Virginia Tech Research Magazine
None of these model airplanes will be under Christmas trees this year — especially with a price tag of nearly $12,000 apiece. However, the hope is that this “toy,” with its ability to detect plant stress caused by pest infestation and disease, will soon find its way into the hands of crop consultants who can help farmers keep their fields healthy.
Currently, farmers spray for pests when they observe a problem. However, that may be too late. Insects and plant pathogens move rapidly and by the time a farmer notices damage in one area damage could be underway in another. Satellite systems have been used to provide images of farmlands, but the most commonly used system, Landsat, is on a 16-day interval cycle — again giving pests more time to attack crops. In addition, the coarse resolution of Landsat images makes it difficult to use these images to pinpoint specific problem areas within individual agricultural fields. Other satellite systems that provide more accurate and timely images are available, however, the images from these systems are much more costly. With a model plane, or Unmanned Air Vehicle (UAV), armed with an infrared and color camera, high-resolution images of fields can be obtained and processed to detect pest and disease damage before problems become severe on one farm or even a portion of a farm.
While the idea of finding pests from 500 feet above the crops may still seem a bit like looking for a needle in a haystack, the low altitude remote sensing system can detect plant stress so that, within a few hours, a farmer can isolate and control problem areas before they spread. The broader perspective gives farmers a visual representation of their fields, identifying where the problems are and where they could be.
“Remote sensing cameras can see problems before the human eye can detect them,” says Carlyle Brewster, assistant professor of entomology at Virginia Tech, who is leading the charge to develop the bug mapping system. One of the goals of Brewster’s research is not just to help one farmer at a time, but also to create a database of information maps, or “bug maps,” that can be accessed for quick reference. It would be a way for farmers to share information and to spot common trends.
To do this, remote control planes, which can fly 8 to 45 miles per hour, are sent to scout for disease and pests, and then coordinate the information with global position systems (GPS), which provides an accurate location of the pest or disease. Photos of the field are taken and then the images are ‘dissected’ and analyzed by blending the red layer from the color photo with the near-infrared photo. In looking at the overlapped images, white indicates healthy green vegetation because of the way plants absorb and reflect light. Black indicates soil or water while varying degrees of gray indicate levels of plant stress. This information would then be stored and used as a reference for other troubles that may arise.
How it began
Brewster began to consider model airplanes as a research tool while looking at the white fly problem in Florida. As a Ph.D. student at the University of Florida in the mid-1990s, his focus was insect population modeling. In Florida, whiteflies were controlled by natural enemies. However, 3,000 miles away in Southern California, whiteflies were destroying crops in the Imperial Valley. The contrast provided a rich context for Brewster’s research.
“We were looking for a way to obtain remotely sensed images that were inexpensive compared with satellite images and that would allow us to take images of fields at any time,” says Brewster. The researchers wanted a quick turnaround time for managing pest problems. About the same time, precision agriculture, or site-specific farming, became trendy in the entomology world. However, most of the research focused on slow-moving problems, such as weeds, soil conditions, and poor crop yields. Precision agriculture wasn’t being used much on fast-moving pest and disease infestations.
The problem with precision agriculture for pest management is that it relies on manual scouting, which is labor intensive, time consuming, and costly, Brewster explains. Finding a way to reduce the labor and costs would allow farmers to have a more accurate and timely picture of the pest and disease infestation affecting their crops. The model plane may be the answer.
Present status of the research
Currently, the system is in the design phase and is being funded by the U.S. Department of Agriculture Southern Region Integrated Pest Management System and the Virginia Agricultural Experiment Station. Researchers have focused on learning how the components work together and what is needed to make the system better. ”
Brewster reports, “All of our flights with the UAV have been successful.” During test flights of the UAV and remote sensing cameras, the major difficulties have been the placement of the cameras on the plane and the quality of the photos.
“The cameras can communicate with the ground video system,” Brewster says. “However, processing the video is time consuming and does not always provide the best quality images. “Modifications are underway to perfect this prototype system.
The next steps involve more test flights over peanut fields in Suffolk, Va., before heading to Florida to study pest problems in citrus.
Having every farmer own one of these model planes is not the intended outcome. Researchers would like farmers to have access to the information quickly and eventually be able to download information directly into pest control equipment. But first, “this system has to go through several modifications to be user-friendly,” Brewster says. Those modifications include higher resolution cameras and the addition of the global positioning component, which would make it easier to process images and control the model plane. The UAV would be programmed with the geographic coordinates of an area to be analyzed. The plane would fly and take images without much outside interference.
“Ideally, there should be real-time analysis of the image that could be downloaded into pest control equipment,” Brewster says. That will take at least five years, as improvements are made in image processing systems, and in video and wireless downlinks. With all of these improvements and additions to the plane, the final hurdle will be to bring the cost down so that the remote control airplane might be found under more Christmas trees in the future.
-  This article was printed as is in the fall 2002 Virginia Tech Research Magazine.