摘要
气象卫星提供的卫星云图可以帮助我们追踪云的变化,从而进行天气分析和预报,目前,人工判读仍然是卫星云图识别、分析的主要方式之一,其具有一定程度的主观因素,因此对卫星云图中各种类型的云进行自动分类一直是遥感领域的研究热点与难点。文章基于VC++环境,结合图像的纹理特征和光谱特征采用支持向量机(SVM)实现卫星云图自动分类,实验最后进行分类精度、分类速度及分类效果分析,其研究成果在遥感云图分类、云量检测行业有广泛的应用前景。
Meteorological satellite imagery provided bymeteorological satellite can help us to trackthe change of cloud, thereby performing weather analysis and forecasting. At present,the manual interpretation is still one of the main methods of satellite image identification and analysis,which has a certain degree of subjective factors,so the various types of cloud in the satellite cloud image automatic classification has been the research hotspot and difficulty in the field of remote sensing.In this paper, based on VC ＋ ＋ environment, using support vector machine （SVM） achievesautomatic classification of＂ meteorological satellite imagery combined with the texture featuresand spectral characteristics of image, finally, analysis of classification accuracy, classification speed and classification effect, the research achievement has a broad application prospect in the classification of remote sensing cloud and cloud testing industry.
出处
《无线互联科技》
2016年第1期120-123,共4页
Wireless Internet Technology
基金
中国国家自然基金项目
项目编号:61263012
U1431118
中国博士后基金项目
项目编号:2012M510593
航空科学基金项目
项目编号:20120156001
关键词
卫星云图
纹理特征
光谱特征
支持向量机
自动分类
satellite cloud chart
texture features
spectral characteristics
support vector machine
automatic classification
