摘要地表的地物分布是认识和了解各种自然和人文现象的基础，特别是土地利用覆盖的信息，是各种生产决策的重要依据，准确地获知这些信息具有重要的科 学和现实意义。现今，遥感技术的应用已越来越广泛，它是根据传感器接收到不 同地物发射或反射的电磁波对地物的一种远距离探测技术。通过遥感影像，按照 一些算法和规则，即可实现对不同地物的分类，遥感技术不仅监测范围广、获取 信息快、周期短，而且避免了由于天气和地域条件所带来的限制。69941
为了分析不同分类方法对地表遥感影像的分类效果,本文以黑河流域资源三 号卫星影像作为实验数据，分别采用最大似然分类法、迭代自组织数据分析算法
（ISODATA）和决策树分类法对影像进行分类,结果表明最大似然分类法和决策 树分类法精度都较高,总体精度分别为 95%和 93.75%，但通过与全球 30 m 地表 覆盖数据的比较，可以发现决策树分类法在水体和居民地分类中更接近实际值。 决策树分类不但能够充分利用地物光谱信息，而且增强了地物非光谱信息对分类 结果的作用。
毕业论文关键词：地表分类 资源三号卫星 最大似然分类法 ISODATA 决策树分类 法
Surface classification based on resource three (ZY-3)data
Abstract Feature distribution surface is knowledge and understanding of natural and human phenomena basis, particularly land use information / coverage, is an important basis for a variety of production decisions and accurately informed of this information has important scientific and practical significance. Today, the application of remote sensing technology has become increasingly widespread, it is received or reflected by different objects emit electromagnetic waves feature a remote sensing technique based on the sensor. By sensing image, according to some algorithms and rules, you can achieve the classification of different objects, not only to monitor a wide range of remote sensing technology, access to information, short cycle, and avoid limitations due to the weather and geographical conditions brought about.
To analyze the effect of different classification methods to classify surface remote sensing images, the paper resources in Heihe River Resources satellite three image as the experimental data were used to maximum likelihood classification method, iterative self-organizing data analysis algorithm (ISODATA) and decision tree classification of images the classification results show that the maximum likelihood classification and decision tree classification accuracy are high, overall accuracy was 95% and 93.75%, but by comparing the data with 30 m of surface coverage of the globe can be found in classification tree water and residents classified in closer to the actual value. Decision tree classification can not only take full advantage of the feature spectrum of information, but also enhances the role of information on non-spectroscopic feature classification results.
Key Words: surface classification Resources satellite three maximum likelihood classification ISODATA decision tree classification
目录
摘 要 I
Abstract II
目录 IV
图清单 V
表清单 V
1 绪论 1
1.1 研究目的和意义 基于资源三号ZY-3数据的地表分类研究:http://www.youerw.com/guanli/lunwen_79032.html