...
首页> 外文期刊>Sensing and imaging >Spectral–Spatial Hyperspectral Image Classification Based on KNN
【24h】

Spectral–Spatial Hyperspectral Image Classification Based on KNN

机译:基于KNN的光谱空间高光谱图像分类

获取原文
获取原文并翻译 | 示例
           

摘要

Abstract Fusion of spectral and spatial information is an effective way inrnimproving the accuracy of hyperspectral image classification. In this paper, a novelrnspectral–spatial hyperspectral image classification method based on K nearestrnneighbor (KNN) is proposed, which consists of the following steps. First, thernsupport vector machine is adopted to obtain the initial classification probabilityrnmaps which reflect the probability that each hyperspectral pixel belongs to differentrnclasses. Then, the obtained pixel-wise probability maps are refined with the proposedrnKNN filtering algorithm that is based on matching and averaging nonlocalrnneighborhoods. The proposed method does not need sophisticated segmentation andrnoptimization strategies while still being able to make full use of the nonlocalrnprinciple of real images by using KNN, and thus, providing competitive classificationrnwith fast computation. Experiments performed on two real hyperspectral datarnsets show that the classification results obtained by the proposed method arerncomparable to several recently proposed hyperspectral image classificationrnmethods.
机译:摘要光谱和空间信息的融合是提高高光谱图像分类精度的有效途径。本文提出了一种基于K最近邻(KNN)的新颖的光谱-空间高光谱图像分类方法,包括以下步骤。首先,采用支持向量机获得初始分类概率图,该图反映了每个高光谱像素属于不同类别的概率。然后,使用提出的基于匹配和平均非局部邻域的rnKNN滤波算法对获得的逐像素概率图进行细化。所提出的方法不需要复杂的分割和优化策略,但仍然可以通过KNN充分利用真实图像的非局部原理,从而提供具有快速计算能力的竞争性分类。在两个真实的高光谱数据集上进行的实验表明,通过该方法获得的分类结果与最近提出的几种高光谱图像分类方法具有可比性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号