图像分割论文：图像分割 模糊聚类 特征提取 分形维数 遥感图像.docVIP

【关键词】图像分割 模糊聚类 特征提取 分形维数 遥感图像 【英文关键词】Image Segmentation Fuzzy Clutering Feature Extraction Fractal Dimension Remote Sensing Image 图像分割论文：高分辨率遥感图像林区提取算法研究 【中文摘要】森林资源调查和监测是研究地表植被覆盖状况乃至全球生态变化的重要内容,而森林区域面积的调查和统计又是森林资源调查和监测的研究重点。利用遥感卫星或航空飞机拍摄的高空间分辨率遥感图像,研究解决森林面积的估测问题,即对高分辨率遥感图像进行分类,进而提取出森林区域的部分,是本文的研究重点和主要内容。当前高分辨率遥感图像的分类主要是以计算机为主要工具的对遥感图像像素进行数值处理,从而达到分类识别地貌特征的。聚类分析技术以其独特的分类优势,成为了本文的主要技术手段。由于遥感图像在成像过程中受到噪声及其它因素的干扰,使得图像具有较高的模糊性,基于传统聚类思想的图像分割算法虽然能处理此类问题,但具有一定的局限性。本文在对国内外高分辨率遥感图像分类方法大量的调研和分析基础上,针对高分辨率林区遥感图像提出了一个自适应的加权模糊C-均值聚类分割算法进行分类,并利用分形维数特征对分割后的数据中的森林区域部分进行提取。实验结果表明,该方法对高分辨率遥感图像的分割与林区提取取得了较好的效果,不仅提高了遥感数据的分割精度,也基本实现了数据处理的自动化。 【英文摘要】Forest resource investigation and monitoring is the main content of studying vegetation coverage comdition of the earth抯 surface and global ecological change currently. And the investigation and statistics of the forest area, which is considered to be the basic content of forest resource investigation and monitoring, turns into just work of the paper. To study and solve the problem of estimating forest area is also to do the problem of forest classification in high spatial resolution remote sensing image. To classify the high resolution remote sensing image and further to extract the forest part, are the concentration and main content of the paper. The current classification in high resolution of remote sensing image is to process the pixel data mainly by the computer as the tools, so as to achieve the purpose of classifying geomorphic characteristics. Clustering analysis technology is the main research content for its unique classification advantage. With the continuous development of new technology, clustering analysis methods emerged in endlessly from the traditional K-Means algorithm to the Fuzzy C-Means algorithm. Although these methods improved the remote sensing image segmentation to some extent and promoted the application of remote sensing technology, different algorithms had their advantages and disadvantages.