Institute of Information and Communication Technologies-Bulgarian Academy of Sciences, Acad. G. Bonchev Str., Bl. 2, 1113 Sofia, Bulgaria.
Sensors (Basel). 2022 Aug 31;22(17):6566. doi: 10.3390/s22176566.
Smart farms, as a part of high-tech agriculture, collect a huge amount of data from IoT devices about the conditions of animals, plants, and the environment. These data are most often stored locally and are not used in intelligent monitoring systems to provide opportunities for extracting meaningful knowledge for the farmers. This often leads to a sense of missed transparency, fairness, and accountability, and a lack of motivation for the majority of farmers to invest in sensor-based intelligent systems to support and improve the technological development of their farm and the decision-making process. In this paper, a data-driven intelligent monitoring system in a cloud environment is proposed. The designed architecture enables a comprehensive solution for interaction between data extraction from IoT devices, preprocessing, storage, feature engineering, modelling, and visualization. Streaming data from IoT devices to interactive live reports along with built machine learning (ML) models are included. As a result of the proposed intelligent monitoring system, the collected data and ML modelling outcomes are visualized using a powerful dynamic dashboard. The dashboard allows users to monitor various parameters across the farm and provides an accessible way to view trends, deviations, and patterns in the data. ML models are trained on the collected data and are updated periodically. The data-driven visualization enables farmers to examine, organize, and represent collected farm's data with the goal of better serving their needs. Performance and durability tests of the system are provided. The proposed solution is a technological bridge with which farmers can easily, affordably, and understandably monitor and track the progress of their farms with easy integration into an existing IoT system.
智能农场作为高科技农业的一部分，从物联网设备中收集大量有关动物、植物和环境状况的数据。这些数据通常存储在本地，而不会在智能监测系统中使用，以提供为农民提取有意义知识的机会。这常常导致错失透明度、公平性和问责制的感觉，并且大多数农民缺乏投资基于传感器的智能系统的动力，以支持和改善其农场的技术发展和决策过程。在本文中，提出了一种在云环境中基于数据驱动的智能监测系统。所设计的架构为从物联网设备中提取数据、预处理、存储、特征工程、建模和可视化之间的交互提供了全面的解决方案。包括来自物联网设备的流数据到交互式实时报告以及内置的机器学习 (ML) 模型。由于提出的智能监测系统，使用强大的动态仪表板可视化收集的数据和 ML 建模结果。仪表板允许用户监测整个农场的各种参数，并提供一种方便的方式查看数据中的趋势、偏差和模式。在收集的数据上训练 ML 模型，并定期更新。数据驱动的可视化使农民能够检查、组织和表示收集的农场数据，以更好地满足他们的需求。提供了系统的性能和耐久性测试。所提出的解决方案是一座技术桥梁，农民可以轻松、经济实惠且易于理解地监测和跟踪他们的农场的进展，并轻松集成到现有的物联网系统中。
