We are pleased to announce the successful completion of the development and evaluation of a selective spraying system incorporating vertically arranged nozzles controlled by solenoid valves and a microcomputer.

We are pleased to announce that we have developed an algorithm that uses hourly weather data to predict the risk of infections by downy mildew, powdery mildew, black rot, and grey rot.

We are pleased to announce that we have developed an LSTM-based AI model to forecast monthly average temperatures in vineyards, using historical meteorological data.

We are pleased to announce that we have implemented a Random Forest model to estimate the rainfall required for healthy vineyards, based on local soil properties and forecast temperatures. This will enable us to estimate the amount of irrigation needed to keep vineyards healthy.

We are thrilled to announce that we have created a dataset with over 8,000 images of infected vine leaves, covering six different diseases. We have also trained a Faster R-CNN AI model for the automatic detection of these diseases.

We are pleased to announce that we have completed our mapping software, capable of creating 2D maps of vineyards with an accuracy of a few centimetres. This will enable precise and selective spraying of pesticides.

#AGRARIAN

We are pleased to announce that we have successfully trained an AI model based on Mask R-CNN for the detection and identification of grapes and vine leaves in vineyards.

We visited AVIPE's vineyards in Palmela and our prototype rover captured 9,000 high-resolution photos of the vines and grapes, along with their exact position, for the WINESMART project.

The WINESMART project team presents its prototype rover equipped with six cameras and RTK positioning, ready to map vineyards.

We're excited to announce that we're collaborating with Agrarian on the WINESMART project. We'll be developing a prototype rover equipped with cameras and RTK positioning to map vineyards at high resolution, allowing for precise selective spraying of pesticides.

FIRECOM in action: high-resolution orthophoto map and 3D point cloud created from images captured by drone.

FIRECOM in Action: Drone Mission Planner. This module allows you to create image capture missions according to specific requirements. The drone then automatically flies the designed mission.

The FIRECOM project was developed by Axtron Systems as part of the EU Horizon CHAMELEON project. Its main objective is to use drones and artificial intelligence to help reduce the risk of wildfires spreading in urban-forest interfaces.

You can find out more on our website (link on our profile).