基于克隆选择的多光谱遥感影像分类算法

Classification of Multi spectral Remote Sensing Image Based on Clonal Selection

为了对多光谱遥感影像进行更精确的分类,提出了一种基于克隆选择(clonalselection)的多光谱遥感影像分类算法。该方法首先应用基于人工免疫系统的克隆选择算法对样本进行自学习来得到全局最优的聚类中心,然后利用学习得到的聚类中心对整幅影像进行分类。由于克隆选择算法具有生物免疫系统自组织、自学习、自识别、自记忆的能力,不仅使得基于克隆选择的多光谱遥感影像分类算法具有非线性的分类能力,而且能够快速准确地得到全局最优解,从而克服了传统分类方法约束条件多、容易陷入局部最优的缺点。实验结果证明,基于克隆选择的多光谱遥感影像分类算法在分类精度上优于传统的分类方法,其总精度和Kappa系数分别达到了93. 63%和0. 915,因而具有实用价值。

In this paper, some initial investigations are conducted to employ Clonal selection for classification of multi spectral remote sensing image. The clonal selection is used to explain the basic features of an adaptive immune response to an antigenic stimulus. The general algorithm, named clonal selection algorithm(CLONALG), is derived from clonal selection to perform machine learning and pattern recognition tasks and it has been adopted to solve optimization problems. In this paper, image classification task by CLONALG is attempted and the preliminary results are provided. The experiment is consisted of two steps: Firstly, the classification task employs the property of clonal selection of immune system. The clonal selection proposes a description of the way that the immune systems copes with the pathogens to mount an adaptive immune response. Secondly, classification results are evaluated by applying three known algorithm: parallelepiped, minimum distance and maximum likelihood. It is demonstrated that our method is superior to the three traditional algorithms, and its overall accuracy and Kappa coefficient reach 93 63% and 0 915 respectively.