Research On Super Resolution And Occlusion Removal In Synthetic Aperture Integrated Imaging

With the increasing demand for display quality,three-dimensional display technology has attracted more attention from researchers.Due to the advantages of providing full parallax,continuous viewpoint,and no need for auxiliary equipment,integrated imaging technology has become one of the most promising 3D display technologies currently.However,due to the limitations of the physical structure of the lens or the acquisition equipment,traditional integrated imaging techniques often have low resolution in reconstructing images.Moreover,when the target object is partially occluded by an obstruction,it not only fails to recognize the full view of the target object,but also causes the pixels of the obstruction and the target object to overlap in the reconstructed image,reducing the quality of the reconstructed image.In order to solve the problem of low resolution of the reconstructed image,a super-resolution method based on Bicubic interpolation is adopted.First,a 3D scene and synthetic aperture integrated imaging acquisition system is built in the 3DS MAX simulation software to collect the 3D scene,obtain the low resolution element images,and then splice the low resolution element images in order to obtain the low resolution element image array,Then Bicubic interpolation is applied to the obtained low resolution element images,and high resolution element images can be obtained by this method;On this basis,in order to further process the obtained high-resolution element images,the high-resolution element images are spliced into a high-resolution element image array in a certain order,and the high-resolution element image array is reconstructed to obtain a Iterative reconstruction stereo image.The simulation results show that the reconstructed image resolution processed by this method has been significantly improved.In order to remove the influence of occlusion on the Iterative reconstruction of the target,a semi global stereo matching based occlusion removal method is adopted.Firstly,the 3DS MAX software is used to build a synthetic aperture integrated imaging acquisition system,collect the element images with occlusion,digitally refocus the acquired element images,and extract the depth information of objects according to the focusing and defocusing characteristics of 3D objects at different depths,combined with the energy gradient function;Secondly,based on the semi global stereo matching algorithm,the disparity maps of each element image are calculated.The depth information obtained by the digital refocusing method is used to set the classification threshold for the target object and occlusion object,which can be used to distinguish them;Mark the occluded areas in the element image with white to remove the occlusions;Then,according to the element Inpainting technology,select a series of element images that are in the same line with the element image to be repaired as the sample images to repair the occluded areas;Finally,Iterative reconstruction is performed on the repaired element image array.The simulation results show that the PSNR of the reconstructed image using this method reaches 43.1286dB.