基于危险理论的影像纹理调制模板的优化

将危险理论引用到Bayesian Network技术产生纹理调制模板的方法中,解决应用Bayesian Network技术产生调制模板中不能保证纹理模板不断优化的问题。文中介绍危险区、危险信号的概念,并阐述怎样从危险区中找出有效数据的途径。详细介绍在基于Bayesian Network的模板优化中引入危险理论的思路、例证以及计算过程,并通过对实际航空影像纹理分类的试验证明了将危险理论应用于模板优化的有效性和优越性。

The Optimization of Texture Tuned Templates Based on Danger Theory

This paper applies the danger theory in the optimization of texture tuned templates based on Bayesian Network technology, and solves the problem of keeping evolution of texture tuned templates, which are used for texture classification. Some concepts like danger zone and danger signal are introduced, and the approach how to find available data from dangerous zones is also presented. This paper discusses the detailed procedure and algorithms of template optimization based on the danger theory. Experimental results on real aerial images are given to show the validity and advantages of the proposed approach.