融合光谱和分形纹理特征的遥感影像分类研究
首发时间:2017-11-22
摘要:针对传统方法提取出的纹理特征与光谱特征融合后的分类效果比较差的问题,该文提出利用分形理论中的毯模型提取遥感影像的纹理特征,融合光谱特征,采用BP神经网络算法对遥感影像进行分类。以北京市五环内区域为研究区,使用landsat8 ETM数据源,实现了基于分形纹理特征、光谱特征的BP神经网络算法分类。实验结果表明:通过毯模型提取的纹理特征可以很好地表达表面特征,辅以该纹理信息的BP神经网络的分类精度相比于只用光谱信息进行分类的精度有一定的提高,改善了分类效果。
关键词: 分形 纹理特征 BP神经网络 毯覆盖模型 遥感影像分类
For information in English, please click here
Remote Sensing Image Classification Based on Fusion of Spectral Features and Fractal Texture Features
Abstract:IAccording to the fact that traditional classification method based on spectral feature is poor, in this paper, the blanket model in the fractal theory was used to extract the texture features of the remote sensing image, and then by combining with the spectral features, the BP neural network algorithm was used to classify the remote sensing images. The experimental results show that the texture features extracted by the blanket model can express the surface features well. When the texture features and spectral features are merged, the classification accuracy of BP neural network is improved compared with that of using only spectral information, and the classification effect is obviously improved.
Keywords: Fractal texture features BP neural network Blanket coverage model remote sensing image classification
基金:
引用
No.****
同行评议
共计0人参与
勘误表
融合光谱和分形纹理特征的遥感影像分类研究
评论
全部评论0/1000