@inproceedings{BasavegowdaMosebachSchleipetal.2022, author = {Basavegowda, Deepak Hanike and Mosebach, Paul and Schleip, Inga and Weltzien, Cornelia}, title = {Indicator plant species detection in grassland using EfficientDet object detector}, series = {42. GIL-Jahrestagung, K{\"u}nstliche Intelligenz in der Agrar- und Ern{\"a}hrungswirtschaft}, booktitle = {42. GIL-Jahrestagung, K{\"u}nstliche Intelligenz in der Agrar- und Ern{\"a}hrungswirtschaft}, publisher = {Gesellschaft f{\"u}r Informatik e.V.}, address = {Bonn}, isbn = {978-3-88579-711-1}, issn = {1617-5468}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:eb1-opus-6656}, pages = {57 -- 62}, year = {2022}, abstract = {Extensively used grasslands (meadows and pastures) are ecologically valuable areas in the agricultural landscape and part of the multifunctional agriculture. In Germany, the quality of these grasslands is assessed based on the occurrence of certain plant species known as indicator or character species, with indicators being defined at regional level. Therefore, the recognition of these indicators on a spatial level is a prerequisite for monitoring grassland biodiversity. The identification of indicator species for the status quo of grassland using traditional methods was found to be challenging and tedious. Deep learning-algorithms applied to high-resolution UAV imagery could be the key solution, where UAV with remote sensors can map a large area of grassland in comparison to manual or ground mapping methods and deep learning-algorithms can automate the detection process. In this research work, we use an EfficientDet based algorithm to train an object detection model capable of recognizing indicators on RGB data. The experimental results show that this approach is very promising in contrast to the difficult and time-consuming manual recognition methods. The model was trained with the momentum-SGD optimizer with a momentum value of 0.9 and a learning rate of 0.0001. The model was trained and tested on 1200 images and achieves 45.7 AP (and 85.7 AP50) on test data set. The dataset includes images of four distinct indicator plant species: Armeria maritima, Campanula patula, Cirsium oleraceum, and Daucus carota}, language = {en} } @article{MouratiadouLemkeChenetal.2023, author = {Mouratiadou, Ioanna and Lemke, Nahleen and Chen, Cheng and Wartenberg, Ariani and Bloch, Ralf and Donat, Marco and Gaiser, Thomas and Hanike Basavegowda, Deepak and Helming, Katharina and Ali Hosseini Yekani, Seyed and Krull, Marcos and Lingemann, Kai and Macpherson, Joseph and Melzer, Marvin and Nendel, Claas and Piorr, Annette and Shaaban, Mostafa and Zander, Peter and Weltzien, Cornelia and Bellingrath-Kimura, Sonoko Dorothea}, title = {The Digital Agricultural Knowledge and Information System (DAKIS): Employing digitalisation to encourage diversified and multifunctional agricultural systems}, series = {Environmental Science and Ecotechnology}, volume = {16}, journal = {Environmental Science and Ecotechnology}, publisher = {Elsevier}, issn = {2666-4984}, doi = {10.1016/j.ese.2023.100274}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:eb1-opus-8012}, pages = {13}, year = {2023}, abstract = {Multifunctional and diversified agriculture can address diverging pressures and demands by simultaneously enhancing productivity, biodiversity, and the provision of ecosystem services. The use of digital technologies can support this by designing and managing resource-efficient and context-specific agricultural systems. We present the Digital Agricultural Knowledge and Information System (DAKIS) to demonstrate an approach that employs digital technologies to enable decision-making towards diversified and sustainable agriculture. To develop the DAKIS, we specified, together with stakeholders, requirements for a knowledge-based decision-support tool and reviewed the literature to identify limitations in the current generation of tools. The results of the review point towards recurring challenges regarding the consideration of ecosystem services and biodiversity, the capacity to foster communication and cooperation between farmers and other actors, and the ability to link multiple spatiotemporal scales and sustainability levels. To overcome these challenges, the DAKIS provides a digital platform to support farmers' decision-making on land use and management via an integrative spatiotemporally explicit approach that analyses a wide range of data from various sources. The approach integrates remote and in situ sensors, artificial intelligence, modelling, stakeholder-stated demand for biodiversity and ecosystem services, and participatory sustainability impact assessment to address the diverse drivers affecting agricultural land use and management design, including natural and agronomic factors, economic and policy considerations, and socio-cultural preferences and settings. Ultimately, the DAKIS embeds the consideration of ecosystem services, biodiversity, and sustainability into farmers' decision-making and enables learning and progress towards site-adapted small-scale multifunctional and diversified agriculture while simultaneously supporting farmers' objectives and societal demands.}, language = {en} }