县域统计分区施肥模型构建与应用平台建设研究

发布时间：2018-04-20 00:22

本文选题：模型 + 施肥决策　；参考：《吉林农业大学》2016年博士论文

【摘要】：本文结合GPS技术、数据库技术和GIS二次开发技术,针对县级农业技术管理部门指导农户科学施肥的实际,运用可视化程序设计语言Visual Basci 6.0和Mapobjects2.4,构建县域统计分区施肥模型和应用平台,建立了符合实际的施肥模型,有效解决了科学施肥的问题。研究区域为吉林省长春市九台区,建立土壤养分数据库、地形图和土地利用现状图,利用九台区测土配方施肥的土壤养分历史数据进行插值形成土壤养分图,以及主要作物的养分需求数据,建立短信平台施肥决策智能专家系统,用户可利用手机或者能发送短信的GSM短信猫,依据格式要求发送用户地块GPS数据、作物目标平均产量等信息,短信平台接收后就可以通过系统施肥模型的运算和优化,为用户发回施肥决策信息,指导用户科学施肥。本系统是对测土配方施肥成果的有效推广,对于农户非常便捷,对解决测土配方施肥推广应用和最后“一公里”的瓶颈问题有很大作用,对解决测土配方施肥实时决策和通讯的技术问题,为农业技术加速度推广有着深远的意义。本文主要进行以下研究。1.利用神经网络对样本数据进行观测和计算,通过氮肥、磷肥、钾肥的加权,同时对玉米产量和肥料用量的实际比较,建立土壤施肥模型,最后利用GSM短信猫和中国移动公司手机接收信号相匹配的应用软件进行施肥决策,将所需要的肥量用量以短信的形式反馈到用户手机上。结果表明,这种将神经网络应用于建立土壤施肥模型,会得到良好的施肥决策效果。2.利用统计分区代替传统的人为分区方法,建立统计分区施肥模型,并结合九台区施肥决策效果分析,检验该模型的科学性和适宜性,旨在为作物施肥决策提供新的可靠决策技术方法,最后通过短信平台软件将施肥结果反馈到农户的手机。结果表明,这种基于地统计学建立的施肥决策模型计算出的高、中、低肥力下的平均施肥量,对农户施肥指导具有良好的科学性。3.本文的核心模型是统计分区施肥模型(Statistical Division Fertilization Model,SDFM),它是依据在一定区域内的某一作物常年产量和土壤养分进行统计分析,划分高、中、低产量和土壤养分的区间而建立的。根据在该区域作物产量与施肥用量的关系,确定高、中、低产量区间的氮、磷、钾肥平均量,再以土壤养分测试值进行校正调整,以此决策施肥用量。用该模型决策施肥量不受所采用土壤养分测定形态的影响,统计分区施肥模型避开了养分平衡法中难以确定的参数,应用区域作物产量的统计量划分耕地肥力级别,根据此值确定平均适宜施肥量,并对土壤养分作相同的统计分析,确定各级土壤适宜养分量,通过产量差异和土壤养分差异双重调整施肥用量。4.通过对神经网络和统计分区建立施肥模型的比较,选择统计分区模型作为唯一的决策模型嵌入到作者研发的测土配方施肥平台中,通过计算、识别、查询等方式,利用通过GSM短信猫,实现自动向农户回复决策短信。5.采用Microsoft Access建立了施肥决策系统的属性数据库,通过ArcGIS Server和ArcGIS Engine建立GIS土壤养分分布图,以及由矢量图层建立空间数据库,结合本研究所需数据的要求,并进行无缝对接。
[Abstract]:In this paper, based on GPS technology, database technology and the two development technology of GIS, according to the practice of agricultural technology management department at county level to guide farmers' scientific fertilization, the application of visual programming language Visual Basci 6 and Mapobjects2.4 is used to construct the county statistical division fertilization model and the application platform, and a practical fertilizer model is established, which can be effectively solved. The research area is the problem of scientific fertilization. The research area is nine districts of Changchun city of Jilin province. The database of soil nutrients, topographic map and the status map of land use are set up. The soil nutrient map of the soil and the nutrient requirement data of the main crops are interpolated and the nutrient requirement data of the main crops are used to establish the decision wisdom of the short message platform fertilization. In the expert system, the user can use the mobile phone or the GSM short message cat that can send short message, and send the information of the GPS data of the user block and the average yield of the crop target according to the format requirements. After the reception of the SMS platform, the application and optimization of the system fertilization model can be sent back to the user, and the user can be used for scientific fertilization. This system is the system. The effective popularization of the results of soil testing formula fertilization is very convenient for farmers, and it has a great effect on solving the bottleneck problem of the popularization and application of soil testing formula fertilization and the last "one kilometer". It has far-reaching significance for solving the problem of real-time decision-making and communication of soil testing formula fertilization, which is of profound significance for the promotion of agricultural technology acceleration. .1. is used to observe and calculate the sample data by using neural network. Through the weighting of nitrogen fertilizer, phosphate fertilizer and potash fertilizer, the soil fertilization model is established by comparing the actual comparison of Maize Yield and fertilizer amount. Finally, the application software of GSM short message cat and China Mobile mobile phone receiving signal is used to make fertilizer decision, which will be needed. The amount of fertilizer is fed back to the user's cell phone in the form of short message. The result shows that the application of the neural network to the establishment of soil fertilization model will get a good effect of fertilization decision.2.. The statistical zoning method is used to replace the traditional artificial partition method, and the statistical partition fertilization model is established, and the results are analyzed and tested in combination with the effect of fertilization decision in nine regions. The scientific and suitability of the model is designed to provide a new and reliable decision technology method for crop fertilization decision. Finally, the fertilization results are fed back to the mobile phone by the SMS platform software. The results show that the fertilizer decision model based on Geostatistics is used to calculate the average amount of fertilizer under the high, medium and low fertility, and to fertilize the farmers. The core model of.3. with good scientific nature is the statistical partition fertilization model (Statistical Division Fertilization Model, SDFM). It is based on the statistical analysis of the annual yield and soil nutrients of a certain crop in a certain region, and is established for the division of high, middle, low yield and soil nutrients. The relationship between crop yield and the amount of fertilizer applied to determine the average amount of nitrogen, phosphorus, potassium fertilizer in the high, middle and low yield regions, and then adjust and adjust the soil nutrient test value, so as to decide the amount of fertilizer. The determined parameters, using the statistics of regional crop yield to divide the fertility level of cultivated land, determine the average suitable fertilizer amount according to this value, and make the same statistical analysis of soil nutrients, determine the suitable nutrient components at all levels, and adjust the amount of fertilizer.4. through the difference of yield and the difference of soil nutrients through the neural network and statistical division. In comparison with the model of fertilizer application, the statistical partition model is selected as the only decision model to be embedded in the soil formula fertilization platform developed by the author. By calculating, identifying and querying, using the GSM short message cat to realize the automatic reply decision message.5. to the farmer, the attribute database of the fertilization decision system is established by Microsoft Access. ArcGIS Server and ArcGIS Engine are used to establish the distribution map of soil nutrients in GIS, and the spatial database is set up by the vector layer, and the requirements of the data required in this study are combined, and the seamless docking is carried out.

【学位授予单位】：吉林农业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP183;S147

【相似文献】