Research On Synthetic Aperture All In Focus Imaging Algorithm Via Foreground Removing

As an important branch in the light field computation imaging,the synthetic aperture imaging algorithm based on multi view to receive light field information overcomes the limitation of traditional single view imaging mechanism.When the target is occluded,the synthetic virtual large aperture has very shallow depth of field,which can effectively occluded the foreground and achieve the "perspective" detection of the occluded target.But the synthetic aperture imaging algorithm adjusts the depth of focus through the relative depth factor in the imaging model.When we detect the unknown scene,we can get the focusing factor range through several experiments,thus affecting the timeliness.Aiming at the above problems,this paper proposes a light field information specified depth surface focusing imaging algorithm based on the calculation of the disparity range objects or according to EPI edge,and the disparity range as focusing factor,through the 4D-EPI shear transform to the specified depth on the surface of focusing imaging.However,because of the synthetic aperture focusing imaging algorithm does not remove the rays from the occlusion,the reconstructed target is blurred and the peak signal to noise ratio is not high.In recent years,domestic and foreign scientific research teams have done some research work on improving synthetic aperture imaging algorithm for reconstruction of target quality,and have achieved some results.However,there are some problems in these methods.For this reason,this paper proposes a method to remove the foreground occlusion in various perspectives.First,the depth range of the scene is estimated according to the edge features of EPI.The foreground edge features are extracted and diffused according to the parameters specified in the rebuilt surface,thus the foreground occlusion corresponding ray is determined,and it is labeled and removed.However,the synthetic aperture imaging algorithm based on foreground removing can only focus imaging on the specified depth.Because of the synthetic virtual aperturebecomes larger and the depth of focus is shallower,the algorithm can only focus part of a single target or target,and the remaining scenes are still blur.Aiming at the problem of lack of multi-targets detection capability and insufficient image information at a certain depth for the above algorithms,a novel synthetic aperture all-in-focus imaging algorithm based on foreground removing is proposed in this paper.This method is mainly based on the scene depth information at various view.The scene is layered in various perspectives and only the specified regions of interest are reconstructed.After focus measuring in each layered reconstruction image,it is preferred to store in the stratified order.Finally,we use the selective measurement based on depth information to calculate the weighted values of the pixels in each position to get the weighted average of the pixels corresponding to the optimal focus measure,and generate a all in focus image with all pixels are clear.By designing experiments on more authoritative data sets in the industry,the experimental results of this method are compared with the results obtained by more comprehensive methods of the industry,the results show that the algorithm proposed in this paper can effectively improve the detection ability of the occluded multi-targets and improve the synthetic aperture imaging quality.