Virtual Viewpoint Temporal And Spatial Motion Estimation Of Integrated Imaging

Integrated imaging is a technology that records and displays spatial 3D information and is one of the research hotspots in the field of 3D display.In the integrated imaging recording stage,the camera array can both obtain very high resolution elemental images and at the same time reduce crosstalk caused by the use of microlens arrays.However,the number of images acquired through the camera array is small,and the loss of details easily occurs.At the same time,increasing the number of cameras also increases costs.Therefore,it is necessary to study how to obtain a large number of viewpoint images through the camera array to solve the above problems.What we do is to acquire a small number of viewpoint images on the basis of the integrated imaging,and then make an estimation of the virtual viewpoint of the image so that any virtual viewpoint can be estimated based on a given viewpoint.The quality of image segmentation has a very important influence on our acquisition depth.In order to obtain high-quality depth images,this paper proposes an optimal entropy threshold segmentation method based on an improved genetic algorithm to obtain depth images because of the depth image,because the quality of the depth image estimate of the virtual viewpoint has played a vital role.Improved segmentation method it can be very fast for global search,it can not only realize the search for groups,as well as among the individual exchange of information,it makes all of the individuals have the chance of a optimization,in this way,to ensure that the entire group is under the condition of the evolution of evolution.Then,an improved bidirectional DIBR technology to obtain images of the virtual viewpoint,on the basis of the traditional two-way DIBR increased the interpolation of image,to improve the image map precision,at the same time on the depth of interpolation processing reference chart and graph to implement two-way mapping mapping precision of the left and right side of the virtual view image.Finally,by studying the relationship between the virtual viewpoint and the reference viewpoint,this paper presents a virtual viewpoint movement model.By establishing this mathematical model,we can estimate any virtual viewpoint in the camera array.The requirement that a large number of virtual viewpoint images can be obtained by capturing a reference image with a small number of camera arrays can achieve the number of images obtained by the microlens array and avoid its defects.