Image Filtering Based on Improved Information Entropy

• 作者: JINGXiaojun LIUYulin XIONGYuqing
• 来源期刊: 电子学报(英文版)
• 学科: 工学
• 关键词: 图象滤波 熔融熵标准 混合滤波 皮尔逊准则
• 年，卷（期）: 2004,（3）
• 页码范围: 419-423
• 页数: 5页
• 分类号: TP391.41
• 语种: 中文

摘要

An image filtering based on improved information entropy is proposed in this paper, which can overcome the shortcomings of hybrid linear and non-linear filtering algorithm. Due to the shortcomings of information entropy in the field of data fusion, we introduce the consistency constraint factor of sub-source report and subsource performance difference parameter, propose the concept of fusion entropy, utilize its amendment and regularity function on sub-source decision-making matrix, bring into play the competency, redundency and complementarity of information fusion, suppress and delete fault and invalid information, strengthen and preserve correct and useful information, overcome the risk of error reporting on single source critical point and the shortcomings of reliability and error tolerating, add the decision-making criteria of multiple sub-source fusion, finally improve filtering quality. Subsequent experiments show its validity and improved filtering performance, thus providing a new way of image filtering technique.