GreenLab模型20余年研究回顾与展望  

Over Two Decades of Research with Greenlab Model

作　　者：康孟珍[1,2] 王秀娟 华净[1,3] 胡包钢 王飞跃[1,2] De Reffye Philippe Kang Mengzhen;Wang Xiujuan;Hua Jing;Hu Baogang;Wang Feiyue;De Reffye Philippe(The State Key Laboratory of Management and Control for Complex Systems,Institute of Automation,Chinese Academy of Sciences,Beijing 100190,China;School of Artificial Intelligence,University of Chinese Academy of Sciences,Beijing 100049,China;Beijing Engineering Research Center of Intelligent Systems and Technologies,Beijing 100190,China;National Laboratory of Pattern Recognition,Institute of Automation,Chinese Academy of Sciences,Beijing 100190,China;UMR AMAP,International Cooperative Research and Development of Agriculture(CIRAD),Montpellier F-34398,France)

机构地区：[1]中国科学院自动化研究所复杂系统管理与控制国家重点实验室,北京100190 [2]中国科学院大学人工智能学院,北京100049 [3]北京市智能化技术与系统工程技术研究中心,北京100190 [4]中国科学院自动化研究所模式识别国家重点实验室,北京100190 [5]法国农业发展研究中心,植物学与植物构造模型联合实验室,蒙彼利埃F-34398

出　　处：《农业大数据学报》2021年第3期3-12,共10页Journal of Agricultural Big Data

基　　金：国家自然科学基金面上项目(62076239);中国科学院与泰国科技发展署合作研究资助项目(GJHZ2076);中国科学院战略性先导科技专项(A类)(XDA20030102)。

摘　　要：【有关概念】GreenLab是器官尺度的植物功能结构模型(Functional-Structural Plant Model,FSPM),采用离散动态系统的形式来描述植物的生长和发育过程,包括植物生物量产生和分配,以及结构形成等,是融合植物学、数学、农业、计算机、自动化领域学科知识的通用模型。【目前研究现状】自1998年以来,基于中法合作,围绕GreenLab模型发展了包括双尺度自动机理论、分枝结构植物的参数反求方法、随机的植物功能结构模型以及理论计算、植物快速建模与可视化算法,开发了SciLab以及MatLab环境下的作物生长模拟与拟合软件,以及基于c++的面向复杂植物计算的软件。目前,GreenLab模型已应用于玉米、小麦、黄瓜、番茄、油菜、菊花、松树、枫树等具有不同特点的几十种植物,涵盖的植物类型从草本作物到复杂的树木。模型特色在于可通过观测植物的器官生物量和数量等数据,反求影响生物量产生和分配的模型内部源库参数;对于单枝或分枝结构、确定性或随机性结构,均能采用通用的器官尺度的数据进行模型校准。【本文的内容概括】本文回顾GreenLab模型的发展历程及其最新进展,介绍模型的基本概念和主要方法,包括双尺度自动机、器官序列,以及通用的植物拟合目标。详细介绍了GreenLab模型中所包含的结构模型(器官数量的计算、器官产生的随机性模拟等)、功能模型(植物和器官需求、生物量产生和分配、器官生长等),以及二者相结合进行参数反求的计算方法。【展望】随着植物表型技术的成熟和普及,GreenLab模型可用于平行农业种植系统的构建,服务植物与环境关系的深入理解,以及生产管理与控制中的智能决策支持。The GreenLab model is an organ-level Functional-Structural Plant Model(FSPM),which simulates plant growth and development processes with the discrete dynamic system,including biomass production,partitioning,and structure formation.It is a generic FSPM that integrates multi-disciplinary knowledge from botany,mathematics,agronomy,computer science,and automation science.Sino-French cooperation around GreenLab since 1998 has led to the development of new methods,algorithms,and software.These include a dual-scale automaton,parameter inversion for plants with branching structure,stochastic FSPM with theoretical computation,plant fast modelling and visualization,a plant growth modelling and fitting tool in Scilab and Matlab,and a simulator for complex structure in c++.The GreenLab model has been applied on dozens of plants with their own features,including maize,wheat,cucumber,tomato,rapeseed,pine tree,and maple tree,covering plants ranging from herbaceous crops to complex trees.The model is characterized by the fact that its source-sink parameters affecting the biomass production and partitioning can be inversely estimated through the measured organ biomass and quantity.It is applicable for single stem or branching structures,deterministic or stochastic cases,with common organ-level target data for parameter identification and model calibration.This paper reviews the development history and recent advances of the GreenLab model and presents the basic concepts and key methods.These include dualscale automaton,organ series,the generic plant fitting.It gives some details on the structural model(the computation on organ quantities and the stochastic simulation on organ production)and the functional model(demand of organ and plant,biomass production and allocation,and organ growth).With the availability of plant phenotype technologies,GreenLab can be used for building parallel agricultural system,supporting the deep understanding of the plant-environment interaction,and the intelligent decision support for management and co

关 键 词：GreenLab 植物功能结构模型 离散动态模型 参数校准 平行农业系统 

分 类 号：TU205[建筑科学—建筑设计及理论]