Thermal Infrared Spectral Analysis Thermal absorptionemissivity structure 3

Thermal Infrared Spectral Analysis Thermal absorption/emissivity structure (3 -20 µm) can be indicative of the structure and make up of materials (particularly minerals). Figure 5. 1 shows variation in laboratory spectra with structure and, 5. 2 shows variation between rock types Digital Imaging and Remote Sensing Laboratory

Thermal Infrared Spectral Analysis (cont’d) Fig. 5. 1 Thermal infrared transmission spectra of silicate minerals showing the correlation between band location (vibrational energy) and mineral structure. Digital Imaging and Remote Sensing Laboratory

Thermal Infrared Spectral Analysis (cont’d) Fig. 5. 2. Emission spectra of various rock types showing resistrahlen minima (Vickers and Lyon, 1967) Digital Imaging and Remote Sensing Laboratory

Thermal Infrared Spectral Analysis (cont’d) The problem in sampling in this spectral region (320 µm) is that the signal is heavily influenced by thermal effects. The spectral radiance signature can be expressed L(l ) = e (l )LT (l )t 2 + (1 - e (l ))Ld (l )t 2 + Lu (l ) Digital Imaging and Remote Sensing Laboratory

Thermal Infrared Spectral Analysis (cont’d) The atmospheric variables can be solved for using LOWTRAN if detailed radiosonde data are known. An alternative (cf. Hook 1992) is to adjust the inputs to LOWTRAN in an interactive fashion and predict the temperatures for targets with known spectral emissivities. When the predicted temperatures are the same in all bands, the atmosphere is assumed to be correct. Digital Imaging and Remote Sensing Laboratory

Thermal Infrared Spectral Analysis (cont’d) This method has been successfully applied to thermal infrared multispectral scanner (TIMS). A sixband line scanner developed and operated by JPL. The relative emissivities can be estimated by first solving for T in one channel where emissivity is assumed constant over the image (i. e. in continuum) and then solving for the emissivity in the other channels. Digital Imaging and Remote Sensing Laboratory

Thermal Infrared Spectral Analysis (cont’d) A second technique that is useful when spectral data over a wide range are available is to fit the surface-leaving radiance to a Planck curve and select the lowest temperature that doesn’t force an emissivity greater than 1. Digital Imaging and Remote Sensing Laboratory

Thermal Infrared Spectral Analysis (cont’d) Experimental results: Figure 5. 16 shows a comparison of six point emissivity spectra from TIMS compared to laboratory spectra (not clear which technique is used). Digital Imaging and Remote Sensing Laboratory

Thermal Infrared Spectral Analysis (cont’d) Fig. 5. 16 (a) Spectra from TIMS data, Mauna Loa basalts. (b) Spectra from laboratory, Mauna Loa basalts. Digital Imaging and Remote Sensing Laboratory