CAJ | 학술논문

高概率椒盐噪声对数字图像的重度污染大量存在，如要消除信息少且噪点集中的噪声存在诸多困难；而低概率椒盐噪声对数字图像的轻度污染虽然可完全滤除，但在实际图像恢复中又缺少普遍意义。本文基于灰度值空间的模糊划分和描述灰度水平的模糊数，采用极值法对高概率噪声实施检测并建立恰当滤波窗口，应用广义重心去模糊化法和非噪声点对应的隶属函数设计一种新模糊滤波器。最后，通过仿真实例获得该滤波器可有效地过滤数字图像中高概率椒盐噪声，并说明它的去噪性能优于其他常见滤波器。
고개솔초염조성대수자도상적중도오염대량존재，여요소제신식소차조점집중적조성존재제다곤난；이저개솔초염조성대수자도상적경도오염수연가완전려제，단재실제도상회복중우결소보편의의。본문기우회도치공간적모호화분화묘술회도수평적모호수，채용겁치법대고개솔조성실시검측병건립흡당려파창구，응용엄의중심거모호화법화비조성점대응적대속함수설계일충신모호려파기。최후，통과방진실례획득해려파기가유효지과려수자도상중고개솔초염조성，병설명타적거조성능우우기타상견려파기。
Severe pollution of salt-and-pepper noise with high probability in digital images is widespread ,such as to eliminate less information and noisy noise has many difficulties .Although light pollution of salt-and-pepper noise with low probability in digi-tal images can be completely removed ,it lacks of universal sense in actual image restoration .In this paper ,aimed at the fuzzy parti-tion in the gray-value spaces and fuzzy numbers describing the gray-value level ,the extreme value method is used to conduct the noise with the high probability test and establish the suitable filtering window .And then ,applying the defuzzification method of cen-ter of generalized gravity and the membership functions corresponding to the non-noise points to design a kind of new filter .Finally , we obtain that the filter can effectively filter the salt and pepper noise with high probability in digital images by simulation example , and its denoising performance is superior to other common filters .