Validation of Weed Detection Systems
Universitat de Lleida (UdL)
Location
At user's premises

Spain

Arable farming
Horticulture
This service evaluates the accuracy and effectiveness of AI-driven weed detection systems, ensuring their accuracy in real agricultural conditions. By combining field trials (both in experimental and real-farm fields), drone-based imaging, and ground truth validation, the service assesses the system’s ability to detect weeds with high precision, minimising false positives (incorrect weed detection) and false negatives (missed weeds). The service also validates the system's adaptability to different environmental conditions, such as varying soil types, weather, and crop growth stages, and it is applicable to herbaceous and arable crops, offering valuable insights for improving targeted weed control and reducing unnecessary herbicide use.
Collection of test data
Performance evaluation
Provision of datasets
Test design
Test execution
Test setup