Remote sensing applications of uncooled long-wave infrared thermal imagers
Loading...
Date
2012
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Montana State University - Bozeman, College of Engineering
Abstract
The commercial development of microbolometer uncooled long-wave thermal infrared imagers in conjuncture with advanced radiometric calibration methods developed at Montana State University has led to new uses of thermal imagery in remote sensing applications. As a result of being uncooled, microbolometer imagers are notably lighter and cheaper than typical cooled imagers, making them ideal for remote sensing. Two novel uses are discussed in the work presented here. The first is the imaging of beehives in order to remotely determine the hive vitality. Bees thermally regulate their hives to a narrow range of temperatures that creates a thermal signature seen in thermal infrared images. For each of the hives imaged, frame counts (or the number of full frames of bees in each hive) were found by manual inspection. Linear regressions of the normalized frame counts of the hives were performed versus the measured hive thermal radiance values. The resulting plots showed a strong relationship between the normalized frame count and the mean radiance of each hive, particularly in images taken just prior to dawn. The second novel use was imaging vegetation exposed to large ground concentrations of CO ₂ over a four-week period in summer for use in leak detection. A CO ₂ leak was simulated in a test field run by the Zero Emissions Research and Technology Center. Thermal infrared images were acquired along with visible and near-infrared reflectance images of the exposed vegetation and healthy control vegetation. Thermal radiance statistics were measured and a regression was performed versus the day of the experiment. The infrared data were found to have a strong R ² value and clearly show the effect of the CO ₂ on the vegetation. An additional regression was run on the infrared data combined with the reflectance data, and this was found to not add any unique information to the vegetation reflectance data. Both methods were found to independently indicate the potential of a CO ₂ leak before it was detected visually.