摘要
纹理是遥感影像处理和计算机视觉领域中的重要研究内容,在植被遥感中有着非常广泛的应用前景。主要讨论在面向对象,多尺度分割的基础上,分别基于光谱信息和纹理特征的最邻近分类法提取南京某新开发区的植被信息。研究结果表明,与传统的基于光谱信息和基于象元特征的分类法相比,辅以纹理特征的最邻近分类法的总体精度分别提高了6.36%和1.16%。
Texture is an important research content in remote sensing image processing and computer vision field,and is has a very wide range of applications in the field of vegetation remote sensing.Basis on the the objectoriented and the multiresolution segmentation,discusses,respectively,nearest neighbor classifications which are based on spectral information and texture features extracted vegetation of a new development zone in Nanjing.The research results show that,compared with the classification based on spectral information and the features pixedbased classification,the overall accuracy of object-oriented nearest neighbor classification supplemented by texture feature is increased by 6.36％ and 1.16％.
出处
《科学技术与工程》
北大核心
2013年第29期8672-8677,8696,共7页
Science Technology and Engineering
基金
华南理工大学亚热带建筑科学国家重点实验室开放研究项目(2011KB11)
国家自然科学基金项目(40701103)资助
关键词
灰度共生矩阵
面向对象
多尺度分割
纹理分析
最邻近分类
GLCM
object-oriented
multiresolution segmentation
texture analysis
nearest neighbor classification
