Skip to main content Skip to main navigation


AI-TEST-FIELD – A Test Environment for the Automated Evaluation of Methods for Robust and Reliable Environment Perception

Jan Christoph Krause; Naeem Iqbal; Mark Niemeyer; Benedikt Thy; Lutz Plagge; Hannes Hollmeier; Arno Ruckelshausen; Sebastian Röttgermann; Alexander Tauber; Jens Herbers; Stefan Menke; Joachim Hertzberg
In: LAND.TECHNIK AgEng 2023. International Conference Agricultural Engineering (LAND.TECHNIK-2023), 80th International Conference on Agricultural Engineering, November 10-11, Hannover, Germany, Pages 547-554, Vol. 2427, VDI Verlag, Düsseldorf, 2023.


In agricultural processes, harsh and changing environmental conditions conflict with a reliable perception of the environment by conventional sensor systems. Especially for safety reasons, the detection of objects like humans in the surroundings of a working machine is crucial. On the AI-TEST-FIELD typical agricultural field characteristics for example after tillage, mowing and harvesting can be represented. On that test field recording of sensor raw data with repro- ducible scenarios and objects under varying environmental conditions is offered. In this paper, different test capabilities are presented to investigate the performance of different sensor modalities, models, configurations, and algorithms. With the proposed test schemes, evaluation data for assessing the reliability of perception system interpreting the environment is gathered