A Diurnal Animation of Thermal Images from a Day–Night Pair

Abstract

Interpretation of thermal images is often complicated because the physical property information is contained in both the spatial and temporal variations of the data and thermal models are necessary to extract and display this information. A linearized radiative transfer solution to the surface flux has been used to derive a function that is invariant with respect to thermal inertia. This relationship makes it possible to predict the temperature variation at any time in the diurnal cycle using only two distinct measurements (e.g., noon and midnight). An animation can then be constructed from a pair of day–night images to view both the spatial and temporal temperature changes throughout the diurnal cycle. A more complete solution for the invariant function, using the method of Laplace transforms and based on the linearized solution, was introduced. These results indicate that the linear model does not provide a sufficiently accurate estimate. Using standard conditions (latitude 30°, solar declination 0°, acquisition times at noon and midnight), this new relationship was used to predict temperature throughout the diurnal cycle to an rms error of 0.2°C, which is close to the system noise of most thermal scanners. The method was further extended to include the primary effects of topographic slope with similar accuracy. The temperature was computed at 48 equally spaced times in the diurnal cycle with this algorithm using a co-registered day and night TIMS (Thermal Infrared Multispectral Scanner) data pair (330 pixels, 450 lines) acquired of the Carlin, Nevada, area and a co-registered DEM (Digital Elevation Model). (Any reader can view the results by downloading the animation file from an identified ftp site.) The results illustrate the power of animation to display subtle temporal and spatial temperature changes, which can provide clues to structural controls and material property differences. This `visual change' approach could significantly increase the use of thermal data for environmental, hazard, and resource studies.