| In the precise micro-operation process,the operator's human eyes and human hands require extremely high positioning accuracy and operation precision,while the micro-operation tool end and micro-operation target are often very small,so that the human eye and the human hand are constantly observed and operated repeatedly.It is easy to fatigue during work,and it is difficult to locate the target position with high precision and high precision.When the positioning operation is performed,the human hand is easily affected by jitter,operational fatigue,proficiency,etc.,resulting in inaccurate position coordinates of the placement,and it is necessary to repeatedly move the positioning and reduce the precision and efficiency of micro-operations even caused the task to fail.Aiming at the characteristics of human eye and human hand fatigue,difficult to maintain high-precision positioning,the human eye assisted positioning and compensation system and the manual assisted positioning and compensation system for micro-operation are designed to enhance the intelligent auxiliary function of the micro-operating system.The thesis introduces the human eye line-of-sight assisting technology and machine learning prediction technology into the micro-operating system.At the same time,it studies the human-eye assisted positioning and compensation strategy,the human-assisted positioning and compensation strategy,reduces the operational burden of the human eye and the human hand,and improves the micro-improvement.Operational accuracy and efficiency.The thesis mainly includes the following aspects:Firstly,the collection and analysis of hand-eye characteristic data based on experiment;through the experiment,the data of the movement mode such as linear motion,polyline motion,arc motion,positioning operation,etc.of the human hand in the micro-operation are collected and analyzed,and the fatigue is caused according to the human eye and the human hand.It is difficult to maintain high-precision positioning and other characteristics,respectively,to establish a human-eye assisted positioning and compensation system based on line-of-sight positioning,and a human-assisted positioning and compensation system based on the SVM(Support Vector Machine,SVM)prediction model.Secondly,use VS2015,VS2017,OpenCV3.1.0 for software development of human-eye assisted positioning and compensation systems,using Python and machine learning module sklearn to complete the manual assisted positioning and compensation system..The hardware scheme and software scheme of the human eye assisted positioning and compensation system are designed,and the human eyesight positioning function is realized by combining the system calibration.The SVM is used to construct the positioning error prediction model for vertical and horizontal directions of human hands.The manual assisted positioning and compensation system are realized,and the prediction results of the model are analyzed experimentally.Thirdly,the human eye assisted positioning and compensation strategy and the human hand assisted positioning and compensation strategy are designed and implemented.The accuracy of human eye positioning is tested experimentally.It is realized by human-computer interaction,secondary auxiliary positioning and auxiliary positioning frame.Auxiliary and compensation for human eye line of sight.The method of predicting and adding an image micrometer with a SVM model to manually input the compensation error value realizes the compensation of the manual positioning error.Finally,the micro-operation experimental platform was built to complete the secondary development of the micro-operating system,and the micro-visual calibration experiment was carried out to measure the measurement ratio k between the pixel distance and the real distance.The human-eye and human-assisted assistance was performed using the micro-operation platform.The positioning and compensation experiments verify the feasibility and effectiveness of micro-operations combined with human eyes and manual assisted positioning and compensation systems. |
