基于直觉模糊熵的红外图像预处理方法

Technique for infrared image preprocessing based on intuitionistic fuzzy entropy

针对红外图像中的脉冲噪声问题以及传统滤波算法的不足,提出了一种自适应直觉模糊滤波算法.该算法利用直觉模糊集理论在处理不确定性问题中的优势,将图像中各个像素点改由一对隶属度函数和非隶属度函数自适应地标定,然后采用修改后的直觉模糊熵定义自适应地确定各像素点的权值以完成滤波.给出了该算法的具体步骤,仿真结果表明:当脉冲噪声密度较大时,采用该算法可有效地滤除噪声;即使当噪声密度很大时,在数学形态学方法的辅助下,该算法也可出色地完成脉冲噪声滤除工作.

In order to overcome the impulse noise in an infrared image and the drawbacks of traditional filter methods,a new adaptive intuitionistic fuzzy filter algorithm is proposed.By use of the advantages of the theory of intuitionistic fuzzy set in dealing with uncertainty problems,each pixel in the image can be adaptively defined by a pair of membership degree and nonmembership degree,then the weight vague of each pixel can be figured out based on the definition of intuitionistic fuzzy entropy which has been modified. The concrete steps of the technique are given.The simulation results show that,on the one hand,the impulse noise can be effectively eliminated by adopting the presented technique if the noise density is not very large;on the other hand,with the help of the knowledge of mathematical morphology,even the noise density is very large,satisfactory effect can also be obtained by using the given technique.