Computational Imaging Base On Microlens Array

Three dimensional imaging equipment now used need to be used in certaincircumstances and conditions due to its complexity and bulky. New devices ofthree-dimensional collecting are a trend of three-dimensional imaging. Andmicro-lens array having a characteristic capable of multi-channel imaging, the objectwill be recorded as an image array through micro-lens array. By processing this imagearray with computational imaging, we can not only restore the extractthree-dimensional information of scene, but also reduce the noise introduced byoptical element. The acquisition device can be designed to portable and light weightby using micro-lens array. Therefore, the study of computational imaging based onmicro-lens array has important significance. At present the study of computationalimaging based on micro-lens array is still in a preliminary stage in domestic. Theprinciples and applications research are relative lack. This paper studiescomputational imaging based on micro-lens array through two typical classifications,light field camera and computational integrated imaging. The main contents andresults include the following aspects:First, research the four-dimensional light field theory. Analysis the imaging processand the principle of the two types of light-field camera through this method. Analyzethe focusing principle of conventional imaging system by this theory. And on thisbasis, the light field camera imaging and re-focusing principle are analyzed. Imagingand re-focusing algorithm are constructed combining the structural features. Thesemake a foundation for subsequent research of light field camera.Second, we research on computational integrated imaging reconstruction methodsby geometrical optics. Image reconstruct based on viewpoint and based on depth areanalyzed respectively. And experiments are conducted, two algorithm ofcomputational integrated imaging are constructed. These works carry out a technicalsupport for the future work.Third, range the objects by using multiple different perspectives sub-images withmulti-baseline stereo ranging method. We analyzed the ranging principle ofcomputational integrated imaging. Construct the algorithm, and verified the feasibilityof our algorithm by simulating. Fourth, we proposed an extraction method based on computational integratedimaging. Targets can be extracted in a complex scene. We study the principle of theprocess of object extracting by computational integrated imaging. Then we verifiedthe effectiveness and practicality of this method by simulation. This methodintroduces computational integrated imaging to object extraction. And this methodprovides a new solution for objecting extracting of multi-objective in a complexscene.