基于多光谱影像的县域农田质量信息提取与快速评价
发布时间：2018-08-19 12:32
【摘要】：农田质量对于保障国家粮食安全和维护社会经济的稳定与持续发展具有非常重要的意义。然而随着城市化的推进,大量优质农田被占用,造成了农田整体质量的隐形流失。为满足新形势下农田质量管理建设要求,亟需借助先进的遥感技术对农田质量进行快速评价,以便实时监测农田综合性状。农田质量内涵在随着人类知识补充和技术进步而丰富,其信息表达也越来越复杂。由于遥感数据本身的局限以及不同指标提取过程中数据源选取、数据处理、提取方法皆有差异,如何通过多光谱影像提取少数关键指标信息来表征农田质量宽泛的内涵,需要采取各种数据处理手段不断完善预测模型并通过更多的地域和尺度的研究来验证。本文以北京市大兴区为研究区,采用定性遥感和定量遥感相结合的手段,在对农田质量信息提取的基础上探索了农田质量的快速评价。研究结果表明:(1)采用基于面向对象的最邻近分类方法能够充分利用SPOT6遥感影像的高分辨率信息及其光谱特征进行分类,且通过多尺度分割可以实现某一类地物的特征显化以便于构建规则集完成信息提取,总之面向对象的分类可以从SPOT6影像中有效识别农田区域各地类;(2)TM8影像的可见光波段上的光谱值及其变化形式与表层土壤有机质含量之间存在相关关系,基本可实现农田表层土壤有机质含量的反演;(3)从区域之间可比和区域内部可比的二维角度构建基于遥感的县域农田质量评价体系,分别从农田本底条件和农田基础设施条件两方面选取遥感可获取的评价因子,可实现基于栅格模型的大兴区农田质量快速评价,最终得到了大兴区农田质量的整体分布格局。本文是在集成目前已有的农田质量所涉信息遥感提取方法的基础上,以遥感技术作为指标信息提取的手段,直接通过遥感技术可获取的指标对农田质量进行评价。该研究以追求快速的目标为主,通过少数关键指标解析了农田质量。可实时更新的评价结果能够更好地服务于农田质量的分级动态管理以及精准农业生产管理,同时可为占补平衡中的质量平衡管理提供有效参考依据。
[Abstract]:Farmland quality plays an important role in ensuring national food security and maintaining social and economic stability and sustainable development. However, with the development of urbanization, a large number of high-quality farmland is occupied, resulting in the invisible loss of the overall quality of farmland. In order to meet the requirements of farmland quality management under the new situation, it is urgent to evaluate farmland quality quickly with the help of advanced remote sensing technology in order to monitor the comprehensive characteristics of farmland in real time. The connotation of farmland quality is enriched with the supplement of human knowledge and technological progress, and its information expression is becoming more and more complicated. Because of the limitation of remote sensing data and the difference of data source selection, data processing and extraction method in different index extraction process, how to extract a few key index information by multi-spectral image to represent the broad connotation of farmland quality. All kinds of data processing methods need to be used to improve the prediction model and verify it through more regional and scale studies. This paper takes Daxing District of Beijing as the research area, adopts the combination of qualitative remote sensing and quantitative remote sensing, and probes into the rapid evaluation of farmland quality on the basis of extracting farmland quality information. The results show that: (1) based on the object-oriented nearest neighbor classification method, the high resolution information and spectral features of SPOT6 remote sensing images can be fully used for classification. By multi-scale segmentation, the features of a certain kind of ground objects can be displayed so as to construct a rule set to complete the information extraction. In short, the object-oriented classification can effectively identify the farmland regional classes from the SPOT6 image. (2) there is a correlation between the spectral value and the change form of visible light in TM8 image and the content of organic matter in the surface soil. The inversion of soil organic matter content in the surface layer of farmland can be realized basically. (3) the evaluation system of farmland quality in counties based on remote sensing is constructed from the perspective of regional and intra-regional comparability. The evaluation factors obtained by remote sensing are selected from two aspects of farmland background condition and farmland infrastructure condition respectively. The rapid evaluation of farmland quality in Daxing area based on grid model can be realized. Finally, the overall distribution pattern of farmland quality in Daxing area is obtained. On the basis of integrating the existing remote sensing methods of farmland quality information extraction, this paper uses remote sensing technology as the means of index information extraction, and evaluates farmland quality directly through the indicators that can be obtained by remote sensing technology. In this study, the quality of farmland was analyzed by a few key indicators. The evaluation results which can be updated in real time can better serve the hierarchical dynamic management of farmland quality and the production management of precision agriculture, and can also provide an effective reference basis for the quality balance management in the balance of occupation and compensation.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：S127;S158
