Cloud Classification by Combining Fractal Geometry Features with Support Vector Machine
We have designed an efficient algorithm to classify cumulus, stratus and cirrus in cloud image. Contrast of a cloud image is perhaps bad or there is some undesired noise in the cloud image, which will result in a bad classification results for different clouds. In order to solve this problem, based on our prior work, the cloud image is transformed to discrete stationary wavelet (DSWT) domain, then detail is enhanced and noise is suppressed in discrete stationary wavelet domain. Based on our prior work, processed cloud image is segmented by continuous wavelet transform and Bezier histogram and a binary image is obtained. Then the binary image is processed by the continuous wavelet transform to extract its edge. Box dimension of fractal geometry is used to extract the features of cloud images. The features are input to support vector machine in order to efficiently classify the cumulus, stratus and cirrus. The proposed algorithm can accurately classify the cumulus, stratus and cirrus. This can provide a good method for weather prediction.
cloud classification fractal geometry support vector machine
Changjiang Zhang Jianping Xu Juan Lu Xiang Zhang
College of Mathematics, Physics and Information Engineering, Zhejiang Normal University, Jinhua,Zhejiang, China
国际会议
第八届国际测试技术研讨会(8th International Symposium on Test and Measurement)
重庆
英文
824-827
2009-08-01（万方平台首次上网日期，不代表论文的发表时间）