随着遥感技术的发展和图像分辨率的提高，高分辨率遥感图像已经被广泛应用于社会的各个领域。高分辨率遥感图像具有细节信息丰富、地理结构明显等的特点，所以遥感图像处理方法应运而生。最重要最常用的遥感图像处理方法是图像的分割，分割出的图像叫做图像样本。样本将存入专用数据库，这些存入数据库的样本供以后与新图像对比分析。样本是图像内容的形状、纹理、灰度等详细信息的表述，样本分割的好坏也将直接影响到后续分析、分类、识别和解译等的精度。同时样本数据库系统的好坏也会影响样本使用的方便性。由于遥感图像的复杂性和图像分割自身的不确定性，遥感图像的准确分割一直是大家非常关注的问题。 上海交通大学遥感科学实验室依托上海市科委重大项目“基于影像内容的自动搜索和特定目标的变化检测与更新技术研究”（NO. 055115018），设计实现了遥感图像处理系统ELU。本文结合该项目而进行，对高分辨率遥感图像样本的提取、分类、和样本数据库做了深入的研究，并完成了ELU系统中样本的提取和样本库模块。本文内容主要介绍三大部分：第一部分样本的提取；第二部分介绍样本提取中的图像分割方法；第三部分：样本库系统的设计和功能实现

With the progress of RS technology and improve of image resolution，the high resolution remote sensing image has been put into wide application by different fields of the society. The high resolution remote sensing image has many characteristics like its detailed information capacity and obvious geography structure, so the process method of remote sensing image emerged. The image segmentation is the main and important process of remote sensing image. The image patterns was obtained by image segmentation. These image samples saved into database and comparison with new remote sensing image. It represent the shape content of an image. the image segmentation method influences the identification and analysis of new remote sensing image, and the image sample database influences convenience using too. Because of the complexity of remote sensing image and the uncertainty of image segmentation, great attention has been paid to the remote sensing image segmentation. A remote sensing image processing system named ELU is designed and developde by Remote Sensing Laboratory of Shanghai Jiao Tong University, supported by the research project of Content-based Search in Image and Change Detection and Auto-updating of Special Target, which is administered by Shanghai Science and Technology Committee. The remote sensing image extraction of image sample and management of image sample dababase is relatively deeply studied in this thesis supported by the project. There are three important parts in this thesis. The first part of this paper, the extraction of remote sensing iamge sample is described. In the second part, image segmentation method with extraction of remote sensing image samples are described. In the third part, the design and implementation management system of the remote sensing image sample database is detailed described development.