基于小波的高分辨率遥感影像纹理分类方法研究

基金项目: 国家自然科学基金资助项目

详细信息

Methods for Classification of the High Spatial Resolution Remotely Sensed Images Based on Wavelet Transform

¹ State Key Laboratory for Information Engineering in Surveying,Mapping and Remote Sensing,Wuhan University,129 Luoyu Road,Wuhan 430079,China

Funds: 国家自然科学基金资助项目

摘要: 在基于小波的纹理分类算法的基础上,提出了逐点特征加权和活动窗口算法,使小波纹理分析能够用于高分辨率遥感影像的分类。逐点特征加权算法用样本的均值和方差构造偏离量,对纹理特征进行自适应加权。实验结果表明,本文提出的算法能够有效地提高分类精度,使地物的内部和边缘的分类效果都得到改善。
- 小波变换 /
- 高分辨率 /
- 逐点特征加权
Abstract: This paper discusses the shortage of conventional algorithms of texture classification based on wavelet transform,presents two improved approaches of point feature weighting and smart windows.The point feature weighting algorithm constructs the deviation vector using means and variances of samples,and uses the deviation vector to restrain the feature units which do harm to the classification accuracy.The subsequent experiment results show that the performance of the presented algorithm is much better than the conventional one.
- wavelet transform /
- high spatial resolution /
- point feature weighting