Airborne hyperspectral imaging, a military technology originally designed to pick out camouflaged vehicles from surrounding foliage, may soon help detect a different concealed enemy: citrus canker. This contagious bacterium causes citrus trees to lose their leaves and drop their fruit; its spread can be halted only by burning infected trees and neighboring vegetation. "Our entire citrus industry is at risk," said Jack Neitzke, director of Florida's Citrus Canker Project. More than 20 million trees had to be killed after a citrus canker outbreak in the 1930s, and 33,000 more had to be destroyed between 1986 and 1992. The disease can be hard to spot from the ground and almost impossible to detect from the air, but airborne hyperspectral imaging can locate infected trees early (shown in red). To nip citrus canker in the bud, all 800,000 acres of Florida's orchards must be inspected, along with any citrus trees in suburban back yards. Each tree must be closely examined, since infected leaves look similar to healthy ones in the early stages. A typical eight-man inspection crew covers only about 150 acres per day, Neitzke said. But the pace may speed up if the hyperspectral sensor system passes field trials. Small enough to fit into a single-engine plane, this instrument uses an imaging spectrometer to measure visible and near-infrared light reflected off vegetation in 3-nm slices, according to Michael Barnes, CEO of the Melbourne, Fla.-based Galileo Group Inc., the manufacturer of the system. Hyperspectral imaging identifies infected trees by the unique spectral fingerprint their leaves emit when releasing chemicals to fight canker. The system, developed in collaboration with 3DI LLC of Easton, Md., incorporates software from Research Systems Inc. of Boulder, Colo., and can print these mathematical data as color-enhanced photographs. Ground-based inspection teams can then use these to home in on suspected canker sites, eliminating the need to examine vast numbers of trees. Better, cheaper, faster Previous aerial systems could collect data only in 150-nm-wide segments, reducing spatial and spectral resolution. In addition, they cost up to $100,000 to operate, took days to process and used "bathtub-sized devices that require large jets or turboprop planes flying at 10,000 to 30,000 feet to carry them," Barnes noted. Similarly, satellites were tried 15 years ago to spot canker, but their finest resolution went down to only a 30-m2 pixel, too large to depict the 6-m2 canopy of an individual citrus tree. Galileo Group's prototype is flown at 3000 feet, resolves down to 1 m2, costs $30,000 and prints out immediately. The system also measures absorption rates of substances such as fertilizer or pesticide, so a farmer can tell if he has spread too much or too little. Before introducing the system to the marketplace, Barnes wants to work on making data interpretation easier. Blight or mineral deficiency can sometimes cause a tree to produce an unusual spectral signature, "making for lots of fingerprints on the doorknob to sort through," he said. Neitzke is enthusiastic about the promise of aerial hyperspectral imaging technology. "If this works, it would reduce our need for manpower and equipment, and speed up our timetable. It would be incredible," he said.