Research On Synthetic Aperture Imaging Algorithms Based On Nonlinear Array

As a significant branch of light field imaging,synthetic aperture imaging technology overcomes the limitation of the mechanism of single aperture imaging system.It is able to obtain more dimensional light information of spatial objects,and then achieve computational focusing and de-occlusion reconstruction,which is a hot topic in the field of vision.Camera arrays are the main means to acquire spatial light field information.Currently,the related theories and applications mainly focus on linear arrangement mode.When the distance between the object and the occlusion in the optical axis direction is small or the arrangement space of the camera array is limited,the nonlinear array performs better in the acquisition of spatial light field information.However,the current linear array light field information processing and reconstruction algorithm can not directly adapt to the light field information processing of nonlinear array.For this problem,the paper carries out research on light field information acquisition and reconstruction algorithm based on nonlinear array,aiming at achieving synthetic aperture imaging and object information de-occlusion reconstruction in the mode of nonlinear camera array.Firstly,aiming at the lack of linear mapping relationship between space and image in nonlinear array,the rearrangement algorithm of multi-view image is deduced,which makes the rearranged image meet the requirements of synthetic aperture imaging algorithm.The performances of linear array and nonlinear array in different scenarios are analyzed by experiments and the advantages of nonlinear array are verified.Secondly,the result of synthetic aperture imaging contains not only the object information but also the occlusion information,which results in the blurred result of object reconstruction.In order to solve this problem,the synthetic aperture imaging algorithm based on removing occlusion is studied in this paper.This algorithm uses the defocus function of synthetic aperture imaging to recognize the foreground and background of the image.It realizes the automation of image labeling,and then achieves the separation of occlusion and object.Synthetic aperture imaging can get clearer reconstruction results by removing occlusion from multi-view images.Finally,the algorithm is validated in different scenarios.The results show that the de-occlusion imaging method proposed in this paper can effectively remove the blurriness in synthetic aperture imaging results and improve the quality of reconstruction.