Research And Application Of Key Technologies Of Go Dropping Simulation System

With the improvement of people’s living standards,the market of go game system is gradually hot.The current go game integrates the current high-level game algorithm,which can help the player complete the high-level game and resumption.At the same time,it can avoid the player’s direct contact with the electronic screen and improve the player’s game experience.However,the go game system in the market is generally developed based on industrial manipulator.The system is large and cost much,which is not easy for popularization.At the same time,for maintaining the stability of visual algorithm,the position of each module in the go game system is relatively fixed and the system flexibility is low.In view of the above problems,this thesis makes an in-depth study on the key technologies in the links of go piece recognition and motion execution system modeling and design,and constructs a go drop simulation system through algorithm design and model simulation,which simulates the player’s drop through laser guidance,so as to complete the process of automatic game.The mainly contents of this thesis include:（1）two kinds of go piece recognition algorithms are studied,The first algorithm completes the positioning and segmentation of the standard chessboard image in multiple scale spaces based on the chessboard positioning mark.According to the threshold detection and corner detection results of the standard chessboard image,the positions of the white chessboard and the black chessboard are determined respectively.Then,after several rounds of morphological processing,the uneven illumination area is determined,and the chess classifier is trained based on support vector machine to detect the chess pieces in this area;The second algorithm determines the pixel position of the chessboard in the image based on target detection,reconstructs the chessboard area image and completes the chessboard positioning,and finally determines the specification and state of the chessboard through probability analysis.We can make choices according to the respective advantages of these two algorithms.（2）The position error compensation method of the end of the motion actuator is studied.Firstly,the corresponding motion actuator and related functions are designed.The motion actuator is constructed by using laser transmitter and high-precision digital actuator,and the telescopic joint is used to simulate the laser light path to complete the kinematic modeling.Then,on this basis,a motion end position error compensation method based on perspective transformation is proposed to obtain the mapping relationship between the ideal value and the actual value of joint variables,The error compensation of the remaining position points at the moving end is completed through simulation calculation.Based on the above research contents and the functional requirements of the go placement simulation system,the development of each functional module is completed,and a complete game system is constructed.Finally,the effectiveness of the chess recognition function of the vision module and the error compensation method of the motion control module are verified by experiments.The experimental results show that the average detection accuracy of the two chess pieces detection algorithms reaches 98.96%and 99.95%respectively under uniform illumination,and the effect is good.However,there is still much room to improve the detection effect under uneven illumination.