Diffractive Imaging Based On Computational Imaging

DOE has the characteristics of light weight and flexible design,which can meet the application requirements of light weight and integration However,DOE has severe dispersion characteristics.When it deviates from its design wavelength,it will bring large-scale dispersion blur to the image.Hence,it is not easy to apply to broadband imaging systems.Aiming at the DOE band’s defects,this dissertation introduced computational imaging technology to combine optical design and image algorithms.This dissertation got a vivid image from a single-chip diffraction element in the visible light band from 415 to 685 nm.In order to obtain the target PSF,this dissertation constructed the Point Spread Function(PSF)model in optical design,encoded the diffractive element structure,and carried out numerical optimization.This dissertation used an image pyramid in the image restoration algorithm and restored the blurred image at multiple scales.Simulation showed that when the traditional phase Fresnel diffractive lens is imaging at 415nm～685nm,the frequency is only 2.6lp/mm when the average MTF value is 0.1.In contrast,the diffractive computational Imaging designed in this paper can reach 113.3lp/mm under the same conditions.This dissertation demonstrated that diffractive imaging technology based on computational imaging could obtain clear images in a wide wavelength band while maintaining a thin and light physical structure.The main research contents of this dissertation are as follows:(1)This dissertation analyzed and discussed the optical field transmission theory of diffractive computing imaging,the principle of phase control of diffractive elements,and the diffraction image degradation model.Some typical restoration algorithms are listed and introduced.Moreover,this dissertation discussed the image quality evaluation method.(2)The achromatic method of the diffractive element is studied in this dissertation.Furthermore,this dissertation constructed a PSF-oriented diffractive element optimization model.The optical field phase is adjusted by coding the step height of the diffractive element.This dissertation also introduced a numerical optimization method based on particle swarm optimization so that the spectral PSF of the diffractive element remains consistent for achromatic purposes.By calculation,the cut-off frequency at 0.1MTF under visible light is 8.6lp/mm,which is higher than the 2.6lp/mm of the Fresnel diffractive lens.This dissertation processed the method of ion beam etching,8mm,20 mm,and 40 mm coded optimized achromatic diffractive lenses.Furthermore,this dissertation carried out the actual star point target,and resolution target imaging tests on the 8mm diameter coded optimized achromatic diffractive lens.And this dissertation discussed the relationship between the aperture and imaging clarity.This dissertation verifies the achromatic performance by the color fidelity test.(3)This dissertation studies a suitable image restoration algorithm according to the designed coding optimized achromatic diffractive lens’ s imaging characteristics.Coding optimization Achromatic diffractive lens has a large blur kernel size,which is easy to introduce artifacts and noise.For this problem,refer to the image pyramid method,denoising and deblurring at low scales and restoring to high scales as cross-scale prior information to gradually restore the image detail.This dissertation compared the performance of the cross-scale prior algorithm to demonstrate its effectiveness.This dissertation combined with the coding optimized achromatic diffractive lens for imaging and evaluating the image quality quantitatively.The result was much higher than the imaging quality of traditional diffractive elements.Finally,the experimental analysis of the computational imaging process of the achromatic diffractive lens is carried out,which shows the effectiveness of the method in this paper.