Research On Super-resolution Imaging System With Bionic Multi-apertures

With the rapid development of photoelectric system,the optical equipment with high spatial resolution and large field of view has become the goal of users and researchers.There is a contradiction between high resolution and wide field of view.Traditional single aperture imaging system is difficult to break through this limitation.The bionic multi-apertures system has many subapertures.The images acquired by subaperture have overlapping information.Image registration and superresolution reconstruction can solve the contradiction between wide field of view and high resolution.So the bionic infrared compound eye imaging system has a good prospect and important research significance.In this paper,a multi-aperture bionic infrared optical system was designed.The system has 3×3 subapertures and uses the half-field overlapping imaging model.The total field angle is 36°×36°.Rotationally symmetric aspheric lenses were used.The basic performance parameters of the optical system were determined and demonstrated.Several structural types of infrared optical systems were discussed.The overall design of the multi-aperture bionic infrared optical system was determined.Nine sub-apertures were reduced to three optical systems based on symmetry.The optical system was designed via CODE V software.The image quality of the optical system was evaluated by MTF and point plot.According to the lens data,the corresponding mechanical structure and diaphragm were designed.The super-resolution reconstruction technology of multi-aperture imaging system includes sub-image registration and image reconstruction.In this paper,a superresolution reconstruction algorithm was established.First,the images was initially registered by Scale-invariant feature transform algorithm.The effects of light transformation and affine transformation were reduced.Then the Keren registration algorithm was used to improve the accuracy to sub-pixel level.The basic requirements for image reconstruction were met.And the reconstruction effect was improved.Finally,the infrared image was reconstructed by the structure adaptive normalized convolution super-resolution algorithm.This algorithm can retain the image details and edge information.The splicing of field of view and the improvement of spatial resolution are realized.Basic research experiments were carried out.The superresolution reconstruction of the image was realized.The feasibility of the multi-aperture bionic infrared imaging system and image processing algorithm designed in this paper was verified.The theoretical basis for subsequent application research in the fields of high-altitude detection,traffic travel and ground monitoring was laid.