CAFA AgMRS (Agricultural Multi Robot System)
a fully automated system that uses a drone and a mobile ground robot for IPM (Integrated Pests Management).
AgMRS uses the drone’s fully automated Dock360 platform and RoboStation for the robot, which consists of a wheeled robot CAFA RAM-20 and its charging station.
AgMRS system sends automatically tasks to both the drone and the robot and plans repetitive flights/rides to detect the threat of pests.
CAFA AgMRS uses both a drone and a robot with a camera (approximately 20MPx) and close-ups to detect the formation or arrival of pests thanks to Computer Vision algorithms that automatically process these photos in near real time.
CAFALyzer is Computer Vision software that uses GPUs to analyze photos collected by a drone and a mobile robot to assess the presence of pests.
Pest data collected and analyzed by the AgMRS system can be exported in common formats for use with other precision farming software used by the farm, i.e. send to location of the detected pests and relevant tasks to tractors.
The core problem– pests appear unexpectedly in the field and therefore it is important to detect the first location as early as possible to prevent crop loss. If the outbreak of pests is controlled in the initial phase, the cost of pesticides will be significantly lower, which will save money.
Currently, pests are detected with a time delay when the farmer goes scouting the field. On average, only a few places in the field can be inspected, even after 5-7 days. Therefore, it is usually not possible to identify pests until they have already expanded significantly. This, in turn, has already had a significant effect on yields. CAFA AgMRS provides an automated overview of each field with a delay of up to 1 day, allowing a quick response, reducing both crop losses and the financial cost of plant protection products. The CAFA AgMRS solution significantly increases the detection of pests in the initial phase and at precise locations, thereby reducing the use of chemical plant protection products.
AgMRS consists of 5 main components:
1. The Dock360 is a fully automated drone flight dock. The drone reviews photos from a height of approx. 60m to detect major changes. At a height of about 2m, the drone takes detailed photos to detect the presence of pests.
2. RoboStation, consisting of a wheeled robot CAFA RAM-20 and its charging station CAFA R-Station, where the battery is fully charged. The robot drives along the tractor tracks in the fields and takes photos close to the ground to detect pests that are not visible to the drone from above.
3. CAFALyzer is Computer Vision software that uses GPUs to analyze photos collected by a drone and a mobile robot to assess the presence of pests. Video analytics requires a lot of computing power and therefore cannot be placed on board a drone. Computer Vision algorithms are constantly being improved.
4. CAFA C2 (Command and Control) control system that sends automatic tasks to both the drone and the robot and plans repetitive flights/drives when the VideoLyzer detects a threat.
5. Map application and user interface for the farmer allows to log in to the map application, which shows an overview of the fields, where AgMRS has detected pests, which pests, the frequency of pests. In addition, an analytical view is displayed on the timeline.
AgMRS Technical details:
1. AgMRS is able to cover an area with a radius of up to 4km because of this distance communication with the drone control panel is ensured in the European Union (higher transmission power is allowed in the USA). The monitored area is therefore 500 hectares.
2. The drone and robot can work 24/7/365 because they can also work at night. The drone cannot fly if the wind strength is more than 12 m/s, but at the same time ground robot can drive to conduct monitoring tasks.
3. AgMRS needs power supply 230V/300W or solar panel.
4. The farmer can give to the system from time to time priority areas to check. If a crop has been harvested from a field, its monitoring is no longer relevant.
5. Early warnings and analytics. Expected emergence of pests related to warm air currents (potato beetle, etc.)
The CAFA AgMRS allows:
• Identification of pests
• Automatic counting of pests
• Setting the control criterion
• Positioning the damage center
• Determining the intensity of damage
• Take immediate pictures of damage, catch insects from the air and quantify them
• Provides input for timely scheduling of plant protection work in the IPM.
• The transformation of geoinformation of damage sites into an automated sprayer control system for crop protection sprayers should allow for precision application of pesticides, reducing the use of toxins against non-target organisms.
