A Method Of Cloud Classification Base On Image Entropy

In this paper a technique to classify cloudy image is introduced base on the image entropy. As beginning, entropy of images are estimated, and then the images with cloud are classified automatically. Secondly 9 thresholds are selected after counting the thresholds of classify. With the 9 thresholds the classification of satellite image is realized. During a heavy rain event on the early July 2007 in Huaihe Basin, the method is applied and results are compared with the enhancement technology of infrared image in GOES manual. As a result, it is shown that the two methods have their own advantages and disadvantages respectively. And the optimal technology is combining the two methods. Finally, relationship of rainfall and FY2C cloud top temperature are concluded during autumn in Huaihe Basin.