FieldCompanion project comes to an end with excellent results.
The FieldCompanion project has reached a successful conclusion with many excellent results being attained. This is despite the COVID-19 pandemic causing many logistical problems. Moving forward, the project’s partners shall continue to invest in post-project efforts on pending issues during the coming months. This will be focussed on polishing the system’s finer details to allow a demonstration to happen when access to test-fields in Spain is possible.
Sundance’s goal for the project was to create a scalable Edge-AI platform to detect and enable a sprayer to efficiently treat pests and diseases in orchards and vineyards.
Using the algorithms developed to let the sprayer “See and Spray” it was demonstrated that we can significantly reduce the total quantity of pesticide being used by only spraying it where it is required. The other benefit to the more efficient spraying is that it will also reduce the amount of diesel fuel used in the process.
This is a short demonstration of the final testing of the ICS system for the FieldCompanion Project.
- Person Danger Detection
- Tree classification
- Fruit counting
- Tree segmentation
- Running the full software stack on VCS-1 board
The ICS system is now a fully integrated solution that receives left and right image streams to compute the next nozzle states and control the nozzles control array.
A state-of-the-art AI algorithm and an advanced vision perception systems enable the ICS system to:
- Estimate the number of fruits per tree
- A human/animal safety awareness system used to stop spraying while the humans/animals are being detected;
- Perform a visual speed estimation for avoiding over-/under-spraying plants
- Detect canopies and trunks for performing selective spraying by controlling different sections of the nozzles spraying arrays via the SCS.
The consortium calls this platform “ICS”, whereas Sundance will be promoting this as the “VCS” (Vision, Control Sensor) system. Full details can be found here.