Creating Sustainable Farming Practices with Machine Learning
The 2050 Robotics Lab at Kansas State University is focused on finding ways to feed the world in a sustainable way by 2050. To do this, the 2050 Robotics Lab is working on a number of agriculture and environmental projects that include using robots to plant crops on high sloped hills, finding and spot treating pests in crops, and locating blue-green algae on open surface water. To make their robots as efficient as possible in the field, the group is using machine learning to train and automate the procedures.
Project Requirements
- Fix a real problem facing the world
- Make farms more sustainable
- Save farmers money
One large problem farmers face is insect infestations that can devastate crops. Currently, most farmers use pesticides on an entire crop when a potential infestation is suspected. Spraying an entire crop requires a lot of chemicals and a large amount of monetary investment. To minimize the cost, many farmers choose to wait as long as they can to spray pesticides (in order to spray as few times as possible) and end up losing some of their crop to insects while waiting for the best time to spray.
The 2050 Robotics Lab is working to create a better solution for both farmers and the planet. Using machine learning, the team is training their system to spot aphids on sorghum. Once the aphids are detected, a mobile robot spot sprays the area. By treating individual sections of the crop as aphids are detected, rather than spraying the entire crop all at once, farmers using this system will be able to save a large amount of chemicals and money while increasing their yield - resulting in a positive environmental impact all around.
Once finalized, this project has the potential to make a huge impact on the world, creating a solution that allows farmers to easily become more sustainable while increasing their yields.
The recent developments in affordable GPS units that record raw data have allowed Sibert to take his passion to a new level. As you can see, Sibert uses the RTK Express in conjunction with Mapillary to record precisely localized street-level pictures (pictures that he then uses later to contribute to OSM). He does this in three ways: by car with an antenna on the roof, and by foot or bike with his handheld device (pictured above). For Sibert, he derives pleasure from seeing pictures stuck to the road and not floating left or right (see screenshot below).
