Search

Significant advances in weed mapping from unmanned aerial platforms have been achieved in recent years. The detection of weed location has made possible the generation of site specific weed treatments to reduce the use of herbicides according to weed cover maps. However, the characterization of weed infestations should not be limited to the location of weed stands, but should also be able to distinguish the types of weeds to allow the best possible choice of herbicide treatment to be applied. A first step in this direction should be the discrimination between broad-leaved (dicotyledonous) and grass (monocotyledonous) weeds. Considering the advances in weed detection based on images acquired by unmanned aerial vehicles, and the ability of neural networks to solve hard classification problems in remote sensing, these technologies have been merged in this study with the aim of exploring their potential for broadleaf and grass weed detection in wide-row herbaceous crops such as sunflower and cotton. Overall accuracies of around 80% were obtained in both crops, with user accuracy for broad-leaved and grass weeds around 75% and 65%, respectively. These results confirm the potential of the presented combination of technologies for improving the characterization of different weed infestations, which would allow the generation of timely and adequate herbicide treatment maps according to groups of weeds.

Assessing the spatial distribution of weeds within a field is a key step to the success of site-specific weed management strategies. Centaurea diluta (knapweed) is an emerging weed that is causing a major agronomic problem in southern and central Spain because of its large size, the difficulty of controlling it, and its high competitive ability. The main objectives of this study were to examine the spatial variability of C. diluta density in two wheat fields by multivariate geostatistical methods using unmanned aerial vehicle (UAV) imagery as secondary information and to delineate potential control zones for site-specific treatments based on occurrence probability maps of weed infestation. The primary variable was obtained by grid weed density field samplings, and the secondary variables were derived from UAV imagery acquired the same day as the weed field surveys. Kriging and cokriging with UAV-derived variables that displayed a strong correlation with weed density were used to compare C. diluta density mapping performance. The accuracy of the predictions was assessed by cross-validation. Cokriging with UAV-derived secondary variables generated more accurate weed density maps with a lower RMSE compare with kriging and cokriging with RVI, NDVI, ExR, and ExR(2) (the best methods for the prediction of knapweed density). Cokriged estimates were used to generate probability maps for risk assessment when implementing site-specific weed control by indicator kriging. This multivariate geostatistical approach enabled the delineation of winter wheat fields into two zones for different prescription treatments according to the C. diluta density and the economic threshold.

This paper aims to evaluate the spatial persistence of wild oat patches in four wheat fields over time to determine the economic feasibility of using late-season wild oat maps for early site-specific weed management (SSWM) next season. The spatial persistence of wild oat patches was analyzed by three tests: land use change detection between years, spatial autocorrelation, and analysis of spreading distance. The temporal trend of wild oat patch distribution showed a clear persistence and a generalized increase in the infested area, with a noticeable level of weed aggregation and a tendency in the new weed patches to emerge close to older ones. To economically evaluate the SSWM, five simulations in four agronomic scenarios, varying wheat yields and losses due to wild oat, were conducted. When yield losses due to wild oat were minimal and for any of the expected wheat yields, some SSWM simulations were more economically profitable than the overall application in most of the fields. Nevertheless, when the yield losses due to wild oat were maximal, all SSWM simulations were less profitable than overall treatment in all the analyzed fields. Although the economic profit variations achieved with SSWM treatments were modest, any of the site-specific treatments tested are preferred to herbicide broadcast over the entire field, in order to reduce herbicide and environmental pollution.