Resolution Research Based On Integral Imaging Reconstruction

Compared with2D images,3D image technology is a great progress in image generationand display area.3D image technology has been applied to many fields, such as film, print,commerce, medicine and so on. It has great academic and practical values for the research of3Dimage technology.Integral Imaging is one of the3D image technology using micro-lens to record and displayimages. Since no special apparatus is needed for observers and no specific requirement isneeded for the recording environment, it is attracting more and more attention and be regardedas one of the most perspective3D display technologies for the3DTV.At the moment, the major factors that restrict the development of this technology arelimited resolution for3D display, limited depth range for corrected reconstruction and limitedobserving range for the viewers. On this aspect, although many research works has been carriedout, the results are not consistent with the practical results. The major problem existed in theanalysis methods restricted by traditional2D images.Unlike a traditional2D image, an Integral Image is composed with many small2DElement Images (EI). The viewer perceives the reconstructed3D optical model of the objectscene rather than each distinctive EI. Unlike previous research method, our work address aboveproblems on the reconstructed3D object space rather than each distinctive2D EI.Depth extraction from the planar recorded Integral Images is the basis of3D construction.According to the characteristics of Integral Imaging, a special method of viewpoint imagesextraction was applied at the beginning. A viewpoint image is corresponding to a2D projectionof the3D scenes along a particular angle. Different viewpoint images corresponding to differentprojection angles. A depth equation is derived through geometrically analysis. Multi-baselinetechnology is adopted to calculate disparity information from different viewpoint images.interpolation algorithm is further used to improve the precision of the depth calculation.Subsequently, using a computer generated software developed by ourselves, we simulatedthe whole Ⅱ recording and displaying processes. The reconstructed position and range for asingle point in the object space is obtained according to its optical principles. The reconstruction resolution in ideal condition is analyzed. Further analysis has also carried out on the effect of thelimited recording resolution and micro-lens parameters.Following conclusions has been obtained: in ideal conditions, any point recorded by anIntegral Imaging system could be reconstructed within a fixed area near the original place. Thesmaller the aperture of the micro-lens, the smaller the reconstructed area. Limited recordingresolution might affect the reconstruction results, causing an enlargement of the reconstructedarea. The critical point for the recording resolution is related with object depth. The further anobject point is, the higher the recording resolution is required. Similarly results also can beobtained for the micro-lens parameters.Our research results are consistent with the practical results, which will provide a goodfoundation for further research and practical application.