针对传统的线性和非线性图像滤波算法对含有丰富细节的图像滤波处理的不足,提出一种基于模糊神经网络的图像滤波算法。该算法利用神经网络的自学习功能训练标准样本以确定模糊贴近度阈值,在基本不改变原有图像的灰度信息的前提下,找出图像受到噪声污染的像素点,采用迭代中值滤波算法得到的中值替换噪声点。仿真实验表明该方法有很好的滤波效果,优于传统的图像滤波算法。 Aiming at the shortage of traditional linear and nonlinear image filtering algorithms for image filtering with rich detail, an image filtering algorithm based on fuzzy neural network is proposed. The algorithm uses the self-learning function of neural network to train the standard samples to determine the fuzzy closeness threshold. The basic principle of the algorithm is to find out the pixels that are polluted by the noise without changing the gray information of the original image. The resulting median replaces the noise point. Simulation results show that this method has a good filtering effect, which is better than the traditional image filtering algorithm.