Basic Research On The Key Technology Of Image Superresolution Reconstruction

Super-resolution reconstruction is an emerging interdisciplinary technology in optics and image processing field. It have a wide range of application in many domain, such as national defense, medical, intelligent transportation, public safety, space exploration, geographic information systems, microscopic measurement, computer vision, pattern recognition, and other fields. So research to super-resolution reconstruction has important significance.This paper introduces the research background and significance of the super-resolution reconstruction and the optical analysis base of the super-resolution reconstruction in detail. This paper also introduces Angular spectrum of Fraunhofer diffraction from the point of Fourier optics analysis; explains the appearance of Airy disk which is the system response by Fraunhofer diffraction effects from limited circular aperture lens; reveals the diffraction aperture’s limited effect to MTF(Modulation Transfer Function) by coherent light and OTF(Optic Transfer Function) by non-coherent light. This paper also discusses some key problems of the super-resolution reconstruction, analyze the imaging degradation model and studied the super-resolution reconstruction algorithm, discussed ways to improve the image resolution. Some super-resolution simulation experiments are carried on based on the algorithm, and the results of the simulation show that the used reconstruction algorithm improves the resolution significantly; several super-resolution experiments are also carried on, the results also show that they breaks the diffraction limit’s cut-off frequency of the optical imaging system and improved the resolution.This paper is organized by the following aspect:(1) Derives the OTF’s expression form by incoherent polychromatic light. Reveals the essence of image reconstruction is to solve the ill-posed inverse problem, and proposes the basic algorithm of super-resolution reconstruction.(2) The two-line simulation of super-resolution reconstruction is also carried on. The mathematical algorithms super-resolution reconstruction is deduced in detail. The results of the simulate experimental demonstrate the significant effect, which the "two-line" image is changed from undistinguishable state by the principle Rayleigh criterion to obvious distinguishable state through the algorithms of super-resolution reconstruction; it is discussed that prior knowledge is effective for improving the reconstructed image quality and reducing’the minimum distance’based on the algorithm proposed in this paper. Several factors which affect the ’two-line’ reconstructed results quality are discussed and analyzed, through these discussions it is more clearly that the theoretical analysis of the imaging system is correct and the algorithm of super-resolution reconstruction is feasible, all of this provide strong theoretical and simulation basis for the actual experiments.(3) According to the theory of the imaging system and the algorithm of the reconstruction, it is carried on a serial of experimental studies and explorations of super-resolution reconstruction. Microscopic image super-resolution reconstruction experiment at the condition of optical diffraction limited imaging system, which is not achieve the mini-resolution limited, is designed by the variable aperture size. In micro-diffraction-limited situation of the mini-resolution with the convention microscope, the image quality of the result is greatly improved to "0.13um" and it demonstrates the attempt to break the microscopic distance of diffraction limited is positive significance. The SRCT experiment is carried out based on the principle of super-resolution reconstruction. From the reconstructed results above, it can be seen clearly that the final image has higher quality, more de-blurring effective and wider extensive spectrum domain than the original image. As a results, the resolution is improved by these experiments, and achieve a super-resolution reconstruction effect.(4) Finally, two probing questions of the super-resolution reconstruction are carried out, which includes a short and a long focal length imaging system experiment, the reason that these experiments did not achieve the desired effect is also analyzed and discussed. The evaluation criterion function of super-resolution reconstruction problem is also discussed, which includes the reference image and non-reference image quality evaluation function, this paper point out that the’ Peak-Saddle-Margin’ and the ’Saddle-Point-Gray-Value’ are two appropriate evaluation criteria to the super-resolution reconstruction at current time.