As satellite imagery is increasingly used to measure and monitor changes on the surface of the earth, there is great need to develop novel methodologies to bridge the gap between small-scale field studies and global-scale satellite observations. We are thus developing a lightweight, remotely controlled hyperspectral remote sensing instrument package to be flown from a large tethered balloon platform. This instrument package will be capable of measuring the visible and near-infrared spectral reflectance of the surface in very high detail. By changing the altitude and view angle of the platform, we will examine how the variation of spectral features can be used to quantify ecological changes by satellite. This platform will enable new research within the state of South Dakota and beyond. We expect that it will help uncover new understanding of how ecosystem structure and function can be quantified and monitored, thereby assisting with developing agronomic and forestry applications of remote sensing.
The SDSM&T tethered balloon is constructed of a urethane coated fabric and is inflated with helium. Its length is 38.9 ft, with a maximum diameter of 14.2 ft. The balloon is designed to fly in steady winds of up to 25 miles per hour, and its maximum free lift of 172 pounds is great enough to hoist a relatively large scientific instrument payload to an altitude of approximately 2 kilometers.
 |
 |
 |
Click thumbnails for larger images |
The instrument package is mounted to the balloon's tether line about 200 meters below the balloon to minimize shading from the balloon. Platform tilt adjustment, data collection, and data transmission is controlled in real time from the ground.
Some of the instruments onboard the SWAMI platform include:
Our spectroradiometer is a lightweight dual spectroradiometer (Fieldspec Dual UV/VNIR, ASD, Boulder, CO). It can rapidly gather contiguous, narrow-band (4nm bandwidth) spectral data between 350 and 1050nm wavelength. Its dual nature allows simultaneous measurement of the total downwelling solar irradiance by an upward-pointing sensor and the upwelling ground radiance by a downward-pointing sensor. The solid viewing angle is adjustable between 10 and 25 degrees. By changing the solid viewing angle and the altitude of the instrument package, we can design the pixel sizes so that they are similar or proportional to satellite images. With the rotation of the platform, the upward-pointing sensor can remain level and the downward-pointing sensor can change its view angles from nadir to 60 degree backward and forward angle.
This camera can capture the same target as the spectroradiometer and transmit the image to the ground in real time. This way we can choose the most suitable pixel to measure and have the video image help us correlate the ground cover characteristics with the spectral characteristics.
The thermal infrared sensor can collect thermal infrared spectral data between 8-14µm.
With the GPS receiver, we can measure the absolute location of the platform, including latitude, longitude and altitude. The GPS data is sent in real-time to the ground control unit so we can keep track of its position at all times.
We have placed various meteorological instruments on the platform that measure the current conditions and send the information in real-time to the ground control units. Some of the instruments include an anemometer, barometer, thermometer, and relative humidity sensor. In addition, we have a tilt sensor that tells how level the platform is.
The spectral information of a pixel represents a mixture of all the ground components (e.g. trees, grasses, bare soil) contained within. Therefore, many ecological processes occur at scales that are too fine to be seen by satellite. It is thus necessary to conduct remote sensing experiments at small and intermediate scales to improve our techniques of “unmixing” the spectral information contained within a pixel, so that we can interpret the relative abundances of ground components using larger scale satellite imagery.
By adjusting the elevation of the platform or the field of view, we can change the pixel size from several meters to hundreds of meters, so that it can bridge the gap between small ground-based scale and larger aircraft- and satellite-based measurements. By adjusting the pixel size, we can measure homogeneous ground covers and find their spectral properties. Then we can measure a large pixel consisted by many kinds of ground covers to research the specific details of spectral mixing and unmixing.
Many satellites currently in orbit measure the ground surface across a very large range of angles. Because natural ground surfaces reflect radiation anisotropically, there is a large amount of information that can be gathered by correctly interpreting angular remote sensing measurements. For example, shadows cast by tree canopies can be used to interpret forest structure via remote measurements. Over the past few years, research on the bidirectional reflectance of vegetation has shown that many structural and functional vegetation characteristics can be derived from multi-view angle remote sensing (e.g. Deering et al., 1992, Vierling et al., 1997, Sandmeier, 2000).
With our balloon system, we will be able to adjust the view zenith angle from nadir to 60° backward or forward angle to measure large structures such as tall forests and urban buildings. We expect that our research results will provide new and useful data to better interpret multi-angle measurements gathered by aircraft or satellites.
|
|
Click thumbnails for larger images |
|
Spectral vegetation indices are very useful in remote sensing research because they allow vegetation density and health to be measured and monitored using remote sensing measurements. One such index, the Normalized Difference Vegetation Index (NDVI), has been widely used as a standard in environmental remote sensing over the past two decades. However, NDVI suffers from several shortcomings as a function of view angle and pixel size. In addition, we also expect to develop and test new hyperspectral vegetation indices across a range of scales not previously attainable. Through NDVI and other spectral vegetation indices, we expect to make advances in estimating stand-level photosynthetic activity using remote measurements.
Many of the SWAMI field experiments will be conducted within South Dakota, and research results will help us to develop new techniques for studying cropland, rangeland, and forestland systems using remote sensing.
The balloon system can fly in the sky and collect continuous data for several hours, and the data can be transmitted to the ground in real time. This system will be cost-effective in relation to data gathered by aircraft.
The pixel size can be adjusted by raising and lowering the balloon, thereby ranging from several meters to hundreds of meters. These pixel sizes correlate well with existing satellite remote sensing platforms.
The spectroradiometer is capable of collecting hyperspectral data, and it can change the view angle by the rotation of the platform. These two capabilities represent two of the cutting edge techniques in remote sensing research.
Major funding for this project is being provided by a National Science Foundation CAREER grant DBI-9985039 to L.A. Vierling. Funding for the field spectrometer and various other system components was provided by the South Dakota NSF EPSCoR program.
