| With the explosive growth of the Internet in recent years,mobile devices and services are also experiencing multiple growth,and behind such a stable transmission of such a large number of data streams is a wireless transmission network system similar to a cellular structure constructed by one base station.The antenna board is an important part of the base station in the wireless transmission system.Most of the production process is basically automated.However,in the production requirements such as product assembly and circuit board wiring,the logic requirements are large due to the large degree of freedom of operation.High,it is difficult to achieve fully automated production,manual operations must be introduced,and manual operations will avoid problems such as operational errors and inefficiencies.In response to this problem,the project team proposed to use AR(augmented reality)technology to guide workers wiring,achieve the effect of man-machine integration,improve production efficiency and reduce the rate of defective.In order to use the AR technology to guide the wiring,it is necessary to achieve a good virtual and real fusion effect,that is,the virtual wire harness should be registered in real time with its real space position,and the wire harness acts as an antenna board object,as long as the antenna plate can be tracked in real time.The virtual antenna board and the virtual antenna board can be combined with the virtual antenna board to solve the problem of the alignment of the wire harness and the real space.In view of this problem,this paper mainly conducts in-depth research on the pose tracking problem of the antenna board in the AR auxiliary wiring project.The specific research work is as follows:(1)Research and analyze the characteristics of manual wiring in industrial scenes,and based on these characteristics combined with augmented reality technology,design a pose tracking structure suitable for the project,establish a coordinate system based on the markless tracking system,and analyze the antenna position in the tracking process.Mathematical model of posture.A solution for initial registration has been proposed.(2)The use of edge contour features of the occluded area on the antenna board for tracking is studied,and two solutions are proposed.One is to use several ROIs manually selected on the antenna board and utilize The edge shape extracted in the region isshape-matched to solve the pose,and the template pool method is used to solve the matching problem when the antenna plate is flipped.The second is to extract the overall outer contour of the antenna plate by using the neural network structure.The three edges are straight lines,and the pose is solved by the polynomial regression fitting method.(3)The tracking of the poses by the feature points on the antenna board is studied,and the feature points are tracked by the KCF-LK optical flow method,which avoids the feature point extraction,matching and removal of mismatching operations every frame.The speed at which the algorithm runs.Based on the pose tracking architecture designed in this paper,experiments show that the above two types of algorithms basically meet the project requirements,and have certain anti-blocking and anti-lighting properties.Among them,the algorithm based on feature points has certain advantages in speed,and there are fewer steps to be manually processed.However,as the wiring steps are deepened,the matching feature points will be gradually reduced,and the robustness of the algorithm is reduced.Combine the previous algorithm to solve the pose. |
PDF Full Text Request