Study On The Imaging And Super-resolution Reconstruction Of The Infrared Optical System

Infrared imaging is an imaging technology that converts infrared radiation emitted by objects into visible thermal images.Benefitting from the advantages of real-time,non-contact,non-invasive and simple operation,infrared imaging has been an effective technology in military affairs,industrial inspection,monitoring system,medical diagnosis etc.While due to the limitation of the semiconductor materials and manufacturing technique,the infrared image has the disadvantages of low resolution,low contrast,less details,and much noise.Therefor the quality of the infarared image is much lower than the visible image.To improve the quality of infrared image by improving the hardware of infrared optical imaging system requires complex technology and expensive cost.Super-resolution reconstruction method can achieve a high-resolution image from one or more frames of low resolution images of the same scene by computing on software.Therefore,it has important research significance and application value to study a simple and effective technology to obtain high quality infrared image by combining image super-resolution reconstruction technology with infrared optical imaging system.In this paper,we study the imaging and super-resolution reconstruction technology of the infrared optical system.Firstly we study the infrared micro-scanning optical system,and design and realize a controllable infrared micro-scanning optical system.Then the low-resolution infrared images with sub-pixel displacement can be obtained by the infrared micro-scanning system,and then the super-resolution reconstruction technology based on the infrared micro-scanning optical system is proposed,whose results are superior to those of the existing super-resolution imaging techniques for infrared images.Finally,in order to obtain high resolution infrared image with larger field of view(FOV),we study the imaging principle and characteristics of the super-large FOV infrared panoramic annular lens(PAL)imaging system,and propose a super-resolution reconstruction method for the infrared PAL image,which provided an effective method to obtain both large FOV and high resolution in the optical imaging system.The main contents of this paper are as follows:1.In this paper we propose a scheme of infrared lens micro-scanning imaging technology,and design and realize a controllable infrared micro-scanning imaging system driven by piezoelectric ceramics.In the traditional lens micro-scanning imaging system,the telescope system and the micro-scanning lens are designed and assembled separately.In this paper,a collaborative designed medium-wave infrared lens micro-scanning system is designed,which can match and optimize the telescope system and the micro-scanning lens simultaneously.By measuring and analyzing the displacements and the Modulation Transfer Function(MTF)of the optimized system,it has been proved that the system we designed can guarantee both the displacements uniformity of the whole FOV and the ideal image quality in the image plane when the micro-scanning system working.2.A frequency domain phase-based super-resolution reconstruction algorithm for multi-frame infrared images has been proposed.In this paper we integrates the phase spectrum virtue into the process of the conventional projection onto convex sets(POCS)method,then a high-resolution infrared image whose quality is much higher than the traditional super-resolution reconstruction result can be obtained.At the same time,we optimize the image denoising method synchronously with the super-resolution reconstruction algorithm.Compared with the traditional method which separates the super-resolution reconstruction algorithm and the denoising algorithm,our method can retain more image details and do denoise better.In this paper we also propose an image quality assessment(IQA)metric based on the frequency domain phase spectrum,which can reflect the super-resolution reconstruction effectiveness in the frequency domain.The proposed IQA metric has certain guiding significance for evaluating the effect of super-resolution reconstruction.Verified by both the software simulation experiment and the actual captured image experiment,the proposed method can effectively improve the image resolution and remove noise,and has better noise robustness and real-time calculation capacity.3.The super-resolution reconstruction method for infrared panoramic annular lens(PAL)image has been proposed.In view of the characteristics of infrared PAL image which is different from that of traditional optical image,firstly an image unwarping and super-resolution(U-S)jointly reconstruction technique based on POCS for multi-frame infrared PAL images has been proposed.In the new method,we optimize the image unwarping and super-resolution reconstruction progress synchronously,and extract pixels in the annular image adaptively for computation.The proposed method can improve the calculation speed,correct the distortion better and achieve higher spatial resolution.Verified by both the simulated infrared PAL image experiment and the captured visible PAL image experiment,the proposed algorithm can effectively improve the spatial resolution of the image,correct the distortion and improve the computational efficiency.