Research On Data Acquisition And Depth Information Extraction Of Light Field Imaging Based On Microlens Array

Three-dimensional imaging and display technology is more and more used inscientific research, entertainment, medical, military and other fields. By the huge rolein promoting economic benefit and social effect, three-dimensional imaging theoryand technology is constantly updating. Combined imaging and calculation, is a newtype of three-dimensional imaging technology. Characteristics such as the simpleoptical system, low cost of hardware, and flexible data post-processing, light fieldimaging got the attention of many researchers and research institutions all over theworld.After an in-depth study at the basic principle and research status at home andabroad of light field of imaging, around the key issues such as imaging method, datacharacteristics, data encoding, depth information extraction and data reconstruction,papers has carried on a comprehensive research work in type of micro lens opticalfield imaging. The main work is as follows:Firstly, papers made a serious research at the methods and theories in depthinformation extraction, introduced the depth information acquisition methods both inactive and passive. And the five steps of stereo vision method in passive depthinformation acquisition method are highlighted. Stereo matching in stereo visionmethod to extract the depth information is introduced in detail, respectively to basedon the area and feature of stereo matching, and made a comparison, pointed out theadvantages and disadvantages of each.Then, analyzed the model parameters of micro lens optical field imaging, theimaging characteristics were studied for micro lens array is placed in the two differentposition, and the matching process of physical parameters in micro lens array andCCD array was studied to provide guidance of the data acquisition and algorithmdispose.Then, papers analyzed the characteristics of CCD image using the light fielddata obtained by Lytro camera, and established a mapping relation at the CCD andmicro lens array using imaging model. Based on those, we puts forward a kind of light field data coding method, using the coding method, the one-to-onecorrespondence relationship between pixels and micro lens were determined.Finally, using the front results of the coding algorithm, and based on gettingimages of different angle of view with a series of sub image, this paper proposes amore effective method to extract depth information against light field imaging, whichmainly based on the thought of image segmentation, color segmentation, the gradientprojection and image fusion. This method can successfully used for depth informationacquisition in disparity map obtained by Lytro camera.Within the framework of the contents of the above research, papers uses Matlabwrote the algorithm of light field data coding, image segmentation, depth informationextraction, digital refocusing and so on. Using those algorithms to process light fielddata, the results shows that the proposed coding and depth information extractionalgorithm in this paper is reasonable and can be used in image segmentation anddepth information acquisition accurately，and compared with the existing algorithms,the accuracy performance of depth information is more superior, which provides astrong guarantee for the precision of3-D data reconstruction.