple kinds of assembly errors.Moreover,a flexible calibration algorithm based on binocular vision is proposed to obtain all the parameters of above-mentioned expressions.This calibration algorithm is flexible and easy to operate,and can precisely calibrate the structured light parameters without complex instruments.This algorithm can also calibrate line-structured light after improvements.Processing algorithms of 3D point clouds are also studied in this dissertation,including 3D point cloud-based form errors evaluation and point cloud registration-based segment scanning method.To solve the form errors evaluation problems,a roundness error evaluation algorithm based on iteration geometrical searching and a spatial straightness evaluation algorithm based on computational geometry are developed,which can realize form errors evaluation with less calculated amount.Point cloud registration-based segment scanning method decreases the dependence of system scanning length on the range of ball screw,and extends the system's application in the measurement of super-length deep holes.At last,the structure design,device selection and system set-up are realized.Inner diameter measurement experiments,roundness and straightness evaluation experiments and inner surface defect detecting experiments are carried out on the real system.Experiment results indicate that both the proposed system and corresponding algorithms are reliable and accurate in inner surface measurement.