GIS辅助遥感影像分类面积提取技术和方法研究

【作者】 刘琦；

【导师】 胡青泥； 赵春江；

【作者基本信息】 大连理工大学 ， 机械制造及自动化， 2004， 硕士

【摘要】 国家863项目“重大行业3S应用示范——农业”是为了发展“精准农业”开展的科研项目。本项目研究的目标之一是开发基于3S(GPS、GIS、RS)技术的作物品质与肥水调优栽培系统。为了开发实用的GIS系统，增强系统遥感影像预处理模块功能，提出本课题要求——“GIS辅助遥感影像分类面积提取技术和方法研究”。 本论文在研究“3S”、模式识别、遥感影像处理理论基础上，提出了在GIS辅助下的遥感影像分类提取作物面积的方法。论文中着重论述遥感影像从转入到预处理，到神经网络分类器分类，再到分类后处理，以及面积统计等环节的设计及实现。 本课题遥感影像模式分类所基于BP神经网络分类器，在GIS功能辅助下，提高了地物分类的精度，能够为本系统的决策及其他操作打下基础。本课题的研究为今后GIS下集成遥感影像处理功能，实现GIS辅助影像分类，提供理论依据和实践经验。 本课题实现了RS作为系统的重要信息源与GIS的集成，克服了传统提取作物面积时必须使用遥感影像处理软件的缺陷。遥感影像处理模块能够实现信息控制、查询、统计等多种功能。 通过采用WebGIS技术、ComGIS技术、组件技术以及公式识别等技术，实现了B／S网络模式的作物品质与肥水调优栽培系统。本系统的开发填补了我国农业科学中作物品质遥感领域的空白，达到国内领先水平。本系统已通过国家863课题专家组的验收，即将在实验区投入使用。更多还原

【Abstract】 "3S experiment in agriculture", a national"863 "scientific item, is being carried out to develop "Precision Agriculture". One task of this item is to establish a system, Crop Quality Monitoring & Decision Making Supporting in Optimizing Fertilization and Irrigation (QM&OFI). In order to complete QM&OFI and strengthen the module of RS-Image pretreating, a task is brought forward. That is "Research in RS-Image Classification assisted by GIS to estimate the area of crop".Based on studying theories such as "3S", Pattern Recognition and RS-Image processing, the way to estimate area is designed and realized. This paper lays stress on the realization of RS-Image classification, from image importing to pretreating, from image classifying to post-treating and area estimating.The classifier of BPNN assisted by GIS improves precision of RS-Image classification and will make decision or other operations reliably. The research provides great references in experience and theory for the development of integration of RS-Image processing in GIS.The software of RS-Image processing, which is indispensable in traditional classification, isn’t needed any more in this system when RS as the important information origin integrates with GIS. The RS-Image pretreating module has various functions, such as information management inquiry and statistic etc.This paper realized QM&OFI, a B/S net mode system, based on technologies of WebGIS,ComGIS and formula recognition etc. At present, QM&OFI is the first system applying RS to monitoring crop quality in agriculture in China and plays a leading role. The system has been accepted by experts, and is going to be used in test.更多还原