The Danish Climate Ministry uses aerial drones flying over oil spills and trash to send signals to sailing drones to clean up the waste. The imagery from aerial drones is fed into a machine learning algorithm accelerated by Kinetica, which automatically identifies waste in the images.
Through data contributed by citizens to the Earth Challenge mobile app and reference-grade sensor data, Kinetica powers efforts to collect and integrate billions of data points on plastic pollution, air quality, food security, insect populations, water quality, and climate change.
San Francisco Estuary Institute
The San Francisco Estuary Institute is capturing thousands of images of the Bay Area via drones in order to pinpoint the location,quantity, and type of trash in waterways. SFEI runs its analysis of the images on Kinetica to more quickly and accurately determine where trash is coming from.
In the initiative codenamed CityShark, the Danish Climate Ministry is partnering with Kinetica to create an autonomous system that will clean up oil spills and trash from the Port of Aarhus. Aerial drones flying overhead send images to Kinetica where they’re rapidly analyzed for the presence of waste. If oil spills or trash are detected, sailing drones are sent to clean them up. Kinetica allows the Danish Climate Agency to build their machine learning models and run their geospatial analyses in real time.
The Earth Challenge initiative is slated to become the world’s largest coordinated citizen science campaign. It empowers individuals to contribute to global scientific data, including by identifying plastic pollutants, documenting fluctuations in air quality, and monitoring the health of bee populations. Data from the Earth Challenge application is stored and analyzed in Kinetica, where scientists can draw new insights about the health of our planet.
Cleaning up our oceans, rivers, and waterways is an immense challenge, so the San Francisco Estuary Institute is leveraging neural networks trained on drone imagery to help identify trash quickly and at scale. However, with tens of thousands of images of the San Francisco Bay and neighboring tributaries, it took SFEI more than a month to get results. By running its algorithms on Kinetica, SFEI is able to complete its analyses in hours.