NaLamKI: AI Technology Project for Sustainable Agriculture
Changing climate conditions, a shortage of skilled workers, and pesticide use are among the wide range of factors affecting the quality and operations of agricultural processes. Researchers at the Fraunhofer Institute for Telecommunications (HHI), also known as the Heinrich Hertz Institute, aim to make agriculture more efficient and sustainable using cloud and artificial intelligence technologies.
As part of the NaLamKI project, they are working with partners to create a "software as a service" platform that collects information from devices and machines to build a database for forecasts and decision-making support. The agricultural sector faces serious challenges: German farmers are already feeling the far-reaching effects of climate change, and in the future they will have to adapt to them to a greater extent.
Rising temperatures and changing precipitation patterns affect all agricultural variables, from crop growth and crop rotation to soil cultivation. Decentralized AI in the cloud and centralized AI on farms can help make the adaptation process to changing conditions more efficient, speed up the process across all areas of agriculture, and thus make the overall ecosystem more flexible and future-oriented.
This is where the NaLamKI project comes in. The project will focus on creating a cloud-based "software as a service" (SaaS) platform with open interfaces for agricultural and industrial providers, as well as providers of specialized crop production solutions.
By combining sensor and machine data collected via satellites and drones, soil sensors, robotics, manual data collection, and inventory data, a data pool can be created that enables agricultural processes to be optimized more sustainably using advanced artificial intelligence methods.
AI applications deployed on the platform help farmers analyze the condition of crops and soil over large areas, and change the organization of plant protection and nutrient processes such as irrigation, fertilization, and pest control, to ensure sufficient crop yields in both quality and quantity, while also reducing emissions and preserving biodiversity.
For example, targeted use of plant protection products increases yields, reduces costs, conserves resources, and actively protects the environment.
“In addition to climate change, the shortage of skilled labor also affects the quality and operations of agricultural processes. As a result, plant condition often can only be checked on a very selective basis. Currently, it is impossible to detect and accurately determine soil water status or pest infestation, for example, across large agricultural areas. Among other things, we are studying the analysis of images obtained from drones, satellites, and robotic cameras, and making the results meaningful for farmers”
To address this challenge, the institute is developing artificial intelligence methods within the project to analyze remote sensing data for modeling agricultural processes and creating 5G networks on farmland. By combining all the data, scientists will gain insights into the characteristics of cultivated areas that were previously nearly unavailable.
Farmers will be able to interact with AI and ask questions. For example, based on current soil moisture and crop disease indicators, the AI will be able to provide action instructions and show the effects of different scenarios. Specifically, a panel showing the area of farmland and the current soil condition will be displayed on a tablet. By tapping specific areas, the farmer will receive information about problems such as low water levels, as well as recommendations on how best to address them.
Training data and AI services will be provided in a decentralized manner using Gaia-X, a European data-secure cloud infrastructure. Moreover, a decentralized AI system with distributed learning will be created, whose data will be stored locally on farms. Farmers will be able to share AI models and transfer them to the NaLamKI platform to continuously improve algorithms. The platform will be open to third-party providers. For example, startups will be able to offer their innovative artificial intelligence solutions on the platform.
The initial data collection for the AI model development process is already complete. In particular, images taken by a robot during apple tree planting on a fruit farm in the Palatinate region of Germany are now available. For this purpose, data from various sensors, such as position sensors, LIDAR, RGB, and multispectral cameras, were collected, analyzed, and combined while the autonomous robot moved through the plantation.
The goal is to create a meaningful representation of the trees in the orchard to determine the number and ripeness of fruit, trunk diameter, the condition of individual crops, and the surrounding soil. This also includes detecting obstacles, in particular living creatures in tall vegetation, along the path traveled.
“We evaluate the data as it is collected across the plantation. The information we gather is combined with a map of fruit tree populations and displayed on a field map. Based on this data, documentation is created for the farmer”
