基于形态学和支持向量的遥感图像混合像元分解

Unmixing Remote Sensing Imagery Based on Morphology and Support Vector Machines

为了更好地解决混合像元问题,将自动形态学端元提取方法与支持向量机算法相结合进行混合像元自动分解。首先利用自动形态学端元提取方法寻找影像的纯净端元,此方法基于形态学理论,结合像素的光谱信息和空间信息,可以更精确地提取纯净端元。然后通过支持向量算法得到像元组分,支持向量机后验概率作为地物的组分信息。实验结果证明,这种方法具有很高的混合像元分解精度。

A method combined with automated morphological endmember extraction （AMEE） and support vector machines （SVMs） is proposed to solve sub-pixel problem. Based on morphological principle,AMEE method which integrates spatial and spectral information to select endmembers is able to provide a relatively good characterization of general landscape conditions. SVM method can be combin.ed with pairwise cou- pling to output probabilities as the abundance fractions of targets. The experiments show that the method can provide better result of abundance estimation as compared with other methods.