Research On Super-resolution Information Acquisition And Reconstruction Method Based On Micro-scanning Technology

For the advantages like good adaptability to environment,strong penetrability and good night vision ability,infrared imaging technology is widely applied to military and civil fields in recent years.Due to the limitation of semiconductor fabrication process,the pixel density of infrared focal-plane arrays is far less than that of visible image sensor,which leads to low resolution and poor visual effect.But when using the superresolution information acquisition technology can get more information in the original scene without changing the sensor,and super-resolution information reconstruction technology can use super-resolution information to improve the spatial resolution of infrared image and improve imaging visual effect.Therefore,it is of great significance to carry out the research on super-resolution information acquisition and reconstruction method based on micro-scanning technology in the infrared field.This thesis is originated from the national nature science foundation of China,and super-resolution infrared image acquisition method and super-resolution information reconstruction technology are studied deeply around super-resolution infrared imaging.Firstly,this thesis designs and realizes the system of super-resolution visible image information acquisition based on 2×2 micro-scanning mode.Piezoelectric ceramics with high precision displacement is used as the micro-scanning actuator,and driver circuit is designed to obtain multi-frame low resolution images combing with existing imaging systems,and the motion estimation and super-resolution reconstruction of the acquired low-resolution images are verified.Secondly,on the basis of the visible light super-resolution information acquisition system,this thesis designs an infrared super-resolution information acquisition scheme based on micro scanning technology,and the driving signal for generating infrared image and controlling micro-scanning are simulated to verify feasibility.Thirdly,this thesis presents the edge preserving super-resolution image regularization reconstruction algorithm,the algorithm extracts the high valued information in the low-resolution image sequence as an important part of the super-resolution reconstruction,which can achieve the desired effect of edge preserving and improving the visual effect of the reconstructed image.Finally,the thesis obtains two groups of low resolution image sequences with sub-pixel shift in the simulation experiment of super-resolution information acquisition process,and simulation experiments of super-resolution reconstruction are carried out.The experiment results show that the average gradient,peak signal-to-noise ratio and structure similarity parameters of the image reconstructed by the proposed algorithm improved to a certain extent than that of the traditional algorithm respectively.When the proposed algorithm is applied to the low resolution infrared image sequence with sub-pixel displacement,and finally the high-resolution infrared image with clear target,clear edge profile and good visual effect can be obtained.