论文标题
使用开放数据和开源软件朝着数据驱动的精确农业
Towards Data-Driven Precision Agriculture using Open Data and Open Source Software
论文作者
论文摘要
储层计算是预测湍流的有力工具,其简单的架构具有处理大型系统的计算效率。然而,其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间,然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用,该系统模拟了湍流的若干特征。我们表明,使用径向基函数作为非线性投影器,即使只有部分观测并且不知道控制方程,也能稳健地捕捉复杂的系统非线性。最后,我们表明,当测量稀疏、不完整且带有噪声,甚至控制方程变得不准确时,我们的网络仍然可以产生相当准确的预测,从而为实际湍流系统的无模型预测铺平了道路。
Information and communications technology (ICT) within the agricultural sector is characterized by a widespread use of proprietary data formats, a strong lack of interoperability standards, and a tight connection to specific hardware implementations resulting from vendor lock-in. This partly explains why ICT has not yet had its full impact within the domain. By utilizing the vast amount of publicly available open data, ranging from topographic maps to multispectral satellite images, the economically and environmentally optimal farming practices can be advanced beyond state of the art. This paper addresses the potential of applying publicly available information sources to improve crop production, with emphasis on yield optimization. This potential is evaluated based on free public data for the growth season 2016 by examining winter wheat production for a selected region in Denmark. Data aggregation is performed by promoting opensource software tools as a foundation for decision support. That allows the farmer, or another domain expert, to query a certain crop type, merge this information with other data sets, and perform analysis on data ranging from sub-field analysis to statistics on national/regional scale. The registration of field polygons and sowed crop types for fields in Denmark, alongside with detailed geographic data and free satellite images, enable us to exploit publicly available data of high quality, which can be applied to perform further analysis.