Abstract

In recent years, a new branch of plant physiology, plant phenomics, which focuses on identifying patterns of organization and changes in plant Phenomes, i.e., physical and biochemical characteristics, considered as a set of phenotypes of a plant organism, has emerged. Phenomics is a postgenomic discipline that actively uses the achievements of the genomic era and bioinformatics. It supplements them with standardized and statistically significant factual material on phenotypes with a high degree of detail. The technique of obtaining and analyzing information about phenotypes in phenomics is called phenotyping. High-performance phenotyping, providing digital automated analysis of large data samples, has become widespread. Recent progress in high-performance phenotyping has been associated with the development of image registration systems in various spectral regions, approaches to cultivating plant objects under standardized conditions, sensory technologies, robotics, and methods for data processing and analysis, such as computer vision and machine learning (artificial neural network). Phenomics technologies have a high information content analysis, surpassing human capabilities, performing measurements in the hyperspectral range using X-ray tomography and ultra-precise “thermal” images, and have a number of other low-invasive and precision approaches. Arrays of data obtained using phenomics technologies are recorded and processed automatically and are free from the problems of subjective assessment and inadequate statistical processing. It is assumed that phenotyping will allow for the creation of digital models of the vital activity processes and the “formation” of plant productivity at the organism level in connection with the dynamics of transcriptomes, proteomes, and metabolomes. Phenomics helps researchers transform a large amount of information received from modern sensors into new knowledge using computer data processing and modeling, reducing the distance from basic science to the practical application of results in crop production and breeding. Phenotyping is actively developing both in laboratory and in greenhouse conditions as well as on open agricultural sites, forests, and in real natural phytocenoses. The review analyzes the current state of plant phenomics with a focus on technical aspects, in particular, the design of hardware-software phenotyping complexes, i.e., phenomics platforms, as well as the use of neural networks in phenotyping of plant organisms.

中文翻译：

---

植物物候学：基础知识，软件和硬件平台以及机器学习

摘要

近年来，植物生理学的一个新的分支工厂型组学，其重点是确定组织的模式和改变植物Phenomes，即物理和生化特性，被视为一组表型的植物生物的，又出现了。Phenomics是一门后基因组学科，它积极利用基因组时代和生物信息学的成就。它以高度详细的表型来补充标准和具有统计意义的事实材料。获取和分析表型学中有关表型的信息的技术称为表型。高性能表型提供对大型数据样本的数字自动分析的软件，已经变得广泛。高性能表型的最新进展与各种光谱区域中的图像配准系统的开发，标准化条件下植物对象的培养方法，传感技术，机器人技术以及数据处理和分析方法（例如计算机视觉和机器）相关联学习（人工神经网络）。物变技术具有很高的信息含量分析能力，超越了人类的能力，使用X射线断层扫描和超精确的“热”图像在高光谱范围内进行测量，并具有许多其他低侵入性和精确性的方法。使用表态学技术获得的数据数组将被自动记录和处理，并且没有主观评估和统计处理不充分的问题。假定表型将允许在转录水平，蛋白质组和代谢组动力学方面建立重要活动过程的数字模型以及在生物体水平上植物生产力的“形成”。Phenomics帮助研究人员使用计算机数据处理和建模将从现代传感器接收到的大量信息转化为新知识，从而缩短从基础科学到将结果实际应用于作物生产和育种的距离。在实验室和温室条件下，以及在开放的农业场所，森林和真正的天然植物素中，表型分型正在积极发展。这篇综述分析了植物表型的现状，侧重于技术方面，特别是硬件-软件表型复合体（即表型平台）的设计，以及神经网络在植物生物表型中的应用。在实验室和温室条件下，以及在开放的农业场所，森林和真正的天然植物素中，表型分型正在积极发展。这篇综述分析了植物表型的现状，侧重于技术方面，特别是硬件-软件表型复合体（即表型平台）的设计，以及神经网络在植物生物表型中的应用。在实验室和温室条件下，以及在开放的农业场所，森林和真正的天然植物素中，表型分型正在积极发展。这篇综述分析了植物表型的现状，侧重于技术方面，特别是硬件-软件表型复合体（即表型平台）的设计，以及神经网络在植物生物表型中的应用。