基于模糊理论的红外图像像元串音抑制算法  

A Restraining Crosstalk Method of Infrared Image Based on Fuzzy Theory

作　　者：唐良瑞[1] 张悦[1] 李守荣[1] 

机构地区：[1]华北电力大学信息工程系,北京102206

出　　处：《红外技术》2006年第5期283-286,共4页Infrared Technology

基　　金：国家自然科学基金(编号:60402004)

摘　　要：红外目标模拟器产生的红外图像都不同程度地存在串音干扰问题。针对该问题提出一种基于模糊理论和模板的图像对比度增强算法。首先确定12种模板,分别计算它们的隶属度,再根据最大隶属原则确定像素点所隶属的模糊集合,最后利用该模板对像素点进行对比度增强的处理。由于背景噪声的存在也会影响图像质量,因此利用阈值去除对比度增强后的图像的背景噪声。实验结果表明,利用该算法不但使图像的对比度有很大的增强,而且抑制了串音,表明该算法有效。There are more or less crosstalk in Infrared Images used widely at present. To solve the problem, we devise a contrast enhancement arithmetic of infrared image based on fuzzy theory and template, assuring 12 kinds of them and calculating their membership. According to maxim membership principle, we can confirm the centre pixel belongs to which template and then has it enhanced by template which belongs to. Because background noise can affect image quality, so we decide to make use of threshold to denoise enhanced image. Experimental result demonstrates that not only the new arithmetic does enhance image obviously, but also restrain crosstalk. Therefore, the arithmetic is effective.

关 键 词：模糊理论 最大隶属度 串音 模板 背景噪声 

分 类 号：TN216[电子电信—物理电子学]