Study of Multi-planar 3D Integral Imaging Sensing and Visualization

We present a 3D multi-planar integral imaging architecture that uses multiple arrays of image sensors or a moving sensor distributed along its x-y-z axis. A corresponding algorithm is developed for 3D volumetric computational reconstruction. Compared to conventional integral imaging, multi-planar integral imaging can reduce 3D image noises to enhance the accuracy of 3D depth detection. Experimental results indicate that the proposed configuration improves the 3D reconstructed image quality at the point of theoretical maximum performance by 0.4346 and 12.42% via SSIM and PSNR, and reduces the 3D depth detection error by 2.5%.