ABSTRACT
In the process of the virtual and the real occlusion coherence processing of digital film and television, the spatial coherence relationship between the real scene and the virtual scene will directly affect the realism-rendering effect of the virtual-real fusion scene. So, the occlusion integration process of the virtual-real fusion plays an important role in the seamless synthesis of a digital movie screen. In order to reduce the complexity of the virtual and the real occlusion consistency process and improve the universality of the virtual and the real occlusion consistency method, a virtual and real occlusion consistency processing method based on hierarchical segmentation is proposed. This method uses three steps to realize the dynamic occlusion rendering of the virtual and real scenes through the experimental analysis. They are hierarchical clustering of depth data, optimization of the layered texture maps, and the consistency rendering of the virtual and real occlusion. Through the experimental results and mathematical analysis, we can see that the virtual-real fusion effect obtained by this method is better for dealing with the virtual and real occlusion consistency. Also, the ideal virtual and real occlusion effect has been achieved, which is totally consistent with the human visual perception mode.
