Research Of A Robot For Hull Surface Weld Identification And Milling Based On Laser Radar

In order to reinforce the unmanned working on the hull surface,a system based on the surface weld identification with laser radar and milling robots was used to take place of manual operation.It could not only improve the automation level of shipbuilding,but also reduce the difficulties in hull surface working.Based on magnetic adsorption structure,a robot for surface treatment was designed to meet the security demands in shipbuilding.The laser radar could identify and locate the weld which was visual through the technology of 3D surface reconstruction.The identification precision was testified and the effectiveness of surface reconstruction algorithm was demonstrated.Firstly,this thesis introduced domestic and international researches about robots for surface treatment and laser radar scanning technology.It also proposed a system scheme including weld identification,surface reconstruction and milling with laser radar.Secondly,considering the live environment for surface treatment and practical application,this thesis divided the whole system into mechanical part and electric control part.After completing the analysis of functional demands and performance evaluation,it designed the mechanical structure.Besides,it completed the design for integral structure and selection for key parts.For several sensitive parts,finite element simulations were conducted to testify their strength properties.Thirdly,it introduced the selection for electrical hardware and the design for software system respectively.The parts,including IPC,motors,laser radar,transducer and limitation sensor,were chosen.In order to meet the functional demands of system,it completed the integral design for software control,and developed the software for weld identification and milling in the MFC framework.Next,the thesis investigated the weld identification and the algorithm for point cloud reconstruction in the system.It introduced the working principle of laser radar.After the researches about extracting the weld data and clustering filtering algorithm,the verification was conducted.The surface reconstruction algorithm for 3D point cloud was analyzed.Besides,the advantages and disadvantages between curve fitting and triangle surface reconstruction were compared.The algorithm for K-point neighborhood search and for triangular mesh construction based on region growing was investigated,which provided the algorithm supports theoretically.Then,according to the integral design and all subsystem design,the thesis completed prototype assembly and its debugging.Experimental methods and inspection methods were designed,and then the tests were conducted.The feasibility of the algorithm was demonstrated and the error resources were analyzed.Finally,the thesis made a systematic conclusion and proposed some suggestions for researches and improvements in the future.