Technology could be the future of crop scouting
Technology could be the future of crop scoutingBy KRISS NELSON
Could the long hours of walking through fields in the hottest and not-so-best conditions to scout for weeds, bugs, diseases and more soon be replaced with technology?
Danny Singh, professor of agronomy at Iowa State University said there is a current trend for the use of automation in crop scouting which will make for more timely inspections done covering a wider area, but also with an intent to provide time sensitive mitigation.
“There has been a huge improvement in our ability to use satellite, drones and robots for scouting,” he said. “For such work, several disciplines are converging to make this happen including plant researchers, engineers and computer scientists.”
Singh said the interdisciplinary Soynomics team at ISU includes professors in agronomy in addition to himself, Arti Singh, and professors in engineering, Baskar Ganapathysubramanian and Soumik Sarkar, who have been working on these topics with several other researchers at ISU and other universities for several years and have received grant support from the USDA, NSF and Iowa Soybean Association as well as internally through ISU. Other members of the team include professors Dr. Daren Mueller ISU pathologist, Dr. Greg Tylka, ISU nematologist, and Dr. Matt O’Neal, ISU entomologist.
“The team also involves more than 30 graduate students, undergraduate students, post-doctoral fellows, and research staff that enable this work geared towards providing solutions to farmers, and base material for innovation for companies,” he said. “We have on-going projects on ground robot and drones based plant phenotyping including for production traits, as well as plant stress traits.”
Looking to the future
Singh said the future will continue to see improvements in the ability of drones that can fly longer and cover wider distances with clearer images as well.
“This means that farmers and scouts can cover wider areas and determine the diseases or plant stresses that may be present on their farms, and can make decision on how to control yield-limiting factors to increase profitability,” he said.
Other improvements are occurring with sensors and computer programming to obtain information that cannot be easily obtained by current methods.
“For example, researchers including us, have shown how hyperspectral cameras and machine learning can be used to identify stresses even before they are visible to human eyes,” said Arti Singh.
Getting the computer to read the results
Singh said the advances have given us the ability to use images taken through drones, robots and satellites and convert the information to field level information on crop health.
While images can tell a lot, Singh said, computers need to be able to process those images and obtain crop information from them.
“Human experts can look at a photograph and identify a specific disease,” he said.
“But the computer needs to be taught to identify the diseases. However, once we can teach the computer to learn the information in images, we can really increase the throughput as the computer can perform tasks that are repeated with more efficiency than humans. All of this has happened due to machine learning which is around us everywhere in our day to day lives,” said Sarkar
When will this technology be available for crop scouting?
Actually, it already is.
“The technology is already available, and will likely become mainstream with time,” said Arti Singh. As technology will continue to improve, more options will become available in the market with an end to end solution that farmers and scouts can use easily use. Off the shelf solution with responsive services will be needed for better adoption.”
Don’t put your crop scouting shoes and tools in the closet quite yet, however.
These technologies, Singh said are not meant to replace in-person scouting, but more to be used as a tool to help improve the crop scouting process.
“We believe that the scouts will move on to more supervisory roles to operate and monitor these AI-based tools rather than being human sensors as they are today,” said Ganapathysubramanian.
Singh said it is important the money is there to continue researching the possibility of using drones, robots and satellites for crop scouting.
“We will have to ensure research investment is continually made in the area because public research is needed to ensure that the ground work is done for private companies and startups to take up the results to build tools for scouting,” he said.
Please Enter Your Facebook App ID. Required for FB Comments. Click here for FB Comments Settings page