Small Unmanned Aircraft Systems For Crop And Grassland Monitoring

K-State has one of the leading small Unmanned Aircraft Systems (sUAS) programs in the U.S. These capabilities have been developed at KSU-Salina Aviation, and now faculty and students in KSU Aviation, Veterinary Diagnostic Laboratory, Agronomy (Environmental and Agricultural Spatial Analysis Laboratory), and Geography are teaming up to develop sUAS remote sensing applications in agriculture and environmental monitoring. This collaborative effort among programs has made great strides within the past year. We are now able to provide remotely sensed multispectral images with a resolution of one inch or less from an unmanned aircraft. And we have received approval in the form of multiple FAA Certificates of Authorization to fly key research areas in the area around Manhattan, Kansas.

We are now able to utilize completely autonomous aircraft, both winged planes and hexacopters (6 rotors), to analyze a wide variety of vegetative conditions more quickly and at levels of detail not possible in the past. This technology, which uses a variety of spectral data, has potential for several uses, including:

•  Early-generation crop breeding line evaluations
• Quick and wide-scale scouting of croplands for crop conditions and weed densities
• Grassland conditions and production evaluation
• Detection of noxious weeds
• Detection of blue green algae blooms in ponds and reservoirs

An example of how this technology is being used in the K-State soybean breeding program was described in Agronomy e-Update No. 387, from February 1, 2013. Some of the other applications of this emerging technology are described in articles 2, 3, and 4 below.

The underlying technology for all these applications is the same. The sUAS is equipped with a still visible camera that measures visible and near infrared light. Then a flight plan is programmed into the computer. An operator puts the sUAS into flight while other operators monitor its flight path to make sure it stays on track. After the flight, the remotely sensed images are downloaded to a computer for display and image processing.

The key to successfully using spectral analysis for any of the purposes mentioned above is the ability to find differences in the remotely sensed measurements made by the camera or imaging system.The remotely sensed imagery used for this purpose may be a specific wavelength from the visible spectrum, near infrared (NIR), or thermal portions of the spectrum, or combinations of wavelengths such as the red and NIR used to compute a vegetation index.

One example of a vegetation index is the Normalized Difference Vegetation Index (NDVI), which is computed using the following equation: (Near Infrared – Red)/(Near Infrared + Red). The NDVI index is used to detect differences in plant pigmentation that are indicative of plant production, health, and growth status. This is the same index that is used at a much larger scale to construct the nine maps that appear toward the end of each issue of the Agronomy e-Update (see, for example, article No. 5 in this issue).

Part of the ongoing research in our project at K-State is to discover improved methods for discriminating among vegetation types. Such methods might include the use of indices such as the NDVI, but additional information about plant characteristic might be extracted using the very fine-scale spectral variation within the plant canopy caused by the varying arrangement of the plants’ leaves. The image below shows a natural color photograph of winter wheat on the right and canola on the left. This picture was taken from about 150 feet above the ground using a hexacopter and Canon S95 camera.