【作者】 王琼；

【导师】 李少昆；

【作者基本信息】 石河子大学 ， 作物栽培学与耕作学， 2013， 博士

【摘要】 新疆是我国主要的产棉区，棉田土壤质量及其评价对生产者和管理者具有重要的意义。本研究利用遥感技术对栅格数据的管理优势，将多时相棉田遥感信息中提取的NDVI等指标与实地获取的棉田理化数据相结合，探索棉田土壤属性信息迅速无损获取的理论依据和研究方法，构建基于遥感技术的棉田土壤质量评价指标体系，并进行评价和检验。拓展了遥感技术在棉田管理方面的应用领域和范围，为棉田管理的信息化提供了新的技术支持。研究利用2009年至2011年的TM数据和HJ卫星数据以及数字化图件、田间实测数据，参考常规土壤质量评价的物理指标、化学指标、生物指标，结合研究区棉田植株生长发育的特点，运用遥感影像分析软件和地理信息系统处理平台，综合使用传统统计学方法、地统计方法、遥感影像处理方法等技术手段（监督分类、相关分析、主成分分析、矢量栅格相互转换等），对获取信息进行充分挖掘，筛选出适合评价棉田土壤质量的指标，构建基于遥感技术的棉田土壤质量评价指标体系，并对棉田土壤质量进行评价，主要研究结果如下：一、通过多种遥感图像分析技术，实现了棉田基本信息的获取，为棉田土壤质量监测奠定了基础。利用不同作物种植类型及其生长发育特点在遥感监测信息上的差别，提取棉田种植区域和面积信息；结合裸土质地样点，实现棉田土壤质地的分类；通过DEM高程数据的分析获得棉田土壤地形状况；并将全国土壤图与研究区卫星数据校正叠加后，提取了研究区土壤类型数据。研究证明用国产遥感影像对棉田土壤质量评价物理指标的获取是可行、有效的。二、研究分析了土壤养分空间变异性和棉花苗期至花铃期的生长状况，利用遥感获取的归一化植被指数NDVI进行密度分割。得到棉田长势分级图，研究表明该分级图能够揭示相应地区棉花的长势状况，并且与土壤有机质、碱解氮含量、速效钾含量、含量在51-63mg/kg范围的速效磷在空间布局上基本一致。根据多时相的NDVI数据能够得到长势持续增加的棉田，并以此作为判断集约化种植措施下棉田土壤质量的依据，为利用遥感技术对棉田土壤质量的监测和评价奠定了理论基础。三、研究表明棉花产量形成关键生育时期的遥感监测指标——植被指数能够用于预测棉花产量，采用多时相指数对棉花产量空间布局进行反演的研究表明，遥感技术获取的植被指数，特别是比值植被指数（RVI）、归一化植被指数（NDVI），不仅与土壤有效氮、有机质等土壤养分具有显著的相关性，而且与棉花单铃重、单产、小区总产具有显著相关性。采用遥感多时相指数能够将棉花产量作为评价棉田土壤质量的关键指标之一，提高利用遥感技术评价棉田土壤质量的科学性和准确性。四、研究构建了基于遥感技术的棉田土壤质量评价体系。该体系利用多时相遥感监测数据与棉花产量、棉田土壤理化性状的相关性，结合常用土壤养分关键指标，通过多种分析方法对棉田土壤质量进行了综合评价，初步构建的棉田土壤质量评价指标体系包括：1、物理指标：作物种植类型、土壤质地、土壤类型以及地形条件；2、化学指标：土壤有机质、碱解氮、速效磷、速效钾含量；3生物指标：棉田关键生育时期多时相植被指数、产量。实测数据的检验证实，本研究构建的基于遥感技术的棉田土壤质量评价体系具有快速、精确、便于管理的特点，适合大规模棉田信息化管理的需求。更多还原

【Abstract】 Xinjiang is the major region of cotton production in China, It has great significance toevaluate the soil quality of cotton fields for producers and managers.This research take theadvantages of grid data management by remote sensing technology, combined with evaluationindex such as NDVI, which extracted by remote sensing technology, and a small amount ofphysical and chemical index of measuring point data. Research a quick, convenient andlossless method to obtain the soil attribute information of the cotton fields, using themulti-spectral remote sensing data to construct cotton soil quality evaluation index system,explore the application of remote sensing technology on cotton soil quality evaluation.From2009to2011, TM data, HJ satellite data, digital maps, field measured data wereused; years study about soil evaluation indexes, include physical indicators, nutrientindicators, biological indicators in the study area were referenced; and remote sensing imageprocessing methods such as: supervised classification, correlation analysis, principalcomponent analysis, vector and raster conversion and other technical means were utilized,using geographical information system and remote sensing image analysis softwareprocessing platform to get information for cotton fields in soil quality evaluation index system,and to evaluate soil quality, the main research results are as follows:(1) Achieved basic information of cotton field by different kinds of RS image analysismethods, establishes the foundation for monitoring cotton field’s soil quality. Study thedifference spectral information between various land use patterns and growing characteristics,and extract information about position and area of cotton field in study area; use real samplesto classified the soil texture; obtain the topography information by DEM data; and gain thesoil type classification by layer-stacking the National Soil map and remote sensing image ofstudy area. The conclusion is that it is feasible and effective to extract physical indicators ofcotton soil quality evaluation by indigenous satellite.(2) Study the space distribution of soil fertility and the growing condition of crops, findout that: the normalized difference vegetation index NDVI classification map is obtained bydensity slicing can reveal the corresponding region of cotton growing conditions, and soilorganic matter, alkali-hydro nitrogen, rapidly-available potassium content, content of51-63mg/kg on the range of available P2O5in the spatial layout are basically identical.Multi-temporal NDVI data can reveal the growing situation of cotton field and it can be aaccording to evaluate the soil quality, meanwhile, it laying a good theory foundation formonitoring and evaluating soil quality.(3) It is concluded that: the key period of cotton growing for production can implyimportant information to predicted yield. Using Multi-temporal vegetation index and somesample data to achieve inversion model. Results shows that: vegetation index, especially theratio vegetation index(RVI) and normalization vegetation index (NDVI), are not only haverelationships with soil nutrient such as available nitrogen, soil organic matter, but also hassignificant correlation with cotton single boll weight, yield and total output of residential area.It will become one of the most important indexes to evaluate cotton field soil quality, andimprove the scientificalness and veracity of the evaluation. (4) Structure the cotton Soil quality evaluation indicator system, tentatively. The systemmake use of the relationship between multi-temporal remote sensing, cotton yield, and soilphysical and chemical properties, comprehensive utilization of soil fertility index, use variedanalysis method evaluated the cotton soil quality. The evaluation index system including:1,physical indexes: land use type, soil texture, soil type and topography conditions;2, nutrientindexes: soil organic matter, alkali solution nitrogen, available phosphorus, availablepotassium content;3, biological indicators: key growth period long, vegetation index, thecotton yield. It also test and verify that the indicator system is rapid, accurate and manageablewhen evaluating and it will be appropriate for informatization management of cotton field.更多还原