The Global Wheat Head Detection (GWHD) dataset was created in 2020 and has assembled 193,634 labelled wheat heads from 4700 RGB images acquired from various acquisition platforms and 7 countries/institutions. With an associated competition hosted in Kaggle, GWHD_2020 has successfully attracted attention from both the computer vision and agricultural science communities. From this first experience, a few avenues for improvements have been identified regarding data size, head diversity, and label reliability. To address these issues, the 2020 dataset has been reexamined, relabeled, and complemented by adding 1722 images from 5 additional countries, allowing for 81,553 additional wheat heads. We now release in 2021 a new version of the Global Wheat Head Detection dataset, which is bigger, more diverse, and less noisy than the GWHD_2020 version.

Society is increasingly demanding a more sustainable management of agro-ecosystems in a context of climate change and an ever growing global population. The fate of crop residues is one of the important management aspects under debate, since it represents an unneglectable quantity of organic matter which can be kept in or removed from the agro-ecosystem. The topic of residue management is not new, but the need for global conclusion on the impact of crop residue management on the agro-ecosystem linked to local pedo-climatic conditions has become apparent with an increasing amount of studies showing a diversity of conclusions. This study specifically focusses on temperate climate and loamy soil using a seven-year data set. Between 2008 and 2016, we compared four contrasting residue management strategies differing in the amount of crop residues returned to the soil (incorporation vs. exportation of residues) and in the type of tillage (reduced tillage (10 cm depth) vs. conventional tillage (ploughing at 25 cm depth)) in a field experiment. We assessed the impact of the crop residue management on crop production (three crops—winter wheat, faba bean and maize—cultivated over six cropping seasons), soil organic carbon content, nitrate (${\mathrm{NO}}_{3}^{-}$), phosphorus (P) and potassium (K) soil content and uptake by the crops. The main differences came primarily from the tillage practice and less from the restitution or removal of residues. All years and crops combined, conventional tillage resulted in a yield advantage of 3.4% as compared to reduced tillage, which can be partly explained by a lower germination rate observed under reduced tillage, especially during drier years. On average, only small differences were observed for total organic carbon (TOC) content of the soil, but reduced tillage resulted in a very clear stratification of TOC and also of P and K content as compared to conventional tillage. We observed no effect of residue management on the ${\mathrm{NO}}_{3}^{-}$ content, since the effect of fertilization dominated the effect of residue management. To confirm the results and enhance early tendencies, we believe that the experiment should be followed up in the future to observe whether more consistent changes in the whole agro-ecosystem functioning are present on the long term when managing residues with contrasted strategies.

Multimodal images fusion has the potential to enrich the information gathered by multi-sensor plant phenotyping platforms. Fusion of images from multiple sources is, however, hampered by the technical lock of image registration. The aim of this paper is to provide a solution to the registration and fusion of multimodal wheat images in field conditions and at close range. Eight registration methods were tested on nadir wheat images acquired by a pair of red, green and blue (RGB) cameras, a thermal camera and a multispectral camera array. The most accurate method, relying on a local transformation, aligned the images with an average error of 2 mm but was not reliable for thermal images. More generally, the suggested registration method and the preprocesses necessary before fusion (plant mask erosion, pixel intensity averaging) would depend on the application. As a consequence, the main output of this study was to identify four registration-fusion strategies: (i) the REAL-TIME strategy solely based on the cameras’ positions, (ii) the FAST strategy suitable for all types of images tested, (iii) and (iv) the ACCURATE and HIGHLY ACCURATE strategies handling local distortion but unable to deal with images of very different natures. These suggestions are, however, limited to the methods compared in this study. Further research should investigate how recent cutting-edge registration methods would perform on the specific case of wheat canopy.

Agroecology represents a pertinent option to improve the sustainability of agriculture. To promote its application, agroecological concepts should be taught to students and professionals in the agricultural sector. However, most agricultural courses are not adapted to teach these concepts due to little interactivity or interdisciplinarity, and a lack of a systems approach to farm management. Serious games help to fill these gaps by simulating complex models in which players can learn by doing. We thus developed a serious computer game, called SEGAE (SErious Game for AgroEcology learning), which represents a mixed crop-livestock farm and assesses impacts of farming practices on indicators related to environmental, economic, and social sustainability. Its pedagogical interest was evaluated through two types of surveys given to university students who played the game during a one-week workshop: A knowledge survey on agroecology, and a feedback survey based on flow theory. Results showed that students increased their knowledge of agroecology significantly, particularly those who had had little knowledge of crop production. More than 86% of the students enjoyed the game, appreciating its interaction and feedback. Thus, SEGAE is an interesting tool to help students acquire knowledge of agroecology in a fun way by facilitating interdisciplinary and collaborative learning.