Method For Improving Image Spatial Resolution Of Airborne Thermal Imager Based On IBP Algorithm

Infrared ray, also called the infrared radiation. Thermal infrared image is the image of the object heat radiation, which is developed with the development of infrared imaging technology. The high-resolution thermal infrared image, by which the information of a scene can be described in details, is extensively used in many fields including computer vision process, medicine, and remote sensing, etc. But the infrared image is always with low spatial resolution, which is lack of consecution. At the same time, the infrared image is easy to be disturbed by the ambient noise, so the thermal infrared images are often accompanied by a variety of noise. Improving the spatial resolution of thermal infrared image by means of hardware is always expensive and is restricted by the technological level. So the super resolution reconstruction, which improving the spatial resolution of thermal infrared image by means of software, is now a research focus.The super resolution reconstruction technology uses the small difference between the sequence images of the same object to reconstruct the thermal infrared image. Compared with the single frame images, sequences images contain redundant information, which also includes some complementary information. We can use these information to implement image super-resolution reconstruction with some target imaging partially known parameters.This paper uses the super-resolution reconstruction algorithms to construct the aerial thermal infrared images. Experiments prove that the super-resolution reconstruction algorithm combined with the phase correlation displacement estimation algorithm and iterative back projection super-resolution reconstruction algorithm can effectively improve the spatial resolution of aerial thermal infrared images. The main technical route and innovation points of this paper are as follows:(1)According to the different processing fields, the common super-resolution reconstruction algorithms are divided into frequency domain method and spatial domain method. Compared with the frequency domain algorithm, the spatial domain reconstruction algorithm can get a better reconstruction result. After investigation this paper choose iterative back projection algorithm to reconstruct the thermal infrared image from aerial photography. We use the iterative back projection algorithm to reconstruct the simulated thermal infrared image. The reconstruction results show that the iterative back projection algorithm can effectively improve the spatial resolution of the thermal infrared image, and it is suitable for the thermal infrared image.(2)The traditional iterative back projection algorithm is often combined with the infrared micro scanning technology, the sequence images of fixed pixel displacement are obtained by infrared micro scanning technique,then we use the traditional iterative back projection algorithm to reconstruct the sequence images. But the displacement between the thermal infrared image sequence of aerial thermography is not fixed, so the traditional iterative back projection algorithm can not reconstruct the thermal infrared image from aerial photography efficiently. To solve this problem, this paper introduces an image registration algorithm.,the displacement information obtained from image registration is introduced into the reconstruction process. Aerial thermal infrared image is shooted with fast frame rate,so the displacement between the images is the sub pixel level,which proposed a high requirement for image registration algorithm. After the research and analysis of the common registration algorithm, this paper use phase correlation algorithm to calculate the displacement between the sequence image.(3)Then the displacement information and the traditional iterative back projection method are combined to realize the super resolution reconstruction of the thermal infrared image. So we can use the improved iterative back projection algorithm to reconstruct the thermal infrared image sequence of aerial thermography.(4)By the proving of experimentation,the improved algorithm is feasible from aspects of theory and experiment. MTF and PSNR are introduced to the quantitative analysis of reconstructed image to verify the effectiveness of the proposed algorithm.