You can read on the Internet how to tell when corn is ripe, how the silk should be dry and the kernels soft. But for farmers who cultivate huge fields and sell their corn as animal feed, every day that goes by outside the ideal harvest window represent a huge sum of money lost. Therefore, many of them prefer to entrust the harvest to professionals and modern technology. How to know the right time to harvest corn has been researched for the third year by electrical engineers from the BUT.
A large drone flies over a green field overgrown with corn silk and takes pictures. When it is done, it navigates to the ground and that is when the slightly less action-packed part of Jiří Janoušek’s research begins. At the Faculty of Electrical Engineering and Communication BUT, he is examining photos taken by a multispectral camera and trying to figure out if now is the best time to harvest the corn. Jiří Janoušek chose this “agricultural” theme as the topic for his dissertation | Autor: Jan Prokopius “Each lens on the camera takes one picture in a narrow spectrum. To our eyes, such image looks like a simple black and white photograph, but when comparing them to each other, big differences can be seen. In the programme, I compose them into reflectance maps and calculate various vegetation indices,” says Jiří Janoušek, pointing to a colour record of the field. The different shades indicate, in a simplistic way, the health of the plant. The infrared camera can, for example, remotely detect from the leaves whether the plant is beginning to wilt.
LENGTHY ANALYSES AND CLOUDY SKIES
Currently, farmers rely on satellite imagery, but this has about ten times lower resolution than a drone and cannot be used when there is cloud cover over the area. The satellite also passes over the site roughly once every five days, and that can result into significant delays. Time is also an issue for chemical analyses, where samples are taken manually from the corn in the field.
"When they take samples, they have to wait for a week for the results, so the harvest time is more estimated than calculated. They thought of trying to find a method that would determine the exact time of the harvest, preferably on the same day, so that they would know if the time is right the very next day. They also considered a method where measurements would be taken a week in advance and they would know if the ideal harvest time will be in about a fortnight, or if it was already too late,” Janoušek points to the time pressure. For farmers, missing the ideal harvest season by even a few days can result in a huge loss. After composing the camera photos, colour maps are created based on different vegetation indices. The aim of the project is to select the one that best captures the maturity of the corn | Autor: archive of Jiří Janoušek At the Department of Theoretical and Experimental Electrical Engineering at FEEC BUT, doctoral student Janoušek has been trying for three years to test whether it is possible to determine the exact chemical composition of corn from images taken by a multispectral camera drone. Fortunately, it turns out that his assumption is valid.
HOW MANY LITRES OF MILK DOES A COW PRODUCE?
The large drone is equipped with a multispectral camera with five lenses | Autor: Jan Prokopius
And what is essential for determining the ideal harvest time? “The dry matter content is the most important factor for corn. Farmers are able to convert it into milking capacity of cows, i.e. up to litres of milk,” explains Jiří Janoušek enthusiastically. “I will fly a drone over a field, find out its vegetation index, and the farmer will know that after feeding that particular corn, the cow will produce that much milk. For them, these are perfectly specific numbers.”
Drones are fast, simple to operate and weather does not affect them as much as satellites | Autor: archive of Jiří Janoušek
Jiří Janoušek chose this topic for his dissertation and since he cooperates with NutriVet, a company that has been involved in chemical analyses for a long time, he had ideal conditions for the research: “We imaged the corn from the beginning of growth and followed it all the way through. The month before harvest is the most intense part. We agreed not to harvest part of the field and leave it for us to examine the differences after harvest. In the most intense part we went to the field twice a week, NutriVet took samples each time and I flew a drone over the field and took multispectral pictures.”
He tried the drone in the fields of large companies near Velke Bílovice, Letovice or Nový Jičín. In addition to various corn hybrids, he also focuses on another feed – alfalfa. In the future, he plans to use the imaging for the harvesting of cereals.
According to Jiří Janoušek, farmers do not have to worry about being financially ruined by drone imaging: “A week’s delay can cost farmers millions, so such a measurement is a negligible amount.”
