首页> 外文会议>IEEE Conference on Industrial Electronics and Applications >A new pansharpening method using multistage joint bilateral filtering
【24h】

A new pansharpening method using multistage joint bilateral filtering

机译:一种采用多级联合双边滤波的泛锐化新方法

获取原文

摘要

Due to the inaccuracy of detail extraction, most of pansharpening methods often suffer from spectral and/or spatial information distortions more or less and produce unsatisfied pansharpened results. To deal with this problem, we present a new pansharpening method using multistage joint bilateral filtering. Our method first separates the multispectral data into intensity and angular components by hyperspherical color transform (HCT). Then the proposed multistage joint bilateral filter is used to capture the different-scale details of both the panchromatic and the intensity images. Specifically, multistage joint bilateral filter performs joint bilateral filtering iteratively while the reference image remains unchanged and the filtering parameters vary progressively. The structure information with different scale is aware and the corresponding spatial details can be effectively extracted. The total spatial details are then combined and injected into the intensity component of the MS imagery. Finally, all the fused MS bands are obtained by the inverse HCT. The experiment is carried out on IKONOS and GeoEye-1 satellite images, respectively. Visual and objective analysis show that our method can produce high-quality pansharpened results and outperform some existing methods.
机译:由于细节提取的不准确性,大多数泛锐化方法通常或多或少地遭受光谱和/或空间信息失真的困扰,并产生不满意的泛锐化结果。为了解决这个问题,我们提出了一种使用多级联合双边滤波的泛锐化方法。我们的方法首先通过超球面色彩变换(HCT)将多光谱数据分为强度和角度分量。然后,将提出的多级联合双边滤波器用于捕获全色图像和强度图像的不同比例的细节。具体来说,多级联合双边过滤器迭代执行联合双边过滤,而参考图像保持不变,并且过滤参数逐渐变化。知道不同尺度的结构信息,并可以有效地提取相应的空间细节。然后将全部空间细节合并并注入到MS图像的强度分量中。最后,通过逆HCT获得所有融合的MS频带。该实验分别在IKONOS和GeoEye-1卫星图像上进行。视觉和客观分析表明,我们的方法可以产生高质量的锐化结果,并且优于某些现有方法。

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