Research On AR Tracking And Registration Method For Cable Assembly Guidance

Augmented reality is a new type of human-computer interaction technology.It superimposes virtual information into real scenes in real time through a computer to provide users with virtual and real fusion image guidance.Among them,the tracking and registration technology is one of the key technologies of the AR cable assembly guidance system.It is the basis for the "seamless" fusion of virtual guidance information and assembly site,which directly determines the performance and user experience of the AR cable assembly guidance system.Therefore,research on tracking and registration has important theoretical significance and engineering value.The main research content of the thesis are as follows:1.The principle of template matching algorithm,the basic model tracking and registration algorithm and the basic idea and workflow of AR tracking and registration method are studied.This provides a theoretical basis for the research of AR tracking and registration method.2.Research on the initialization method for tracking and registration of cable assemblies based on LINE-MOD template matching.Aiming at the problem that the traditional LINE-MOD template matching method can only identify rigid objects at a fixed scale,and the existing tracking and registration method requires artificial assembly guidance coordinate system initialization.First,this method collects multi-view cable assembly images in the offline stage,gradient direction extraction and quantization and gradient direction expansion.Then at the online stage,the template matching is used to realize the identification and initialization of the cable assembly.This method introduces depth information to obtain improved LINE-MOD template matching.Finally,experiments show that this method can identify the cable assembly in multi-scale and realize the automatic initialization of the assembly guide coordinate system.3.The tracking and registration method of the cable assembly based on the edge model is studied.For the existing unmarked tracking and registration method for complex assembly background.There are poor robustness and mechanical parts with insufficient texture features on the surface.It is difficult for the existing tracking and registration method to extract stable and strong discriminating feature points,resulting in tracking and registration method is invalid.First,this method analyzes and samples the three-dimensional model of the cable assembly during the initialization process,and derives the formula for updating the pose parameters.Then the normal distance iterative least squares pose estimation method is used to estimate the pose of the cable assembly.Due to the influence of the complex assembly assembly background on the robustness of the algorithm,pose optimization and pose verification of the algorithm are required to improve the effect of tracking cable assemblies.Finally,experiments show that this method can effectively eliminate the impact of complex assembly background on the robustness of the algorithm,avoid defects that rely on texture features,and improve the practicality of the system.4.The method proposed in this thesis is verified on the existing AR cable assembly guide prototype system of the research group.The experimental results show the feasibility of the theory,method and key algorithm proposed in this thesis.