基于松弛因子的快速独立分量分析算法的遥感图像分类技术

多光谱遥感图像反映了不同地物的光谱特征,其分类是遥感应用的基础。独立分量分析算法利用了信号的高阶统计信息,对于多光谱遥感图像而言,算法去除了波段图像之间的相关性,获得的波段图像是相互独立的。但是独立分量分析算法有一个缺点,即计算量太大,影响了在多光谱遥感图像分类上的应用。M-FastICA算法同FastICA算法一样,它们的收敛依赖于初始权值的选择。通过在M-FastICA算法中引入松弛因子,使算法可以实现大范围收敛,得到更稳定的收敛效果。应用BP神经网络对独立分量分析算法预处理后的图像进行自动分类,其分类精度较原始遥感图像的精度均高,并且三种独立分量分析算法的分类性能也相当。

Remote Image Classification Based on Loose Modified Fast Independent Component Analysis Algorithm

The multi-spectral remote sensing images reflect the spectral features of diverse surface features,and the classification is the base of remote sensing applications.The ICA algorithm uses the high-order information of signals of multi-spectral remote sensing images,it not only removes the correlation of images,but also obtains the new band images that are mutual independent.But the computational complexity of ICA is too big,and it influences the application of ICA in remote sensing field.M-FastICA algorithm has one flaw same as FastICA,which is the convergence dependent on initial weight.Improved loose gene in M-FastICA algorithm,the new algorithm(LM-FastICA) can implement convergence in large-scale.BP Neural Network is used in classification of the images,which are pre-processed by independent component analysis,the right rate of ICA images is higher than source remote images,and the performance of classification of three kinds of ICA algorithm is near.