Is digital imaging cost-effective?

Several agriculture-oriented companies use multispectral digital camera systems and specialized software to assess the health of farm crops. It’s a technique that some federal officials think could help reduce crop insurance fraud and facilitate fast, accurate claims processing.

“We’re trying to apply technology in a way that provides a tool to insurance adjusters and producers,” said Don Brown, chief financial officer of Vega Imaging Solutions, a company that sells near-real-time digital imaging services.

In 2005, the federal crop insurance program lost about $117 million to waste, fraud and abuse, according to Agriculture Department figures. The USDA’s Risk Management Agency (RMA), crop insurance companies and the House Agriculture Committee are working together to reduce those losses.

“We agree with the thrust of the technology and the thinking,” said Garland Westmoreland, RMA’s director of strategic data acquisition and analysis.

Rep. Jerry Moran (R-Kan.), chairman of the committee’s General Farm Commodities and Risk Management Subcommittee, said computer systems and loss adjustment are two areas in which additional technology would be mutually beneficial for RMA and federal crop insurance recipients.

“If the loss adjustment process utilizes the latest technology to ensure an accurate and efficient assessment process, it will also better ensure that all dollars are spent wisely and justly,” he said.

A crop insurance official who is not authorized to speak on the record said that everyone recognizes the potential benefits of near-real-time digital imaging. “How you harness it is probably the biggest question on everyone’s mind,” he said. Lack of funding, for example, might prevent RMA from immediately realizing the potential benefits.

“It all comes down to a cost/benefit ratio of what our resources will allow us to do in this area,” said Eldon Gould, RMA’s administrator.

The agency’s use of such technology is at least two years away, Westmoreland said. As a first step, RMA would work with the crop insurance companies it reinsures, he said. Then it would conduct tests in limited areas to assess the technology’s effectiveness.

“If we could determine how to employ this with the cost efficiency as well as the quality that we need, then we would work with our reinsured partners to figure out how to do it at that point,” Westmoreland said.

Since 1938, crop insurance adjusters have conducted manual inspections at farms where producers filed claims for crop losses. “It’s pretty labor-intensive, and it takes quite a while,” Westmoreland said, but he added that new technology is no substitute for on-site inspections.

The USDA has combined remote imaging with on-site inspections since the early 1980s, said Jim Hipple, a remote sensing and geographic information systems adviser at RMA. The images help the USDA verify claims and settle disputes between farmers and adjusters. Archived images are important because RMA often handles contested claims one to three years after they are filed, he said.

The USDA mostly relies on satellites to produce aerial images because they are readily available and cheaper to use than planes. RMA obtains satellite imagery from the U.S. Geological Survey and NASA and gets some aerial photography from the USDA’s Farm Service Agency.

“What Don Brown is proposing is going a step further in using remote imaging,” Westmoreland said. “It would be doing something that’s real time.”

Vega Imaging Solutions uses four digital cameras on computer-controlled planes to capture images of a farm two to four weeks after a reported crop loss. Planes usually fly for about 20 minutes in parallel lines over a designated area, with the cameras capturing images every four seconds. That information is immediately stored in a computer.

All four cameras capture identical images through different filters: red, green, blue and false infrared. Using specialized software to analyze the images, researchers can measure levels of crop damage in areas as small as one-tenth of an acre, Brown said.

Because the images are digital, analysts can use data-mining algorithms. Data mining has helped the USDA save millions of dollars in fraudulent claims by detecting abnormalities and triggering a claims-review process, Gould said.

But whether the USDA’s current data-mining system can handle the additional multispectral images is uncertain. “We may be expecting 20 or 30 terabytes of data per year, which would be a massive amount of data to handle in terms of data mining in addition to all of our crop data,” Hipple said.

Daniel Bertoni, acting director of the Government Accountability Office’s Natural Resources and Environment Division, said he believes in using technology to extend the enforcement arm of government. GAO has not analyzed the technology the USDA and others are considering, but he thinks the concept could be useful.

“We should have gone way beyond relying on anecdotal statements,” he said.