Key Technologies Research Of Information Acquisition And Processing In Long-wave Infrared Hyperspectral Imaging

Long-wave infrared hyperspectral has the prominent ability of target recognition and detection efficiency, mainly applied in precise detection of objects radiation spectrum, such as chemical gas detection, resources survey, astronomical observations, etc. Research significance and value of long-wave infrared hyperspectral is significant.In view of the problem of big noise and low signal-to-noise ratio in long-wave infrared system, this paper used three-dimensional noise model to analyze noise from the perspective of system, developed several key techniques of information acquisition and processing system according to the analysis result and improved the signal-to-noise ratio of the long-wave infrared system and image quality finally.Main research contents and innovations are presented as follows:1) For the reason that noise analysis and calculation method of infrared imaging system is not comprehensive, this paper introduced 3-D noise model theory from the perspective of system, expounded the mathematical definition of 3-D noise model, deduced noise component specific calculation formula, applied the model to the conventional design of the long-wave infrared imaging system and completed the quantitative analysis of noise. 3-D noise model was a powerful tool for noise analysis of infrared system. At the same time, 3-D noise model theory was scientific, rigorous and comprehensive, especially suitable for being an evaluation criterion for the noise performance of infrared image.2) Studying the characteristics of infrared detector output signal through theoretical analysis and simulation in this paper, it was found that high frequency components of signal mainly worked on signal rising time but little impact on signal amplitude. Through selecting the optimized system bandwidth and controlling the ADC sampling location, it can not only achieve the effective access to information, but also dramatically compress system bandwidth, so it greatly reduced the time noise in the system compared to the conventional design method.3) According to Mooney’s classic theory of pixel response model, pixel linear response model was deduced and used as theoretical basis. An analog front-end hardware real-time non-uniformity correction technique was put forward and researched in this paper. The correction principle, implementation method and calibration error was analyzed. Finally, the experiment results verified the correction effect of this technology. By comparing images before and after the correction and spatial noise values, it was found that the correction effect of this technology was good and improvement of non-uniformity was obvious. At the same time, with advantages of real-time, high speed, large array, its engineering practical value is huge, this technique offers a new way for non-uniformity correction methods and has a certain guiding significance.