A Call for the CASI
A multi-million dollar technology transfer program introduces the CASI airborne system to Brazilians.
By Mary Jo Wagner

If the Compact Airborne Spectrographic Imager (casi) had lungs, it would probably be breathing a huge sigh of relief. It just recently ended a grueling three-year test of its capabilities in Brazil.
      Its data revealed sedimentation that was leaking into a reservoir behind a dam, assessed water quality and detected algae blooms, identified stress in trees, helped build forest inventories and discriminate species, indicated when fertilizer should be applied to tobacco crops, provided inventories to managers of a major apple orchard, determined land conditions and distinguished crops for banks financing farmers and mapped micro basins for future monitoring projects.
      The casi was busy.
      It was all part of a multi-million-dollar technology transfer program to develop remote sensing and GIS technology in Brazil. Aerodata S/A, a major Brazilian aerial survey company, wanted the casi technology and a Canadian consortium was only too happy to oblige them.
      In May of 1992, the program was launched. Itres (Imaging, Technology, Research, Excellence and Service) Research Ltd. in Alberta, provided the casi and training, Dipix Technologies supplied image analysis software and Intera Tydac provided the GIS software. Substantial funding was supplied by the Canadian International Development Agency. And the entire program was technically coordinated by Dr. Fl‡vio Kirchner, formerly of Aerodata, and organized by Emery Miller & Associates.
      Fifteen pilot projects were defined covering land use, reservoir monitoring, forestry and biodiversity classification in key regions of the country. Additional projects came and went as well once the program got underway. The most recent project ended last August.
      "Generally, what we were trying to prove, in all of these projects, was the efficacy or commercial applicability of multi-spectral airborne techniques in remote sensing," said Emery Miller, now vice president of business development at Worldsat in Ontario, Canada. "And the ability to locate, detect or determine certain conditions that would be vastly more expensive to find otherwise."
      All the projects were designed to prove what the casi could and could not do. To the casi's credit, it performed well in the majority of its 15 challenges.

The Cash
Itres designed the casi, combining the better features of aerial photography and satellite imagery with the detailed data analysis of a spectrometer. Launched in 1990, Itres has sold about 12 systems to users worldwide.
      The casi is a pushbroom imaging spectrometer that measures radiation from 400 to 1000 nanometers of the electromagnetic spectrum. Very mobile, installable in a variety of aircraft, it can operate in two modes: the spatial mode where it collects data in 19 bands (maximum), and the spectrometer mode which provides continuous spectral measurements for 288 intervals. Fully programmable, the casi has a swath of 512 pixels and the resolution can range from .5 meters to 10 meters.
      Casi imagery can be geometrically rectified using accurate aircraft movement, GPS measurements and digital elevation models.
      But perhaps what sets the casi apart from other optical airborne systems is the ability to tune the wavelength and width of its spectral bands to meet the needs of a particular application. That provides much more detailed information than optical satellite imagery, says Steven Mah, vice president of Itres.
      In fact, had it not been for the casi's use of narrow bands, a unique "bump" may have gone undetected in one project in Guanabara Bay, Rio de Janeiro. The bump resulted from red tide algae. That water feature may have been more difficult to identify with satellite imagery, says Mah.
      "We had a band positioned in the location which highlighted that feature," he says. "Satellite imagery generally uses very large spectral bands and the resolution is coarser so you would never see such a narrow feature."
      That characteristic of the casi was one of the deciding factors for its use in the technology transfer program.
      "There are no current satellites capable of providing adequate spectral data to do the projects we did," adds Miller.
      Aerial photography was used in the majority of projects for verification of the positioning and geometry of the casi imagery, and for ground control. In some cases, Landsat Thematic Mapper imagery was also used as a base map on which to overlay the imagery.
      Ironically, the first project that was conducted was a failure. The objective was to detect leakage from underground storage tanks, such as petroleum and septic tanks, and to analyze how any leaks could effect the surrounding vegetation. Miller says there was not enough vegetation in the region to locate tanks that were leaking because the majority of the areas of concern were in the city. In short, it wasn't practical.
      There was a valuable lesson in that failure, however. It clearly demonstrated that the casi is not a universal instrument for all applications and cannot be stretched beyond its capabilities.
      But when the casi "fit" the application, the results were impressive. Two notable projects were deemed particularly successful, according to Mah and Miller. One involved orchard management, which was carried out by Hugo Braga, a Santa Catarina University Ph.D. student in Brazil, and the other involved micro-basin studies.

