提出了一种基于DNA计算的高光谱遥感数据光谱匹配分类新方法。该方法利用DNA编码提取各类地物光谱所携带的物理吸收与反射特征信息, 将地物光谱特征转换为DNA编码空间特征, 通过DNA计算基因操作寻找各类地物最典型的DNA信息链。在此基础上, 利用DNA计算原理建立一系列模糊规则, 对高光谱数据进行光谱匹配分类。通过与传统的光谱匹配算法(二值编码, 光谱角, 光谱差分特征编码)的分类结果进行比较, 证明该算法分类精度优于传统高光谱数据的光谱匹配分类方法, 具有实用价值。

Some initial investigations are conducted to employ DNA computing for hyperspectral remote sensing data classification. As a novel branch of computational intelligence, DNA computing expresses rich information of spectral features with DNA encoding, and acquires the most typical DNA encoding of each class by DNA modulating and controlling mechanism. For each pixel of the hyperspectral image, computing the distance between the pixel and the typical DNA sequence, finding the class property of the minimum distance, set the class property of each pixel as the minimum distance class. An experiment was performed to evaluate the performance of the proposed algorithm in comparison with other traditional image matching classification algorithms: binary cording, spectral angles and spectral derivative feature coding (SDFC). It is demonstrated that the proposed algorithm is superior to the three traditional hyperspectral data classification algorithms based on the experiment results.