基于物理模型的污染云团红外光谱精细仿真

发布时间：2018-06-25 17:18

本文选题：红外光谱 + 云团扩散　；参考：《红外与激光工程》2017年01期

【摘要】：鉴于获取实测光谱成本高、获取量少和可测量云团种类少等限制因素,研究污染云团的红外光谱仿真,对于利用仿真光谱进行光谱识别的算法研究显得至关重要。以往的研究多利用经验的概率模型和经验的或半经验的参数来模拟污染云团的实时扩散,在此基础上再进行污染云团的红外光谱仿真。文中将利用基于物理的模型来精细地模拟污染云团的扩散,以弥补概率模型的欠精确性。研究了基于物理的云团扩散的机理,以及基于此扩散模型上红外光谱的生成方法,最终将生成的物理模型下的仿真光谱序列、概率模型下的仿真光谱序列和实测的光谱序列进行比较,得到了更为精准的仿真结果:就光谱残差而言,最高可提高14%,并指出了两种模型的适用范围。文中建立的基于物理模型的污染云团扩散及其红外光谱的实时仿真方法,对于高精度的云团红外光谱仿真及高质量的光谱识别算法研究具有重要意义。
[Abstract]:In view of the limited factors such as the high cost of obtaining the measured spectrum, the small amount of acquisition and the small number of measurable cloud clusters, it is very important to study the infrared spectral simulation of the contaminated cloud cluster for the study of the algorithm of spectral recognition using the simulated spectrum. Previous studies have used empirical probabilistic models and empirical or semi-empirical parameters to simulate the real time diffusion of contaminated clouds, and then simulated the infrared spectra of contaminated clouds. In order to compensate for the inaccuracy of the probabilistic model, a physical model is used to simulate the diffusion of contaminated clouds. The mechanism of cloud diffusion based on physics and the generation method of infrared spectrum based on this diffusion model are studied. Finally, the simulated spectral sequences under the physical model are generated. By comparing the simulated spectral series with the measured spectral sequences under the probabilistic model, a more accurate simulation result is obtained: in terms of spectral residuals, the maximum value can be increased by 14%, and the applicable range of the two models is pointed out. The real time simulation method of pollution cloud diffusion and infrared spectrum based on physical model in this paper is of great significance for the research of high precision infrared spectrum simulation and high quality spectrum recognition algorithm.
【作者单位】：中国科学院安徽光学精密机械研究所;中国科学技术大学科学岛分院;
【基金】：国家自然科学基金(41505020)
【分类号】：X87

【参考文献】