Data Processing Of Fourier Transform Imaging Spectroscopy

According to the whole procedure of data acquisition, processing and applicationof the Fourier Transform Imaging Spectroscopy (FTIS), some key technologies areanalyzed. The results are satisfactory and might be useful for the engineering practiceof FTIS. This thesis mainly includes the following contexts:1. The concept, classification, data processing and applications of imagingspectroscopy are briefly described.2. Some principle theories about FTIS are introduced, and the data amount of theFourier Transform Imaging Spectrometer is acquired. According to the character ofinterferogram, a Spectral Interpolating Prediction and Quantization Coding (SIP&QC)compression method is presented.3. The distortions in spectral radiometric calibration, which resulted frominstrument line shape (ILS) and signal to noise ratio (SNR), are described. Someprinciples are refined which are useful for the instrument design. A new radiometriccalibration method is presented for the low SNR data.4. Spectral data unmixing technology is described and linear mixing model isinduced by the feature space analysis of the spectral data. A Feature-space-analysisBased Endmember Determination (F-BED) method is presented and applied tospectral data unmixing, which is used for extracting the endmembers exiting inspectral data.5. The characteristics of the Temporarily Modulated Fourier Transform ImagingSpectrometer (TMFTIS) are discussed. For the necessary of the real-time targetdetection, a real time target interferometric detection system is proposed, and theadvantages and disadvantages are analyzed. The computer simulation of the target interferometric detection and optical path difference (OPD) setup is given.The main innovations in this thesis include:1. The data amount acquired by Fourier Transform Imaging Spectrometer isanalyzed theoretically and compared with the data acquired by traditional dispersiveimaging spectrometer.2. A Spectral Interpolating Prediction and Quantization Coding (SIP&QC)method is presented for data compression, and the 5:1 compression ratio is acquiredwith the Relative Spectral Quadratic Error (RQE) less than 1ï¼….3. The distortions in radiometric calibration is analyzed which arisen frominstrument line shape (ILS) and signal to noise ratio (SNR). Some principles arerefined which may be useful for the instrument design.4. An effective radiometric calibration method which is useful for the FourierTransform Imaging Spectrometer data with low SNR is proposed.5. The linear mixing model is induced by feature space analysis, and aFeature-space-analysis Based Endmember Determination (F-BED) method ispresented for endmember extraction.6. A real-time target interferometric detection system is proposed for a specialapplication, and the feasibility and optical path difference (OPD) design arediscussed.