Finnish researchers have presented an affordable, multi-sensor solution to enable high-precision positioning and alignment of agricultural equipment. It is being developed under the ESA-funded PAALI project (‘Precision agriculture – high precision coordinate and alignments transfer’).
Led by the Finnish Geospatial Research Institute (FGI), the system integrates GNSS, stereo and infrared cameras, and inertial measurement units (IMUs). The system aims to address challenges in precision agriculture, offering reduced costs and greater efficiency.
Precision agriculture minimizes harmful pesticide and fertilizer use, mitigates soil depletion and erosion, conserves water, and lowers energy and labor costs. However, equipping with current GNSS-RTK-based solutions whole collections of individual farm tools such as trailing tillers, box blades and mowers, can be prohibitively expensive.
To overcome this, the PAALI project developed a unique coupling unit that mounts between a tractor and its trailing tool. This system uses multiple sensors and sensor fusion algorithms to determine both the relative positions of the tractor and tool and the absolute positions of tool components.
FGI Research Group Manager Tuomo Malkamäki explained the objective: “We wanted to be able to estimate the pose of the trailed vehicle with minimal or no components placed directly on that vehicle.” The result is a cost-efficient, adaptable prototype for various agricultural applications. The Precision Agriculture Demonstrator (PAD) was successfully tested in a number of real farming scenarios.
Design choices
Among other components, the PAD prototype includes:
- Septentrio SBi3 Pro+ with IMU: A high-precision GNSS/INS receiver with RTK positioning and robust anti-jamming.
- Flir Grasshopper cameras: Monochrome cameras with onboard image processing capabilities.
- Flir thermal camera: Captures infrared radiation to display temperatures and temperature changes.
- Sick Visionary B stereo camera: Provides 3D vision for complete scene capture in outdoor environments.
Tests demonstrated high accuracy in absolute orientation and real-time performance at 20–30Hz frame rates, with 80Hz available for recording. Visualization output matched trailer movements seamlessly, with no noticeable latency. The PAD withstood mechanical stress and vibrations, delivering precise, low-noise camera-based pose estimations. Some weaknesses were also identified, particularly with regard to calibration-related errors.
Malkamäki noted the broader potential of this technology: “The system has significant applications within the agricultural field, but going beyond agriculture as well, in logistics, things like trailer hitching, also marine approaches and docking, and in many other autonomous operations.”
The PAALI project was funded under ESA’s NAVISP program, aimed at supporting new, commercial developments in the European PNT sector.
