| Micro-assembly technology is an indispensable method for manufacturing micro-electromechanical systems composed of micro-parts of different scales and materials.It has been widely used in the fields in micro-opto-electromechancial system,micro/nano robot technology,lab-on-chip,cell operation and so on.The three-dimensional(3D)surface morphology information of micro-parts to be assembled is the key parameter that needs to be analyzed and tested when micro-assembly system preforms assembly operation.In order to ensure the correctness and accuracy of micro-assembly system in clamping,alignment and approaching operations,it must be tested that the 3D surface morphology information of micro-parts has been assembled.At present,shape from focus(SFF)is the most commonly used method for non-contact passive 3D surface morphology information measurement based on visual images,but traditional SFF method has many problems.On the one hand,the approximate method can not guarantee the accuracy of the recovery when recovering the accurate depth information of the measured object point.On the other hand,the object to be measured needs to be moved when acquiring visual images,which is not suitable for micro-assembly system.Therefore,it is significant to study the depth information acquisition method of microscopic vision images for micro-assembly systems.Aiming at the above problems,this thesis proposes and implements a depth information acquisition method for microscopic vision tomographic scanning images of micro-assembly space based on 3D spatial interpolation technology.This method can achieve high-precision depth information acquisition of microscopic vision tomographic scanning images of micro parts to be assembled in a micro-assembly system.Based on the principle of the depth information acquisition method,simulation and micro-assembly system were designed to obtain microscopic vision tomographic scanning images and verify the depth information acquisition method.The main research work and results of this thesis include the following aspects:1.The system composition of micro-assembly system and the imaging principle of microscopic vision system are analyzed.In order to adapt to the characteristics of micro-assembly system and to ensure the simultaneous acquisition of local detail information and global morphology information of micro-parts,a microscopic vision tomographic scanning technology is used to obtain microscopic vision tomographic scanning image sequence containing complete 3D morphology surface information of micro-parts.Microscopic vision tomographic scanning technology drives microscopic vision system to carry out microscopic vision tomography in micro-assembly space through precision positioning platform system.2.The basic principle of depth information acquisition method of microscopic vision tomographic scanning images of micro-assembly space based on 3D spatial interpolation technology is proposed and studied.Firstly,based on the imaging principle of the microscopic vision system,and establishing the relationship between microscopic vision tomographic scanning images and spatial position of micro-parts,a local extremum method of focusing evaluation function is proposed to obtain the clear focusing contour of microscopic vision tomographic scanning images.The 3D space of each microscopic vision tomographic scanning image is constructed by the depth assignment of the clear focusing contour of each microscopic vision tomographic scanning image,so as to obtain the 3D space construction of all the microscopic vision tomographic scanning images.Secondly,a spatial depth interpolation method is proposed on the basis of obtaining the 3D space construction of all microscopic vision tomographic scanning images.By spatial depth interpolation between the clear focus planes in the adjacent three-dimensional space construction of microscopic vision tomographic scanning image,the depth information of the micro-parts to be assembled is accurately recovered,and the depth map and the 3D model of the micro-parts are obtained.3.A simulation experiment of depth information acquisition based on simulated images is designed for method verification and parameter analysis.By calculating the focusing and defocusing process of the optical simulation camera,the microscopic vision tomographic scanning image sequence of the simulation 3D micro-parts is obtained,and the simulation experiment of depth information acquisition is carried out.On this basis,the influences of various calculation parameters in depth information acquisition method on the results are analyzed,and the analysis results are verified by simulation experiments.The simulation experimental results show that the proposed depth information acquisition method can calculate the depth information correctly,which has better accuracy than the traditional SFF method,and can obtain more accurate depth information results by determining the influence parameters in the depth information acquisition reasonably.4.A micro-assembly experiment setup was established and used to verify the proposed depth information acquisition method.In the experiment,the cone cavity and the end gripper of the micro-gripper are taken as the experimental objects,the microscopic vision tomographic image sequence of the experimental objects are obtained by the microscopic tomography technology and the depth information acquisition calculation was completed to obtain the depth map and the 3D model of the experimental objects.The 3D model of the experimental objects and its own CAD model are analyzed for accuracy by ICP registration.The experimental results show that the method can obtain high-quality 3D surface morphology information of micro-parts to be assembled,and has a good guiding role for micro-operation of micro-assembly systems.The research work in this thesis has laid a theoretical foundation for the development of the automatic detection algorithm for 3D surface morphology information of the micro-assembly system,and promoted the widespread application of automatic micro-assembly online detection. |
PDF Full Text Request