基于机器视觉的葡萄生长调控机理与虚拟建模方法研究

The grape growth regulation mechanism and visualized modelling method based on machine vision is aim to apply systematic model theory and experimental validation methods at the nutrition growth and prarpgate period on main cultivating variety of grape in Hexi corridor of Gansu, through several years of experiments for grabbing images of grape stem and leaf and sampling data of grape structual morphology, explore and proclaim interactive mechanisms between physical growth models of leaf, stem and spikelet of red-globe grape and agricultural ecological factors of water,fertilizer and heat from environment. Combining with image processing technology, grape water requirement compurized model and water content recognition approach are determined, and grape growth dynamic models and organ morphological models are set up and simulated. Then grape growth topological strcture, organ morphological builtup simulation approach and cluster of grape forming mechanism are implemented. Furthermore, the coordination function relation between water and fertilizer with grape growth morphology, virtual simulation approaches of grape stalk and vine shoots and tendrils will be finished. finally, virtual grape forming changes and reality method will be studied under rain or wind condition. Above all of experimental study on growth mechanism simulation and virtulization modelling approaches will provide new technology and methods for pricision horticultural agriculture, which has an important significance for agricultural cultivation informatization and intelligentization both theory and practice development. The experimental results will be widely used in other horticultural plant growth simulation research.

基于机器视觉的葡萄生长调控机理和虚拟建模方法研究，旨在基于甘肃河西走廊地区主栽葡萄延迟收获栽培条件下，探索应用系统模型理论与实验验证相结合的研究方法，在葡萄物候期实验采集多轮葡萄茎蔓、枝叶与花穗图像样本和结构形态数据，揭示红提葡萄茎叶穗生理生长规律与光热水肥等农业生态环境相互作用调控机理；结合图像处理技术，研究葡萄需水量模型和水分动态智能判识方法，实验研究葡萄生长动态过程模型与各器官形态生长模型模拟。探索葡萄茎蔓枝叶穗虚拟生长拓扑结构、器官结构模型构建和虚拟生长建模方法，实验探索建立影响葡萄器官形态发育的水肥调节关系规律，研究葡萄主干蔓和枝叶果穗等虚拟生长技术方法，并研究在风雨条件下的葡萄生长形态变化规律和虚拟实现机理，这些葡萄生长发育机理和虚拟生长建模方法研究将为精准园艺农业提供新技术和方法。本研究对农业栽培信息化和智能化学科发展具有重要理论和实际意义，研究成果可推广应用于其他园艺作物

针对西部现代智慧农业科技研究热点问题，并结合甘肃特色农业生产实际，本项目探索进行了基于机器视觉的葡萄虚拟建模与生长调控机理研究。项目在甘肃农业大学葡萄园实验地和甘肃永登教场沟温室实验地选择主栽红地球葡萄为研究对象，在2016-2020年期间，采用田间正交实验、3D扫描、数据分析、图像处理、模型模拟、虚拟仿真、比较实验验证技术相结合的科学方法，探索了甘肃主栽红地球葡萄分别在露地和温室栽培条件下其生长发育期的叶枝穗的形态建成模型与器官形成规律；运用数字图像处理技术、3维扫描成像技术和虚拟仪器技术方法等进行葡萄生长发育过程模拟和可视化仿真；采用智能图像分类回归技术方法进行葡萄叶片含水量、含氮量与含钾量等矿物质与水分含量无损检测分析方法研究；针对葡萄生长过程的监测问题，进行了基于无人机、机器视觉、无线物联网等天地空一体化监测技术研究；还对甘肃露地红地球葡萄和温室葡萄生长的水、肥、光需求优化调控规律进行了多轮正交实验和虚拟生长建模及可视化验证研究。本项目研究实施期间，先后发表了11篇科研论文和8篇硕士毕业论文，培养提高了8位硕士生和6位青年教师的科研能力，本研究对于农作物模型模拟与精准农业技术提供了园艺作物的葡萄建模仿真领域的新内容，促进了现代园艺作物表型组学与栽培信息技术进步，对智慧农业和数字农业以及农业电气化与信息化发展具有重要科学意义。