遥感傅里叶变换红外光谱层析技术中的重构算法

Reconstruction Algorithms of Computed Tomography of Remote Sensing FTIR

在遥感傅里叶变换红外光谱层析技术中 ,对于是否能够快速、经济地得出气体空间浓度分布图及预测浓度峰位置 ,重构算法的选择是至关重要的。理想的重构算法即使在仪器噪声较大、存在测量误差、光线数量少以及待测气体在空间的分布不稳定情况下 ,也能够通过有限的测量数据和非对称的光线构型 ,得出理想的重构图。本文对遥感傅里叶变换红外光谱层析技术中采用的几种重构算法 ,如ART法、MART法、SIRT法和SBFM法等 ,作了扼要描述。与传统的ART法相比 ,MART法在迭代速度和迭代精度上有较大的优越性 ,SIRT法克服了采用代数重构技术所带来的模糊性 ,SBFM法使重构结果噪声明显减小。

Reconstruction algorithms of computed tomography-remote sensing fourier transform infrared spectroscopy plays a key role in fast and cost-effective acquirement of gas concentration distribution. Ideally, a reconstruction algorithm can be used to obtain good reconstruction maps only by limited number of measurements and non-symmetric ray configurations, even in the presence of instrument noise, measurement error, poor beam geometry, and variations in the contaminant spatial distribution. Several conventional reconstruction algorithms used to reconstruct two-dimensional concentration maps of the vapor contaminants by remote sensing FTIR, such as ART, MART, SIRT and SBFM, have been reviewed in this paper. Comparing to ART, MART has the advantages of iterative speed and iterative precision, while SIRT overcomes the problem of illegibility and SBFM can reduce the noise sharply.