POMIFRAI
The Pomifrai apple orchard in southern Brazil was chosen as the test site for the orchard management project because of its size and apple production volume. Covering over 575 hectares of orchards, managers maintain about 182,000 apple trees of six different species. Although that's a lot of ground to cover, Pomifrai had mapped the orchard well in previous years which provided a good source of data to compare results with.
      The objective was to use casi data to build a comprehensive inventory of the trees in the orchard, including the number and location of trees by species, how many acres they covered, and the health condition of those trees. In addition, they wanted to assess the ability to estimate the yield of each of the six species.
      Data were collected in the late spring of 1992 in Brazil. First the casi was flown in spectrometer mode to acquire a full spectral signature of selected sites and species of concern. Analyzing that information, they chose and positioned the spectral bands and flew the area again. Aerial photos were acquired as well.
      Using the casi data, Braga was able to extract five of the six species at 3 m resolution. They weren't able to detect the sixth type of tree because they were all young, immature trees with small canopies, which allowed the surrounding vegetation to overpower the signal.
      All of the data was put into a GIS, providing detailed maps such as change detection, damage and stress assessment, which managers could use in their daily operations. What took Pomifrai years to build with the resources they had, the casi provided in 10 days.
      Miller says that data were also acquired more cheaply - about 25 percent cheaper - than using traditional methods of air photos and ground surveys, and are about 20 percent more accurate because the data provide complete coverage.
      This database can now be used for managing their crews, planning for new species and most importantly, as a useful tool for assessing damages after a storm or infestation - a monumental task at present.
      "The orchard managers were very impressed with the results," says Miller. "They felt it was a very cost-effective process and could be very valuable to them."
      Braga was also pleased with end result. He has been in contact with Mah and it seems that Braga is trying to establish an operational environmental monitoring program.

Micro-Basin Study
The micro-basin study was the beginning of the end for the technology transfer program. The last project carried out, it perhaps, more than any of the other projects, exemplified what a technology transfer program is all about: teaching, learning and sharing.
      A group of undergraduate and graduate students and professors at Santa Catarina University contacted Itres in September 1994. They had a research grant and wanted to learn more about the casi technology they had been exposed to in the previous years of the program.
      In February 1995, Mah went down to Brazil to educate the students on casi technology.
      The site chosen for this project was a 28 sq. km. micro basin near the city of Cocal, in Santa Catarina. The micro basin is surrounded by a mixture of native forest, rolling terrain, various agricultural crops and hog farms. If the basin is not carefully monitored and managed, the entire ecosystem and surrounding urban centers could suffer.
      Prior to the project this micro basin had not been properly mapped or documented. And this is just one in about 525 micro basins in Santa Catarina that are in need of updating.
      The objective of the study was to delineate areas of soil exposure for agricultural development and planning, and to identify streets and highways to monitor urban growth, using casi data and image analysis techniques.
      Obtaining casi data that had been previously acquired in November 1992, the students learned about the casi, its modes of acquisition and specific image analysis methods and applied what they learned to the micro basin site.
      Through a particular analysis technique called spectral linear unmixing, they were able to detect soil exposure areas at a resolution of 3.5 m. Urban streets and highways were also identified and mapped.
      Mah says it was a good way to end a positive program for now.
      "This type of technology was quite new to the undergraduates," he says. "The graduates had some previous exposure to the casi so they really took to it. And the professors were very impressed. It was a great boost for us because we could see how Brazil is really eager to have this technology."
      Both Miller and Mah say the program was a success and exceeded their expectations. Miller says out of the 20 pilot projects conducted, nine of them "had very positive results and would be economically viable without question." Of those nine, he feels the Pomifrai was the best success all around, thanks to the casi.
      "I'm a big fan of the casi," he says. "It has the ability to do things that no other commercial instrument can today. The ability to collect the data at a spectral location that you want ,with a band width that you want, I believe, is directly responsible for the success of the Pomifrai project, as well as most of the other successful projects."
      And just as Brazil is eager for casi technology, Itres is equally as anxious to get the technology to them. In fact, says Mah, that's a goal Itres is actively working on. At present, Brazil is left to work with old casi data. But Itres has continued to enhance the data collected from several of the projects, such as the micro basin, and would like to show the project group how the casi data and the casi itself have been improved.
      "One of our objectives is to try and establish a link with the universities so that they can spread knowledge," says Mah. "The people there are very interested and capable of using this technology. They just need the tools and equipment to act on it."
      Until then the casi can enjoy the silence and prepare for the future.

About the Author:
Mary Jo Wagner is a freelance writer/ editor who writes about GIS and remote sensing. She may be reached at 415-291-8292.

Back