The Study on the Segmentation of Remote Sensing Cloud Imagery

The researches on weather system are very complicated because the weather system includes many factors, among which the cloud is a very important factor. Cloud appears with various shapes, different gray scales and not clear boundaries in the remote sensing images. For the active contour model and its ex- tensions proposed by Vese and Chan can detect objects whose boundaries are not clear and not necessarily defined by gradient, we employ the active contour model in cloud segmentation in remote sensing cloud imagery. After the contours evolu- tion, the image is automatically decomposed into a set of coherent regions with similar gray scale. The segmentation results are helpful to obtain the distribution of cloud at different levels as a whole, and moreover, the segmentation results could lay a good foundation for further image processing.