新闻动态   News
联系我们   Contact
你的位置:首页 > 新闻动态 > 行业新闻

美国农学学会报道AgPixel在精准农业中的应用

2015-1-4 15:10:20      点击:

Improved crop monitoring takes flight

By Madeline Fisher November 26, 2013 -- Kevin Price is a rangelands ecologist and remote sensing expert who spent 19 years in the University of Kansas geography department working with satellite images on projects funded by NASA. But Price also grew up on a farm, so he welcomed the chance to “get back to his ag roots” when the Kansas State University agronomy department offered him a remote sensing professorship in 2008.
Two scientists standing next to a display of unmanned aircraft
Kansas State University researchers Kevin Price (left) and Deon van der Merwe pose next to two of their small unmanned aircraft systems (UAS). The pair is part of a larger team that has been working on agricultural applications for UAS for the past two years. Photo courtesy of Kevin Price.
His agronomy colleagues were just as enthused to have Price on board. Then he began showing them satellite images of cropland taken from 450 miles above the earth. “Most agronomists, at least in this department, are used to looking at individual plants on the ground,” Price says. “And when they can’t see an individual plant in a LANDSAT pixel that’s one-fifth of an acre in size, they don’t get very excited.” Fortunately, one of Price’s graduate students approached him two years ago wanting to work on applying unmanned aircraft systems (UAS) in agriculture. Commonly known as drones, small UAS equipped with cameras can be flown just above crops to monitor conditions such as nitrogen status or disease. Their impact on farming is projected to be huge. According to a recent study by the Association for Unmanned Vehicle Systems International, for example, small unmanned aircraft will be an $85 to $100 billion industry by 2025, with 80% of that money being spent in agriculture. Price didn’t know much about UAS at the time. But he soon connected with the aviation program at Kansas State’s satellite campus in Salina, as well as with a less likely collaborator: a colleague in veterinary medicine. Deon van der Merwe is a major UAS enthusiast who flies in competitions as a hobby and knows a great deal about building small airplanes, explains Price. With the help of the Salina aviation group and other agronomists, Price and van der Merwe are now wowing their colleagues with images snapped from UAS that can resolve individual crop plants and even single leaves. But getting there has meant much more than simply building the best airplane. The team has also had to learn how to fly the planes remotely, trigger cameras to take images at the right times and shutter speeds, and standardize image data collected under varying levels of cloud cover and solar angles. They’ve needed to invent a process for stitching thousands of images seamlessly together, and to navigate Federal Aviation Administration (FAA) regulations. “There are a lot of components to this technology that I never dreamed of when I started this work,” Price says. “It’s a whole new area of research that’s blowing wide open.” One of the thorniest questions is how to transform simple aerial images of fields that a farmer might use as “refrigerator door art” into information that can help management decisions, Price says. “That’s what we’re working really hard on.” For example, how much nitrogen should be applied? Does a certain section of a field need replanting? What are the projected yields? Working closely with the companies RoboFlight Systems, Inc. and Goldfinch Technologies, LLC, the team has now created new software called AgPixel. The program converts the pixel data in digital photos into a key measure of plant growth and vigor: the Normalized Difference Vegetation Index (NDVI). With “just a few clicks of the computer mouse,” Price says, farmers can then have AgPixel compare NDVI values across the entire field to those from a reference strip where crop plants are given all the nitrogen fertilizer they can handle.
Images of a corn field taken with an unmanned aircraft
One big challenge has been learning to convert image data into information farmers can actually use to manage their fields. On the left is a composite or “mosaic” image of a corn field (see the truck in the upper right for perspective), which was stitched together from many infrared images taken by a UAS. On the right, the image data has been converted into yield data. Note the swath of sandy soils, which is also a region of lower yields. Image courtesy of K. Price.
The final output is a map of variation in plant growth (i.e., NDVI) linked to crop nitrogen status that can be used to apply fertilizer where it’s needed to boost crop yield. And unlike tractor-based or handheld instruments that do the same thing, a UAS can collect the needed image data from a 640-acre field in about 20 minutes, depending on wind conditions. Since first becoming available commercially in August 2013, AgPixel has captured the attention of more than 130 companies in 40 countries, Price reports. “There’s really no other software out there that works well with the kind of imagery we’re collecting.” That’s why AgPixel was developed in the first place, he adds. At the same time, he advises people not to take any single advance as a sign to dive in, since many more details still need to be worked through. Take the airplanes, for example. Many small aircraft are already on the market, and some of them are very expensive. But that doesn’t mean they work as well as they should, Price cautions. The planes he favors are built by a company called Ritewing. They’re able to fly low and very slowly over farm fields without stalling. Plus they’re more likely to bounce than break when they crash, because they’re made of a foam-type rubber that absorbs much of the shock of impact. That leads to another point: Like all pilots, farmers and crop advisers who want to fly UAS need to practice first on flight simulator software. Then they should join their local flying club to get further training from a flight instructor. “The last thing you want to do is buy a really expensive system and learn to fly with that system,” Price says. “You’ll be picking it up in about five minutes as rubble.” Another reason not to rush is that the FAA isn’t expected to have regulations in place governing commercial (non-military) uses of UAS until 2015. As part that process, the FAA will be creating certification standards that commercial UAS must pass to ensure their safety and airworthiness. The Kansas State-Salina aviation program was recently selected by the FAA to test the new standards. In the meantime, Price and his colleagues continue to tackle each new hurdle; one immediate goal is to work some kinks out of AgPixel so it will be even easier for farmers and crop consultants to use. Despite all the questions and problems the team has had to face, however, Price has no regrets about leaving satellite imaging behind for the moment to take a closer look at croplands. “This technology is going to revolutionize the way work is done in agriculture,” he says. “Precision ag will never be the same.”