Research On Segment Robot And Vision System For Shield Machine

The assembly of the tube is a very important step in the construction of the shield tunnel.It has played a role in protecting the excavated tunnel,preventing the settlement of the tunnel and preventing the seepage of the groundwater.At present,most of the domestic pieces of the assembly,mainly use the piece assembly machine,the whole process of the assembly and assembly of the pipe assembly need artificial close range manipulation of the assembly machine,so the speed and precision of the assembly of the pipe is mainly dependent on the proficiency of the workers.In order to improve the efficiency and accuracy of segment transportation and installation,the following researches were carried out in this paper.Firstly,the process of assembly and assembly is studied,and the design of the structure and visual system of the piece assembly robot is proposed.The driving force of the key components in the assembly process is measured by ADAMS simulation,and the rationality of the structure design of the assembly robot is verified.The control model of the hydraulic cylinder and the motor of the rotary motion is established.The joint simulation of AMESim and Matlab is used to control the displacement of the oil cylinder of the translation motion system.On this basis,a displacement following system based on the three power is designed,and the following effect is verified by simulation.Secondly,in order to solve the problem that the location of the tube is difficult to be identified accurately when the noise of the single recognition algorithm is more noisy,this paper studies the fast tube recognition system based on YOLO,and uses the 106 layer convolution depth learning network to learn the slice image.The experiment shows that the accuracy of the recognition algorithm is 93.3%,and the average processing speed of each image is 2.78s,which proves that the system can meet the requirements of rapid assembly.Thirdly,in order to accurately detect the position of the bolt,this paper studies the position of the pipe bolt relative pipe,and puts forward the location extraction algorithm based on the feature point matching.The experiment proves that the accuracy rate of the feature point matching algorithm is 96.4%and the average operation time is 3.83s.Experiments show that the total recognition time of location extraction by camera is 6.61s.With the position control system,the segment assembly robot can capture the tube in 1 minutes,and it takes 5 minutes to use the manual positioning,which proves that the use of the visual system can improve the efficiency of the work.Finally,this paper designs a double eye distance measuring system for the shield machine segment assembly robot,and establishes the coordinate transformation relationship between the bolt,the camera and the piece of the assembly robot.The accuracy of the ranging system can be calculated and verified through calibration experiments.The research shows that using the feature point extraction and binocular range finding system can extract the position coordinates of the pipe bolt with a lot of background noise and dark light,and combines the structure design and the control system design of the piece assembly robot,which lays the foundation for the development of the auto piece assembly technology in the future.