光电技术应用, 2014, 29 (3): 34, 网络出版: 2014-07-18
基于光谱与纹理特征的高空间分辨率图像分类算法
Classification Algorithm of High Spatial Resolution Imagery Based on Spectrum and Texture Features
摘要
针对遥感高分辨率光谱图像的特点, 提出了一种将纹理信息与光谱信息相结合的分类算法。对传统的局部二值模式纹理提取方法(LBPV)进行改进, 并应用到高分辨率图像的土地覆盖分类中。结果表明, 加入LBPV纹理特征的分类算法具有很好的空间连续性以及较高的分类精度。
Abstract
According to the characteristics of high resolution remote sensing spectral images, a classification algorithm with texture and spectral information is described. Traditional local binary patterns variance (LBPV) is improved to apply in land-cover classification of high resolution images. The results show that LBPV has better spatial continuity and higher classification accuracy.
黄立贤, 沈志学, 骆永全, 张大勇. 基于光谱与纹理特征的高空间分辨率图像分类算法[J]. 光电技术应用, 2014, 29(3): 34. HUANG Li-xian, SHEN Zhi-xue, LUO Yong-quan, ZHANG Da-yong. Classification Algorithm of High Spatial Resolution Imagery Based on Spectrum and Texture Features[J]. Electro-Optic Technology Application, 2014, 29(3): 34.