Research On Super-resolution Reconstruction Of Sub-pixel Imaging Base On Linear CCD

Human civilization has developed into the information age. Remote sensing technology can efficiently obtain geographic and spatial information which occupies 80 percent of the total information in human activities. Many countries and companies are developing remote sensing technology, in which the research of high resolution image is the core and hot topics. The quality of the image is increased limitedly only by changing the software or improving the algorithm of software. The sub-pixel sampling technology combines advantages of hardware and software, which can improve the image resolution efficiently with low cost.Sub-pixel technology is introduced and analysised. Simulate the sampling process and image reconstruction in many classical sub-pixel sampling modes. Different sampling modes have different characteristics in terms of the technical requirement, efficiency, and cost and so on. Taking into account all aspects, the super sampling mode has advantages of low cost, high efficiency and facile realization. Finally, introduce the approaches to realize the dislocation between the two linear array detectors in super sampling mode, which is namely cal ed sub-pixel sampling mode.For super- mode, this dissertation introduces and simulates the conventional methods of the image super-resolution reconstruction. These methods have the same characteristics, which results in the shortcoming of non-continuous gray values. In this dissertation, a new sub-pixel image reconstruction method is proposed to overcome this kind of shortcoming, which is named by all- interpolation method, which is proved to effectively overcome the problem of gray discontinuity.Based on the main idea of all- interpolation method, it is improved by combing with different interpolation. In terms of the fuzzy edge caused by the mean processing, all- interpolation method combing with adaptive edge detection is proposed. Analyze the advantages and disadvantages of different reco nstruction methods as well as the causes of all kinds of advantages and disadvantages, in aspects of HVS(Human Visual System), objective evaluation index and time complexity.Experimental results show that all- interpolation method and its improved methods can reconstruct sub-pixel images quickly and effectively.