Research Of Virtual And Real Occlusion Technology And Experimental Verification

Augmented reality technology is one of the technologies that are currently researched and applied in the advancement of intelligent technology and application.And the augmented reality system is one of the key research objects in the field of intelligent manufacturing.In the augmented reality system,how to restore the correct virtual and real occlusion relationship has become one of the research key points of augmented reality technology.Based on the analysis of the commonly used virtual and real occlusion techniques,a better accuracy occlusion technique is proposed.Four parts have been studied in this thesis for the virtual and real occlusion technique,which is the background contour of the object,the judgement of virtual and real occlusion relationship,the occlusion culling,and the virtual and real occlusion processing in the augmented reality assembly.Firstly,the background modeling and the shadow removal are improved and optimized for the acquisition of the background contour of the object.For the problem of the PBAS method for background modeling,that is the judgment threshold and the background model update rate require a long calculation time in adaptive update.An improved background modeling algorithm is obtained based on the update formula of threshold and update rate is improved.The example application experiment shows that the improved background modeling method can can obtain a more accurate real object contour.Secondly,for the problem of shadow removal of background contour,the chromaticity and brightness of the image have beed studied.An improved shadow removal method is obtain based on the combined of the maintain original shadow judgment of(r,g,I)and the brightness change rate ?I.The shadow removal method is used for detecting and removing shadows.The example application experiment shows that the improved shadow removal algorithm can obtain better background contour shadow removal effect than the(r,g,I)shadow removal method.Thirdly,for the processing of virtual and real occlusion illusion and culling,firstly the depth distribution method has been improved based on the depth gradient hypothesis.And then the improved depth distribution method and the pixel traversal method are combined to optimize the occlusion relationship judgment algorithm of virtual and real objects.The judgment of virtual and real occlusion is not only quick but also accurate.The template cache technology in OpenGL is used to complete the virtual and real occlusion culling operations,thereby the correct virtual and real occlusion relationships is restored,and the virtual and real occlusion problem has been solved.Experiments show that the proposed algorithm has high effectiveness in solving the virtual and real occlusion problems.The example application experiment shows that the improved algorithm proposed in this thesis has a good application effect in solving the virtual and real occlusion problems.Finally,experimental verification study is carried out by a simple mechanical augmented reality assembly problem.For the virtual and real occlusion problems in the assembly,combining the virtual and real object contour depth figure and the improved virtual and real occlusion processing method,the contour depth figure is used to judge the depth relationship,and the pixels that meet the requirements are redrawn,and a correct and effective virtual and real fusion assembly scene has obtained.