融合噪声信息函数的改进FCM图像分割方法研究

Research on improved FCM image segmentation method based on fusion noise information function

针对模糊C均值图像分割方法存在的噪声干扰问题,提出一种融合空间噪声信息的改进模糊C均值图像抗噪分割方法。该方法通过定义空间噪声信息函数并将其引入隶属度函数中,若图像中出现噪声像素点,则其空间噪声信息函数值增大,相应隶属度函数和目标函数值也随之增大,当目标函数值超出设定阈值时,则剔除该噪声点。实验结果表明,提出的方法可以有效解决利用FCM算法进行图像分割时存在的噪声干扰问题,实现图像的精准分割。

Aiming at the noise interference problem of fuzzy C-means image segmentation method,an improved fuzzy C-means image anti noise segmentation method based on spatial noise information is proposed.In this method,the spatial noise information function is defined and introduced into the membership function.If there are noise pixels in the image,the value of the spatial noise information function increases,and the value of the corresponding membership function and objective function also increases.When the value of the objective function exceeds the set threshold,the noise point is removed.Experimental results show that the proposed method can effectively solve the problem of noise interference when using FCM algorithm for image segmentation,and achieve accurate image segmentation.