RESUMEN
Weed is a major biological factor causing declines in crop yield. However, widespread herbicide application and indiscriminate weeding with soil disturbance are of great concern because of their environmental impacts. Site-specific weed management (SSWM) refers to a weed management strategy for digital agriculture that results in low energy loss. Deep learning is crucial for developing SSWM, as it distinguishes crops from weeds and identifies weed species. However, this technique requires substantial annotated data, which necessitates expertise in weed science and agronomy. In this study, we present a channel attention mechanism-driven generative adversarial network (CA-GAN) that can generate realistic synthetic weed data. The performance of the model was evaluated using two datasets: the public segmented Plant Seedling Dataset (sPSD), featuring nine common broadleaf weeds from arable land, and the Institute for Sustainable Agro-ecosystem Services (ISAS) dataset, which includes five common summer weeds in Japan. Consequently, the synthetic dataset generated by the proposed CA-GAN obtained an 82.63% recognition accuracy on the sPSD and 93.46% on the ISAS dataset. The Fréchet inception distance (FID) score test measures the similarity between a synthetic and real dataset, and it has been shown to correlate well with human judgments of the quality of synthetic samples. The synthetic dataset achieved a low FID score (20.95 on the sPSD and 24.31 on the ISAS dataset). Overall, the experimental results demonstrated that the proposed method outperformed previous state-of-the-art GAN models in terms of image quality, diversity, and discriminability, making it a promising approach for synthetic agricultural data generation.
RESUMEN
Integrated weed management (IWM) is currently the most appropriate and effective method of agricultural weed control. To determine the most effective strategy, it is necessary to compare the effects of different control options and their rotation. Avena fatua (common wild oat) is one of the most common and economically threatening grass weed species of cereal crops worldwide. To examine the effects of non-chemical weed management options (farmland use, delayed sowing, and summer irrigation) on control of A. fatua, we recorded coverage levels and field conditions in 41 sites during the spring growing season of winter wheat for about 10 years. A transition matrix model was then constructed to project coverage levels of A. fatua under each management option using ordinal logistic regression. The results showed that farmland use had a remarkable effect on coverage; notably, planting of paddy rice and vegetables, which respectively eliminated the effect of coverage in the previous year and facilitated rapid convergence of coverage to 0%. Thus, although 90% of fields under continuous wheat cultivation were found to be at risk of A. fatua colonization, the risk was reduced to almost 0% with rotation of effective farmland use. As summer irrigation was also effective, more than 50% of wheat fields with the option continuously converged to no risk for A. fatua colonization. When the different management cycles were repeated, the effects were observed within 3 years, with a steady state reached in less than 10 years. Overall, these results suggest that simplified monitoring data could help decision-making on IWM, thereby helping to improve the efficiency of agricultural production.
