摘要
随着数字图像处理技术的发展,多角度、深层次挖掘卫星云图数据信息己成为一种必然发展趋势。根据卫星云图的特点应用了多种图像特征提取算法,包括图像的一、二阶距,分形维数特征和灰度梯度共生矩阵等算法。给出了K—means的改进算法,算法的出发点是确保发现聚类中心的同时使同一类内的相似度大,而不同类之间的相似度小。文中又引入了“事务模式”这一全新概念,基于它对Apriori算法改进后,算法用于云图和雨量数据间的关联规则挖掘,不仅获得了较好的实验结果,也满足了效率需求。
With the rapid development of the digital image processing technology, this makes it as an inevitable trend to dig out the information of satelite cloud images (SCI) and from the deep graduation and multi-angle. Various feature extraction algorithms are presented for satellite cloud images, which are, namely, the first and second order momentum, fractal number and grayscale co-existing matrixes, etc. It proposes an improved algorithm K-means, based on litde comparability between different clusters and large comparability in the same cluster, to set number K and the initial centroids. The problem is greatness of image data.Apriori algorithm, a classical algorithm, don't meet the need of effficiency.In order to deal with the problem, a new concept of transaction is introduced. Base on the concept, Apriori algorithm can be improved so to meet the need of efficiency.
作者
王波静
柳楠
朱琳琳
WANG Bo-jing, LIU Nan, ZHU Lin-lin (1.63061 Troops, People's Liberation Army, Shenyang 110027,China;2.65022 Troops, People's Liberation Army,Shenyang 1101623,China; 3.65042 Troops,People's Liberation Army, Shenyang 110001,China)
出处
《电脑知识与技术》
2009年第7期5242-5244,共3页
Computer Knowledge and Technology
关键词
卫星云图
数据挖掘
图像挖掘
关联规则
聚类
satellite cloud image
data mining
image mining
clustering
association rule