The Research of Satellite Cloud Image Recognition Base on Variational Method and Texture Feature Analysis

Abstract

Recently, the development of satellite cloud image processing technology has become very quick; the research aspects concentrate on judge the cloud type and classify the cloud mainly. These image processing methods relate to the subject category like image processing and pattern recognition etc; it has become one of the fields of most quickly development in the research of satellite image processing technology. In satellite cloud image, texture is an very important feature, since satellite cloud image has clear texture structure, the computer texture analysis provide perfect future for study and analyze all kinds of satellite cloud image. Variational method is a new image segmentation method development in recent years, which is adapt to modeling and extract deformable contour of random shape. In satellite cloud image, recognize the target object has great application meaning. In this paper, using computer image texture analysis technology combine with variational theory, extract and analyze the texture feature of familiar cloud and clear sky in satellite cloud image, adopt co-occurrence matrix base on statistic feature, compute the texture parameter like energy, entropy, contrast, correlation and local steady, at last, carry through the automatic recognition experiment which using the active contour model base on variational theory, and acquire better effect of recognition to satellite cloud image.