本文编号：2191660
[Abstract]:Farmland quality plays an important role in ensuring national food security and maintaining social and economic stability and sustainable development. However, with the development of urbanization, a large number of high-quality farmland is occupied, resulting in the invisible loss of the overall quality of farmland. In order to meet the requirements of farmland quality management under the new situation, it is urgent to evaluate farmland quality quickly with the help of advanced remote sensing technology in order to monitor the comprehensive characteristics of farmland in real time. The connotation of farmland quality is enriched with the supplement of human knowledge and technological progress, and its information expression is becoming more and more complicated. Because of the limitation of remote sensing data and the difference of data source selection, data processing and extraction method in different index extraction process, how to extract a few key index information by multi-spectral image to represent the broad connotation of farmland quality. All kinds of data processing methods need to be used to improve the prediction model and verify it through more regional and scale studies. This paper takes Daxing District of Beijing as the research area, adopts the combination of qualitative remote sensing and quantitative remote sensing, and probes into the rapid evaluation of farmland quality on the basis of extracting farmland quality information. The results show that: (1) based on the object-oriented nearest neighbor classification method, the high resolution information and spectral features of SPOT6 remote sensing images can be fully used for classification. By multi-scale segmentation, the features of a certain kind of ground objects can be displayed so as to construct a rule set to complete the information extraction. In short, the object-oriented classification can effectively identify the farmland regional classes from the SPOT6 image. (2) there is a correlation between the spectral value and the change form of visible light in TM8 image and the content of organic matter in the surface soil. The inversion of soil organic matter content in the surface layer of farmland can be realized basically. (3) the evaluation system of farmland quality in counties based on remote sensing is constructed from the perspective of regional and intra-regional comparability. The evaluation factors obtained by remote sensing are selected from two aspects of farmland background condition and farmland infrastructure condition respectively. The rapid evaluation of farmland quality in Daxing area based on grid model can be realized. Finally, the overall distribution pattern of farmland quality in Daxing area is obtained. On the basis of integrating the existing remote sensing methods of farmland quality information extraction, this paper uses remote sensing technology as the means of index information extraction, and evaluates farmland quality directly through the indicators that can be obtained by remote sensing technology. In this study, the quality of farmland was analyzed by a few key indicators. The evaluation results which can be updated in real time can better serve the hierarchical dynamic management of farmland quality and the production management of precision agriculture, and can also provide an effective reference basis for the quality balance management in the balance of occupation and compensation.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：S127;S158
【参考文献】
相关期刊论文 前8条
1 乔贤哲;张超;杨建宇;朱德海;;高分辨率遥感影像中农田灌溉设施的提取[J];测绘通报;2012年S1期
2 吴克宁,郑义,康鸳鸯,申g,程道全,李玲,葛树春,李继明,穆兰,赵华甫,吕巧灵;河南省耕地地力调查与评价[J];河南农业科学;2004年09期
3 彭玉魁,张建新,何绪生,卢恩双;土壤水分、有机质和总氮含量的近红外光谱分析研究[J];土壤学报;1998年04期
4 骆东奇,白洁,谢德体;论土壤肥力评价指标和方法[J];土壤与环境;2002年02期
5 刘钦普;GIS和SPSS技术支持下的许昌市耕作土壤肥力综合评价[J];土壤;2002年02期
6 李华;毕如田;乔显亮;;基于GIS技术的耕地土壤环境质量综合评价研究——以山西省永济市为例[J];土壤;2007年04期
7 张凤荣,安萍莉,胡存智;制定农用地分等定级野外诊断指标体系的原则、方法和依据[J];中国土地科学;2001年02期
8 江华;;基于eCognition面向对象技术的高分辨率遥感影像土地利用分类——以福州琅岐岛为例[J];海峡科学;2014年08期
相关博士学位论文 前1条
1 孙家波;基于知识的高分辨率遥感影像耕地自动提取技术研究[D];中国农业大学;2014年
相关硕士学位论文 前1条
1 于晓静;基于GIS和RS技术的肇东市耕地质量评价研究[D];东北农业大学;2012年
，本文编号：2191660
本文链接：https://www.wllwen.com/kejilunwen/nykj/2191660.html
