[{"data":1,"prerenderedAt":96},["ShallowReactive",2],{"content-en-/en/data":3},{"id":4,"title":5,"body":6,"description":16,"extension":89,"meta":90,"navigation":91,"path":92,"seo":93,"stem":94,"__hash__":95},"content/en/data.md","Data Collection & Monitoring: The Digital View of Grasslands",{"type":7,"value":8,"toc":81},"minimark",[9,13,17,22,25,28,32,35,38,41,45,48,51,68,72,75,78],[10,11,5],"h1",{"id":12},"data-collection-monitoring-the-digital-view-of-grasslands",[14,15,16],"p",{},"![Image placeholder: Close-up of a drone or camera system hovering over a meadow]",[18,19,21],"h2",{"id":20},"precision-through-modern-sensor-technology","Precision Through Modern Sensor Technology",[14,23,24],{},"The basis for any targeted care measure in the aimGRASSLAND project is highly precise data collection. To accurately assess the condition of grassland areas, we rely on a so-called multimodal approach. This means that we don't trust just a single information source, but combine different technologies to obtain as complete a picture of the vegetation as possible.",[14,26,27],{},"High-resolution 2D cameras are used to provide detailed visual information. In addition, we evaluate and integrate 3D cameras as well as hyperspectral sensors. While conventional cameras only capture visible light, hyperspectral sensors can collect information in wavelength ranges that are invisible to the human eye. This helps to reliably distinguish plant species not only by their leaf shape but also by their spectral signature – that is, the way they reflect light.",[18,29,31],{"id":30},"flexible-carrier-platforms-and-precise-positioning","Flexible Carrier Platforms and Precise Positioning",[14,33,34],{},"The sensor technology is flexibly deployed on different carrier platforms, depending on requirements and terrain conditions. For fast and comprehensive aerial coverage, we use drone technology (UAVs). For detailed ground-level recordings, the sensors can also be mounted on tractors or autonomous robots.",[14,36,37],{},"The knowledge of exactly where a plant is located is crucial for later control. Therefore, all image data is linked with high-precision RTK-GPS data. This technology enables centimeter-accurate positioning of each individual detected plant. This creates a digital map on which every ragwort and every valuable indicator species is precisely located.",[14,39,40],{},"![Image placeholder: Schematic representation or graphic: Drone scans field -> Data transmission -> Digital map]",[18,42,44],{"id":43},"training-the-artificial-intelligence","Training the Artificial Intelligence",[14,46,47],{},"The collected raw data is not yet usable on its own – it must be interpreted. This is where artificial intelligence (AI) comes into play. For the algorithms to reliably distinguish between a toxic problem plant and a harmless meadow herb, they must be trained.",[14,49,50],{},"This learning process is based on two pillars:",[52,53,54,62],"ul",{},[55,56,57,61],"li",{},[58,59,60],"strong",{},"Real Data:"," In field trials on the farms of our partner farmers, we collect thousands of images under the most varied weather and light conditions. These images are \"annotated\" by experts, meaning the plants are manually marked and named so that the AI learns what it sees.",[55,63,64,67],{},[58,65,66],{},"Synthetic Data:"," Since manual marking is extremely labor-intensive, we take an innovative approach and additionally use synthetic data. Here, plants and growth processes are simulated on the computer or 3D models are used. These artificially generated images help the AI learn faster and also recognize rare scenarios for which sufficient field data is not yet available.",[18,69,71],{"id":70},"from-image-to-decision-support","From Image to Decision Support",[14,73,74],{},"The goal of data collection is not to gather data for its own sake, but to provide action-relevant information for agriculture. The data analyzed by the AI is processed in a data management system.",[14,76,77],{},"Farmers gain access to this information via a user-friendly platform or app. The results include, for example, clear \"heat maps\" (infestation maps) that show at a glance which parts of a field are heavily infested with problem plants and where valuable biodiversity hotspots are located. This data forms the basis for deciding whether and where control is necessary and enables complete documentation of field development.",[14,79,80],{},"![Image placeholder: Person (farmer or researcher) viewing data/map on a tablet directly in the field]",{"title":82,"searchDepth":83,"depth":83,"links":84},"",2,[85,86,87,88],{"id":20,"depth":83,"text":21},{"id":30,"depth":83,"text":31},{"id":43,"depth":83,"text":44},{"id":70,"depth":83,"text":71},"md",{},true,"/en/data",{"title":5,"description":16},"en/data","aZd1_b_f6y0IEYsve_TvOHZ-BRzEhjc18hN5DB523jg",1792588245609]