Mapping the world's food production with AI and satellites
Being a farmer has always been a bit like being an intelligence officer. In order to succeed in agriculture, you need to keep a close eye on what growers in your region and around the world are up to. Vital information like crop diversity, fertilizing schedules, and expected yield can have a significant impact on your strategy and your bottom line.
As agriculture becomes increasingly technology-driven, it should come as no surprise there's now an app for that.
OneSoil, a precision farming startup, is launching an international map that applies its AI algorithms to satellite images of farms in 43 countries around the world. The AI is able to analyze the satellite images to determine information like crop type and phenostage, which is the development cycle of plants.
"The world's population is constantly growing, but the amount of agricultural land remains the same. In order to provide food to meet the needs of the population, farming must be more effective than ever before," said Slava Mazai, OneSoil CEO and co-founder. "The OneSoil Map was built to reveal local and global trends in crop production, and make those trends available to everyone from the agriculture industry. It helps to predict market performance both at region and country scale and fosters smart decision making for agronomists, farm managers and traders."
In addition to real-time information, OneSoil's map allows farmers to see how fields have changed over the past three years. The company pulled off the feat by using public data from the European Space Agency and Mapbox, an open source mapping platform.
SEE: The Future of Food (ZDNet/TechRepublic special feature)
Agriculture has been an eager adopter of many AI technologies. Predictive analytics and technologies including autonomous robots and crop soil monitoring have transformed how farmers of all sizes make growing decisions.
OneSoil's map uses AI, deep learning models, and computer vision to autonomously detect field boundaries and predict what crops are growing based on certain visual signatures.
Soon it may be possible to predict yield based on visual analysis of satellite imagery. In the not-too-distance future, it's likely that such a system will be able to give farmers detailed insights and suggestions about fertilizer and watering schedules, taking some of the guesswork out of a notoriously unpredictable industry.