Development Of Automatic Loading And Unloading System For Grinder Based On Machine Vision

Aluminum substrate is the main raw material for the manufacture of mechanical hard disks.For multi-disk hard disks,the consistency of the thickness of the aluminum substrate is an important factor affecting the performance of the hard disk.Grinding is the most important process that affects the thickness accuracy of aluminum substrates.At present,domestic related companies still use manual loading and unloading operations during the grinding stage.Manual methods have a series of problems such as low efficiency,high cost,and high scrap rate,which seriously affect the company's reliability.Continuous development.In response to the above problems,the paper developed an automatic loading and unloading system based on machine vision positioning and suitable for aluminum substrate grinding machines,which improved the automation level and production efficiency of the grinding process.The main research work of the thesis is as follows:(1)According to the grinding process requirements of the enterprise,the production environment and the structural characteristics of the grinding machine,the overall realization scheme design of the system is completed;for the special needs of system data transmission,the design of multiple devices based on Modbus protocol,Ether Net/IP protocol and Socket technology data exchange program;elaborated the overall process flow of the automatic loading and unloading system in detail.(2)Designed a Mark point recognition and positioning method based on machine vision technology,introduced the implementation process of the algorithm,and introduced in detail the pre-image processing work such as camera calibration,distortion correction,image preprocessing,etc.;by comparing the differentiation of multiple edge detections The pros and cons of the operator,the Sobel operator is used to extract the edge feature points of the image instead of the gray value to measure the similarity,the non-maximum value suppression is used to obtain a stable matching window,and the angle difference estimation method is used to obtain the initial rotation angle between the images.,The use of normalization processing to reduce the impact of light and noise;a template matching algorithm based on the image pyramid layered acceleration algorithm and the least square method is proposed to meet the high-speed and high-precision positioning requirements of the system.(3)Designed and completed the assembly and debugging of the system experimental prototype,briefly introduced the characteristics of the main components of the system and the selection process;introduced in detail the design principle and working process of the fork frame recovery,automatic loading and unloading trolley,and visual acquisition and processing subsystem.The adjustment process of the Mark point image is controlled by building a visual servo system based on fuzzy control.(4)Completed the development of host computer human-computer interaction,image processing,data processing and main control software of the system.Based on the designed experimental prototype,the system function verification experiment is completed and the experimental data and error sources are analyzed in detail.The experimental results show that the designed automatic loading and unloading system has reached the design goal and can meet the production needs of the enterprise.