摘要
图像的分数维特征描述了纹理的复杂度和粗糙度。图像的灰度梯度信息则检出了图像中灰度跳变的部分，将图像的梯度信息加进灰度共生矩阵，则使得共生矩阵更能包含图像的纹理基元及其排列的信息。将４ 种基本云类（卷云、积雨云、积云和层云）的分数维和灰度梯度共生矩阵（ＧＧＣＭ）的二次统计特征结合起来，对云类进行分类与识别。样本的试验表明，分数维和灰度梯度共生矩阵的二次统计特征结合起来，气象云图能有效地识别。
The image fractal dimension feature describes the complexity and the roughness of texture.The image gray gradient checks up the skip part of image gray.Combining the gradient information with gray level co occurence matrix,it will be better to include the information of texture elements and its arrangement.Combined the fractal dimension with the 2 order statistical features of Gray Gradient Co occurence Matrix (GGCM) of 4 kinds of basic clouds pictures (Cirrus,Cumulus,Cumulonimbus and Layer Cloud),the clouds are classified and recognized.The samples experiments showed that with the combination of the fractal dimension and the 2 order statistical features of GGCM,the meteorological cloud pictures can be recognized effectively.
出处
《南京理工大学学报》
EI
CAS
CSCD
1999年第4期289-292,共4页
Journal of Nanjing University of Science and Technology
关键词
卫星云图
图像处理
纹理
气象云图
分数维
GGCM
satellite cloud pictures,pattern recognition,image processing
gray gradient co occurence matrix,texture analysis
