|Norman, J - UNIV OF WISCONSIN|
Submitted to: Remote Sensing of Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: February 23, 2001
Publication Date: N/A
Interpretive Summary: Surface temperature measurement is important in many research disciplines, but it is difficult to do with sufficient accuracy. Currently available infrared thermometers (IRTs) typically have an uncertainty of 1-2 degrees C due to inherent limitations (primarily the effect of changes in sensor temperature), and even this level of accuracy requires frequent recalibration. To avoid these problems, we developed a new approach, known as differential infrared thermometry. It uses a conventional IRT that is rotated between two positions. In one position it views the target surface, while in the other position it views a small calibration chamber for which the temperature is precisely controlled and accurately measured with thermocouples. By comparing the sensor signal in these two positions it is possible to determine surface temperature with much greater accuracy, avoiding the problems that affect conventional IRTs. We obtained accuracy to within 0.04 degrees C in laboratory tests, and 0.17 degrees C in field tests, substantially better than conventional IRTS that were tested at the same time. This new instrument will be useful in a number of areas including study of climate anomalies related to sea surface temperatures and monitoring of foliage temperatures to detect crop water stress.
Technical Abstract: Surface temperature is a crucial variable, but it is difficult to measure accurately. Remote measurement by infrared thermometry is often the only viable choice, but is plagued by problems that limit its absolute accuracy. Primary among these are calibration shifts and an inability to eliminate the influence of detector temperature on the measurement. We have developed da new approach that avoids these difficulties by making the measurement differentially. The system uses a conventional infrared thermometer coupled to a solenoid so that its field of view can be periodically switched from the target of interest to a blackbody cavity. The temperature of the cavity is controlled and measured with carefully calibrated thermocouples. The blackbody is controlled at a temperature close to that of the surface and the detector output is measured while it views first the target area, then the blackbody. This allows accurate determination of surface brightness temperature, from which true surface temperature can be calculated with knowledge of surface emissivity and background longwave radiation. The system was tested in the laboratory by using it to measure the surface temperature of a mineral oil reservoir that was cycled over a range of temperatures and independently monitored with calibrated thermocouples. Over a 24 C temperature range the mean absolute error of the instrument was 0.04 C. The system was also field-tested for four days over a pond in which surface temperature was also measured with thermocouples. The mean absolute error for the period was 0.17 degrees C. A conventional IRT produced a mean absolute error of 2.2 degrees C. We conclude that the concept of differential infrared thermometry offers substantial improvement in the absolute accuracy of non-contact surface temperature measurement.