摘要
BP神经网络具有收敛速度快和自学习、自适应功能强的特点,能最大限度地利用样本集的先验知识,自动提取合理的模型。本文采用Landsat TM遥感图像作为数据源,以山西省定襄县为研究区,通过主成分分析方法来压缩输入数据,并结合NDVI和纹理特征来建立BP神经网络的土地利用分类模型,将分类结果与基于光谱单元信息的神经网络分类和基于纹理特征的神经网络分类结果进行定性和定量比较分析。结果表明:该方法总精度达到了80.50%,分别比基于光谱单元信息的神经网络分类和基于纹理特征的神经网络分类提高了18.89%和6.23%,能够有效地解决地物光谱混淆、分类精度不高等问题。
Remote sensing image classification technology is an important segment of image analysis and interpretation. B PANN has the characteristic of faster convergence speed and stronger capabilities of self-study and self-adaptation, it ＇ s able to utilize the priori knowledge of data sets furthest to automatically determine reasonable models, and model prediction results can be realistic to reflect the real surface features. This paper adopted Landsat TM images for data sources, Dingxiang country of Shanxi province was taken as the research area, principal component analysis was used to compress the data, a comprehensive model of B PANN integrated NDVI with texture features was adopted to classify regional land use information, finally, the result was compared with the classification based on the method of spectrum cell information and Texture Features-based B PANN separately. The result showed that the accuracy of this method is 80. 50%, which increases byl8. 89% compared with spectrum cell information based ANN classification, and 6. 23% compared with texture features based ANN classification separately. It could effectively solve the problems of spectral confusion and low classification accuracy.
出处
《测绘科学》
CSCD
北大核心
2012年第6期140-143,共4页
Science of Surveying and Mapping
基金
中央高校基本科研业务费(QN2009040)
陕西省自然科学基金项目(2011JM5007)
