Typhoon cloud system segmentation with multichannel images using vector-valued Chan-Vese model

Kun Wei, Yuanxiang Li, Zhongliang Jing, Chunxiang ShiProceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE 2009 · DOI

Abstract

Cloud segmentation is an important step and a very difficult problem in typhoon image processing. There are many works on cloud image segmentation, but few are carried out on typhoon primary cloud system (galaxy) segmentation. Typhoon satellite images are always multiple channels whose properties are very different, so that the appearances of these channels are different as well. In order to segment out primary cloud systems accurately, multiple channel images are employed in this paper. The image data is from MERSI (short for MEdium Resolution Spectral Imager) of Chinese FY- 3A meteorological satellite launched on March, 2008. The scalar multiphase Chan-Vese (CV) model is extended for the vector-valued images, so as to partition out typhoon cloud systems. The experiment results show that the multi-channel segmentation is more accurate, more complete and more effective than that of usually using only one image, with multiple channel images being treated as a vector one input into the CV model. The multi-channel segmentation integrates the distribution information of cloud systems in all channels, so information fusion of multiple channels are realized when segmenting.
BibTeX
@inproceedings{wei2009typhoon,
  author = {Kun Wei and Yuanxiang Li and Zhongliang Jing and Chunxiang Shi},
  title = {{Typhoon cloud system segmentation with multichannel images using vector-valued Chan-Vese model}},
  booktitle = {Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE},
  year = {2009},
  doi = {10.1117/12.833129},
}

Loading related content...

Figure