超像素分割已成为遥感影像预处理研究的新热点,但易形成过分割。为解决该问题,提出一种结合超像素和图论的高空间分辨率遥感影像分割方法。首先,采用简单线性迭代聚类(simple linear iterative clustering,SLIC)算法对遥感影像生成超像...超像素分割已成为遥感影像预处理研究的新热点,但易形成过分割。为解决该问题,提出一种结合超像素和图论的高空间分辨率遥感影像分割方法。首先,采用简单线性迭代聚类(simple linear iterative clustering,SLIC)算法对遥感影像生成超像素过分割影像;然后,采用图论算法对超像素进行合并,通过计算得到每次合并后的分割数对应的局部方差,确定合适的影像分割数;最后,根据合适的影像分割数用图论算法对超像素重新聚类合并。实验数据为4幅不同空间分辨率、不同场景的遥感影像,采用定性和定量相结合的方法评价实验结果。实验结果表明,该方法能有效地解决遥感影像过分割结果的问题,获得了良好的分割结果。展开更多
Contourlet is a new 2-D signal analysis tool as well as wavelet analysis.It provides a flexible multi-resolution local and directional image expansion,and is more suitable for image processing than other math tools.In...Contourlet is a new 2-D signal analysis tool as well as wavelet analysis.It provides a flexible multi-resolution local and directional image expansion,and is more suitable for image processing than other math tools.In this paper,an IKONOS image with forests is decomposed by Contourlet transformation,then a method for extracting the boundary of forests from high-resolution remote sensing imagery is proposed associated watershed transformation and texture analysis.The experiment showed that the method could extract the boundary of forests accurately.展开更多
文摘超像素分割已成为遥感影像预处理研究的新热点,但易形成过分割。为解决该问题,提出一种结合超像素和图论的高空间分辨率遥感影像分割方法。首先,采用简单线性迭代聚类(simple linear iterative clustering,SLIC)算法对遥感影像生成超像素过分割影像;然后,采用图论算法对超像素进行合并,通过计算得到每次合并后的分割数对应的局部方差,确定合适的影像分割数;最后,根据合适的影像分割数用图论算法对超像素重新聚类合并。实验数据为4幅不同空间分辨率、不同场景的遥感影像,采用定性和定量相结合的方法评价实验结果。实验结果表明,该方法能有效地解决遥感影像过分割结果的问题,获得了良好的分割结果。
文摘Contourlet is a new 2-D signal analysis tool as well as wavelet analysis.It provides a flexible multi-resolution local and directional image expansion,and is more suitable for image processing than other math tools.In this paper,an IKONOS image with forests is decomposed by Contourlet transformation,then a method for extracting the boundary of forests from high-resolution remote sensing imagery is proposed associated watershed transformation and texture analysis.The experiment showed that the method could extract the boundary of forests accurately.
