当前位置: X-MOL 学术Precision Agric. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Forecasting seasonal plot-specific crop coefficient (Kc) protocol for processing tomato using remote sensing, meteorology, and artificial intelligence
Precision Agriculture ( IF 6.2 ) Pub Date : 2022-05-23 , DOI: 10.1007/s11119-022-09910-6
Ran Pelta , Ofer Beeri , Rom Tarshish , Tal Shilo

The crop coefficient (Kc) is a key parameter in irrigation scheduling decision-making and depends on local conditions, e.g., crop type, weather, and topography. Kc protocol or curve, and the derived growth stages, describe generally the expected behavior of the crop during the growing season and growers use it to different extents. However, Kc protocols are usually experimentally determined and hence spatially limited. This study shows an approach to generate an estimated plot-specific Kc protocol in a more cost-effective way that is not spatially limited. To that end, data for almost 600 commercial processing tomato plots were collected. The data included the normalized difference vegetation index (NDVI) from Sentinel-2 and Landsat-8, meteorological data, and plot properties such as country, and the season start date. Then, an artificial intelligence model was trained on the 2017–2019 growing seasons and validated for 2020. At the beginning of the season, the model estimated the crop behavior in terms of Kc for the entire season. Additionally, a piecewise regression model was employed to estimate the crop growth stages in terms of days from the season start. The results of this study show improvement in both Kc and growth stage estimation, compared to experimental Kc protocols. The results can help design the irrigation regime (when and how much irrigation is needed) at the plot level and thus improve the ability to allocate the required water amounts between plots in real-time and even to plan it before the season starts.



中文翻译:

使用遥感、气象学和人工智能预测用于加工番茄的季节性特定作物系数 (Kc) 协议

作物系数 (Kc) 是灌溉调度决策中的一个关键参数,取决于当地条件,例如作物类型、天气和地形。Kc 协议或曲线,以及衍生的生长阶段,一般描述了作物在生长季节的预期行为,种植者在不同程度上使用它。然而,Kc 协议通常是通过实验确定的,因此在空间上是有限的。本研究展示了一种以不受空间限制的更具成本效益的方式生成估计的特定于地块的 Kc 协议的方法。为此,收集了近 600 个商业加工番茄地块的数据。数据包括来自 Sentinel-2 和 Landsat-8 的归一化差异植被指数 (NDVI)、气象数据以及国家和季节开始日期等地块属性。然后,人工智能模型在 2017-2019 生长季节进行了训练,并在 2020 年进行了验证。在季节开始时,该模型以 Kc 为单位估计了整个季节的作物行为。此外,采用分段回归模型以从季节开始的天数来估计作物生长阶段。与实验性 Kc 协议相比,本研究的结果表明 Kc 和生长阶段估计都有所改善。结果可以帮助设计地块级别的灌溉制度(需要灌溉的时间和多少),从而提高在地块之间实时分配所需水量的能力,甚至可以在季节开始之前进行计划。该模型根据整个季节的 Kc 估计作物行为。此外,采用分段回归模型以从季节开始的天数来估计作物生长阶段。与实验性 Kc 协议相比,本研究的结果表明 Kc 和生长阶段估计都有所改善。结果可以帮助设计地块级别的灌溉制度(需要灌溉的时间和多少),从而提高在地块之间实时分配所需水量的能力,甚至可以在季节开始之前进行计划。该模型根据整个季节的 Kc 估计作物行为。此外,采用分段回归模型以从季节开始的天数来估计作物生长阶段。与实验性 Kc 协议相比,本研究的结果表明 Kc 和生长阶段估计都有所改善。结果可以帮助设计地块级别的灌溉制度(需要灌溉的时间和多少),从而提高在地块之间实时分配所需水量的能力,甚至可以在季节开始之前进行计划。与实验性 Kc 协议相比。结果可以帮助设计地块级别的灌溉制度(需要灌溉的时间和多少),从而提高在地块之间实时分配所需水量的能力,甚至可以在季节开始之前进行计划。与实验性 Kc 协议相比。结果可以帮助设计地块级别的灌溉制度(需要灌溉的时间和多少),从而提高在地块之间实时分配所需水量的能力,甚至可以在季节开始之前进行计划。

更新日期:2022-05-24
down
wechat
bug