论文部分内容阅读
传统高光谱图像分类方法主要使用图像的光谱特征信息,没有充分利用高光谱图像的空间特性及样本的其他信息。本文提出了一种基于空间特征与纹理信息的高光谱图像半监督分类方法。首先,将高光谱图像每一像素的光谱特征与其邻域范围内的光谱特征进行结合,得到了这一像素的空-谱特征;然后用灰度共生矩阵提取了高光谱图像的纹理特征,并与空-谱特征进行了融合;最后,用基于图的半监督分类算法进行了分类。通过在Indian Pines数据集和Pavia U数据集上进行试验,结果表明本文提出的方法能取得较高的分类结果。
The traditional hyperspectral image classification method mainly uses the spectral characteristic information of the image, and does not make full use of the spatial characteristics of the hyperspectral image and other information of the sample. This paper presents a semi-supervised classification method of hyperspectral images based on spatial features and texture information. Firstly, the spatial-spectral features of each pixel in the hyperspectral image are combined with the spectral features in the neighborhood of the hyperspectral image. The texture features of the hyperspectral image are then extracted using the gray-level co-occurrence matrix And air-spectral features were fused; Finally, the use of graph-based semi-supervised classification algorithm for classification. Experiments on Indian Pines data set and Pavia U data set show that the proposed method can achieve higher classification results.