| 李智峰.基于灰度共生矩阵的图像纹理特征地物分类应用[J].地质与勘探,2011,47(3):456-461 |
| 基于灰度共生矩阵的图像纹理特征地物分类应用 |
| Application of GLCM-Based Texture Features to Remote Sensing Image Classification |
| 投稿时间：2010-06-27 修订日期：2010-09-26 |
| DOI： |
| 中文关键词: 遥感影像 纹理 灰度共生矩阵 地物分类 |
| 英文关键词: remote sensing images,texture characteristics,gray level co-occurrence matrix(GLCM),classification of ground objects |
| 基金项目:中国地质调查局地质调查项目“甘肃中东部重点成矿带与西藏昌都等矿集区矿山开发多目标遥感调查与监测冶地质调查项目
(121201916062)资助。 |
|
| 摘要点击次数: 2988 |
| 全文下载次数: 2222 |
| 中文摘要: |
| [摘要]针对传统遥感影像分类方法的分类精度不高,在分析图像的光谱信息的基础上,对基于
灰度共生矩阵的纹理特征在地物分类中的应用进行了研究。本研究利用原始图像进行主成分分析后的
前两个主成分,经过编程运算,提取了基于灰度共生矩阵方法的不同测度的纹理特征,将提取的纹理特
征作为新的波段,与原始波段进行组合,再对组合图像进行监督分类,探索了利用纹理特征进行地物分
类的可行性,并将分类结果与最大似然法分类结果进行定性和定量比较分析。结果表明,综合了纹理特
征和光谱特征的地物分类方法,能够有效地提高地物分类精度,证明了基于纹理特征的遥感影像分类的
有效性。 |
| 英文摘要: |
| Abstract:In order to solve the problem of low-accuracy in the conventional classification of remote sensing image classification, a new method based
on gray level co-occurrence matrix(GLCM) texture features is presented and utilized. After the principal component analysis, the first two principal com鄄
ponents were selected to extract the texture features of different measurements based on gray level co-occurrence matrix. As a new band, the extracted tex鄄
ture features with the original bands were combined by supervised classification methods to classify the images. The classification result was compared with
the maximum likelihood classification qualitatively and quantitatively. The research results show that a combination of texture features with spectral charac鄄
teristics can enhance the accuracy of ground object classification, and prove the feasibility and effectiveness of the GLCM-based classification method
presented in this paper. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
