基于DMN的高光谱图像分割方法研究

曹建农; 关泽群; 李德仁

doi:10.11834/jrs.20050587

浏览量 : 0 下载量: 1230 CSCD: 4

PDF
导出
分享
收藏
专辑

基于DMN的高光谱图像分割方法研究
Study on the Approaches for Segmentation of Hyper-spectral Images Based on DMN
2005年第5期页码：596-603
纸质出版：2005
DOI： 10.11834/jrs.20050587
稿件说明：

移动端阅览

[1]曹建农,关泽群,李德仁.基于DMN的高光谱图像分割方法研究[J].遥感学报,2005(05):596-603. DOI： 10.11834/jrs.20050587.

CAO Jian-nong~. Study on the Approaches for Segmentation of Hyper-spectral Images Based on DMN[J]. Journal of Remote Sensing, 2005, (5): 596-603. DOI： 10.11834/jrs.20050587.

摘要

简述高光谱遥感光谱特征体系（包括光谱曲线特征、光谱变换特征和光谱度量特征3个层面）。研究马尔科夫网的概念和方法

生成基于光谱角（SA）特征度量的DMN

并以DMN为证据对高光谱图像进行分割;研究和实验表明基于SA信息的马尔科夫网可以很好地综合高光谱数据空间特征与光谱特征间的关系

为进一步数据处理提供优化控制（其实质是概率神经网络）。最后提出未来应用和研究方向。

Abstract

Three parts of property systems in hyper-spectral remote sensing data are simply expounded. The paper studies concept and method of DMN

constructs it based-on SA

and conducts segmentation of image in line with it. The researches and experiments indicate that the DMN based-on SA information may synthesize relation between spatial and spectral features existed in hyper-spectral remote sensing data; and provides control for further treatment of image such as artificial neural network. Lastly

we point out future directions of its application and research. (