基于蚁群算法的多光谱遥感图像分类

Classification of Multispectral Remote Sensing Image Based on ACO

该文应用蚁群算法和支持向量机实现多光谱遥感图像分类。首先提取出多光谱遥感图像的光谱特征、纹理特征和形状特征,然后利用蚁群优化算法从提取出的多维特征空间中选择最优的特征子集向量,最后将特征子集作为支持向量机分类器的输入量实现分类。实验结果显示,较传统的K均值方法文章给出的方法能够提高遥感图像的分类精度。

This paper uses a port vector machine （SVM） classification method based on ant to implement multispectral remote features, textural features and shape features are extracted from colony optimization（ACO） algorithm and sup- sensing image classification. Firstly, spectral the image, and then use ACO algorithm to se- lect subset multi-feature vector from the multi-dimensional feature space. Finally, use SVM classifier to classi- fy the multispectral remote sensing image based the subset features. The experimental results show that the method presented in this paper can obtain higher classification accuracy than K-means.