Automated Feature Localization And Assembly Control For Miniature Magnetic Parts

As one of the core components of the accelerometer,the micro-miniature magnetic component provides a stable magnetic field for the accelerometer.The electromagnetic force is balanced with the acceleration inertial force,so as to achieve the purpose of accurately measuring the acceleration.The assembly accuracy between the magnetic component and the base will affect the uniformity and stability of the magnetic field.In order to make the accelerometer achieve the expected performance and accuracy,the assembly accuracy is required to be high.At present,this assembly link is mainly completed by manual or semi-automatic assembly equipment.The production capacity and accelerometer’s performance are restricted by human factors.To the above situation,an automatic feature positioning method and an assembly control method are designed,the control software is written to realize the unattended automatic assembly of magnetic components.Firstly,the micro magnetic components are studied.The characteristics of the components as well as the requirements and problems in the assembly process are analyzed.Build fully automated assembly equipment and analyze its hardware structure and connection methods.The whole equipment is built with the idea of modularization.It provides the hardware basis for the realization of unattended assembly.The layered software architecture is designed to improve the maintainability and scalability of the program;through the multi-thread control,each module can work in parallel,the automatic loading and unloading and assembly work is realized,and the system assembly efficiency is improved.In order to meet the practical needs of batch precision assembly for microscopic vision guidance,a microscopic vision localization algorithm that can adapt to batch assembly is proposed.The algorithm extracts and fuses the gradient histogram and local binary pattern features of the target image,and trains the support vector machine.The sliding window is used to search for the highest score to achieve the localization of target features.In order to improve the efficiency of the algorithm,the pyramid search strategy is used to search layer by layer to reduce the amount of computation.The stability and accuracy of the method are verified by experiments,which shows that the algorithm can meet the requirements of batch assembly.Since the existence of gaps between the components after assembly and their own strong magnetic field,the magnetic interference during the assembly process will cause a slight offset at the end of the robotic arm.In view of this,a control scheme for the assembly of magnetic components is proposed.It uses visual servo to resist the influence of magnetic force,and an accurate positioning algorithm is proposed to realize the position positioning of magnetic components quickly and accurately.The extraction of edge information and the use of least squares fitting enable subpixel-level localization of magnetic components.Using the Kalman filter algorithm,the image Jacobian matrix is estimated in real time.This algorithm realizes the adjustment of the base position in the assembly process.Finally,the software implementation of the above algorithm is carried out,and additional assembly experiments are performed.The precision assembly process is designed to realize unattended automatic loading and unloading,feature extraction and positioning,assembly locking and other operations.Experiments to complete the batch assembly of magnet components in automatic precision assembly equipment are performed.The results show that the assembly coaxiality is better than Φ20μm with high consistency,and the assembly accuracy of the components can meet the requirements.