摘要
变形体的变形通常是多个变形子模式综合作用的结果,因此变形子模式的识别和分离是变形数据分析处理中的一个重要问题。基于独立分量分析原理,提出了一种新的变形数据分析处理方法;该方法应用独立分量分析中的最大负熵法,将变形观测数据作为观测信号,对变形观测信息进行独立分量分析,可以获得彼此独立的变形子模式。实验结果表明独立分量分析方法可以从变形观测数据中有效地识别和分离出相互独立的变形子模式。
Deformation of distortion-body is usually the combining result of several deformation sub-patterns, so deformation sub-pattern recognition and separation is an important problem in deformation data analysis and processing. Based on independent component analysis principle, this paper presents a novel deformation data analysis and processing approach, in which the maximum negentropy is used as the criterion of independent component separation, and the deformation data are used as observation signals, then deformation sub-patterns can be separated in the way of independent component analysis. The experiment shows that independent component analysis can be used to find out the independent deformation sub-patterns from deformation data, and this approach promotes the development of the theory and the method of deformation data analysis and processing.
出处
《金属矿山》
CAS
北大核心
2005年第4期46-49,共4页
Metal Mine
基金
山东省基础地理信息与数字化技术重点实验室开放基金资助项目(SD2003-10)。
关键词
独立分量分析
变形子模式
最大负熵法
模式识别
Independent component analysis,Deformation data analysis and processing, Deformation sub-pattern, Negentropy, Pattern recognition
