Temperature and emissivity retrieval from remotely sensed images using the "Grey body emissivity" method

• Published in: IEEE transactions on geoscience and remote sensing Vol. 34; no. 3; pp. 681 - 695
• Main Authors: Barducci, A., Pippi, I.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.05.1996; Institute of Electrical and Electronics Engineers
• Subjects: Areal geology. Maps; Atmospheric modeling; Data mining; Earth; Earth sciences; Earth, ocean, space; Exact sciences and technology; External geophysics; Geologic maps, cartography; Geophysics. Techniques, methods, instrumentation and models; Image retrieval; Land surface temperature; Mathematical models; Meteorology; Ocean temperature; Oceanography; Q1; Remote sensing; Sea surface; Temperature; Temperature sensors; Emissivity; Accuracy; Temperature; Radiance; Multispectral remote sensing; Refraction index; Infrared spectrum; Absorptance; Imagery; Angle of incidence
• ISSN: 0196-2892; 1558-0644
• DOI: 10.1109/36.499748

The problem of extracting temperature and emissivity information from thermal infrared multispectral radiance data is of great importance in remote sensing applications. The Earth surface temperature is a relevant parameter for climatology, meteorology, oceanography, fire mapping, wind field determination, and energy balance at the atmosphere-Earth interface. In addition, ground emissivity is very important for lithological mapping, geological inspection, petrography, the dating of lava flows, etc. A new method for emissivity and temperature retrieval has been developed in the authors' laboratory which is based on the assumption that emissivity is a slowly varying function of wavelength. It is shown that this assumption holds true for many kinds of natural targets. The resulting model is called "grey body emissivity", and may be used to retrieve the emissivity of soils and other natural materials. The problem with using an effective wavelength or a bandwidth integral in order to represent the sampled radiance is discussed, thus leading to the implementation of two algorithms for the grey body emissivity. It is demonstrated that one of them, extensively discussed throughout the paper, provides the maximum likelihood values for temperature and emissivity. The other reveals two drawbacks: it produces noisy processed images and works well only for target temperatures greater than 274 K. Comparisons with other models are carried out utilizing remotely sensed images, laboratory measurements, and numerical simulations. The main approximations and the resulting errors, contained in the considered models, are briefly discussed. The influence of the atmospheric effects on utilized remote sensing data is taken into account. Finally, the authors investigate the emissivity dependence on the viewing angle. Theoretical predictions which rely on Fresnel's and Kirchhoff's laws are compared with the emissivity maps extracted from TIMS images over sea. Starting from the image data alone, the procedure allows them to compute the refraction index and the absorption coefficient of the concerned marine water.