Understanding complex process sensor signal may be the source for deep Learning: A smart IoT delving into a tomato weight sensor

文献类型： 外文期刊

第一作者： Wang, Xiaotong

作者： Wang, Xiaotong;Shi, Junlei;Sun, Shaokun;Wang, Dandan;Li, Zhulin;Sun, Zhouping;Li, Tianlai;Sun, Shaokun;Wang, Dandan;Sigrimis, Nick

作者机构：

关键词： Smart sensing; Plant modelling; Hybrid twinning; Context understanding; Smart greenhouses

期刊名称：COMPUTERS AND ELECTRONICS IN AGRICULTURE （ 影响因子：8.9； 五年影响因子：9.3 ）

ISSN： 0168-1699

年卷期： 2025 年 231 卷

页码：

收录情况： SCI

摘要： With the rapid advancements of Internet of Things (IoT) technology, understanding sensor signals has become the key to acquiring plant physiological data and realizing the concept of "speaking plants". This study developed a real-time and intelligent monitoring system using IoT-enabled weight sensors to continuously track the total weight of the shoot and root systems of tomato plants. A newly introduced variable, the photosynthetic leaf area index (LAIp), which plays a crucial role in capturing solar radiant heat, was used to balance energy components in the modeling of tomato transpiration processes. In this framework, a canopy transpiration model was developed and trained using data collected by sensors to enhance the online accuracy of tomato transpiration predictions. Supported by this model, in-depth analysis of weight sensor signals enabled the non-destructive, data-driven, and real-time extraction of LAI and LAIp. The parameter LAIp, the radiation absorbing portion of LAI, also serves as a measure for photosynthetic efficiency when related to the concurrently assessed growth rate. The results demonstrated strong consistency between simulated and measured values of LAI and LAIp (LAI simulation: R2 = 0.99, NRMSE = 0.04; LAIp simulation: R2 = 0.98, NRMSE = 0.04). This new WEP (Weight Effectiveness Photosynthesis) based modeling provides the required biofeedback for robust crop management, expanding the validity of model precision in cases when non-optimal environmental conditions dominate due to economic considerations. This study reveals that multiple critical insights can be obtained from complex process signals monitored by a single sensor, thereby providing a new perspective on multi-parameter estimates of plant growth and management.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序