Research On Improved Integral Imaging Reconstruction Based On Sensing Structure

Integral imaging is a three-dimensional imaging technology with full parallax,full color,and continuous viewpoints.Integral imaging does not need to any wearable devices and the imaging structure is simple,without coherent light and other stringent requirements.This makes 3D integral imaging technology considered to be the best 3D imaging technology.However,due to the limitations of material and imaging environment,conventional integral imaging based on microlens array in three-dimensional reconstruction has the problem of distortion.To solve this problem,two methods of improving the reconstruction quality based on the integrated imaging technology are proposed from two aspects: image algorithm and imaging system.Aiming at the problem of image reconstruction quality reduction caused by occlusion in three-dimensional synthetic aperture integral imaging system,this paper proposes a computer reconstruction enhancement technique.First,this paper discusses the feasibility of stereo matching for depth estimation in synthetic aperture integral imaging system.Then we propose to estimate the depth information of the scene by using the stereo matching method based on the minimum spanning tree,and the position of the occlusion in the scene is detected by the depth information.Then,the image inpainting technology are used to repair the occlusion area.Compared with the previous methods,the proposed method has two advantages: the stereo-matching algorithm based on the minimum spanning tree is more accurate than the block matching method,and it is more accurate to detect the occlusion;The use of pixel restoration and exemplar-based image inpainting technology to remove the occlusion region,breaking the limitation of the distance between the occlusion and the target regions.Aiming at the problem of reconstruction quality reduction caused by the shortcomings of traditional integral imaging system,this paper presents a three-dimensional integral imaging system based on non-uniform distribution sensors along the circle.This non-uniform distribution sensor system not only has the advantages of the perspective enhancement,but also has the advantages of enhanced system resolution.In addition,this paper proposes a non-uniform distributed sensing system to solve the problem of flexibility reduction of the system in the plane synthetic aperture integral imaging system.Finally,the computer reconstruction algorithm of traditional integrated imaging system is improved to fit the proposed system.Geometric projection method is used for image reconstruction,and the image distortion is avoided caused by non-uniformly distribution sensing.