摘要图像滤波是图像预处理过程中不可缺少的部分，其结果的好坏直接影响到后续的图像处理和分析的有效性和可靠性，其中利用脉冲耦合神经网络的图像滤波具有广泛的应用前景。脉冲耦合神经网络(Pulse-Coupled Neural Network,简称PCNN)是基于对猫的脑视觉皮层试验中所观察到的神经元脉冲串同步振荡现象，研究发展而来的新式神经网络。34565
本文介绍了基于PCNN模型的滤波算法。PCNN的神经元同步点火性质，使的它能够由各神经元的不同点火方式来查找噪声点的具体位置，再针对噪声采取相应的算法来去除，这样不但能够在去噪的同时可以尽量的保持图像细节，从而获得较好的图像质量。
由于PCNN网络参数只能通过手动调节，最优参数难以确定的缺陷，本文在PCNN简化模型的基础上作出了改进。针对不同点的像素和相邻像素之间的耦合关系的不同，改进内部活动项的连接系数。通过实验得出，改进后的模型处理效果要优于改进前的效果。
毕业论文关键词: 脉冲耦合神经网络 图像滤波 连接系数
Abstract
Image filter is an integral part of image processing, the result has a direct impact on the subsequent image processing and analysis, validity and reliability, using pulse coupled neural networks for image filtering has broad application prospects. Pulse-coupled neural networks (Pulse-Coupled Neural Network, called PCNN) is based on the brain's visual cortex of cats in this experiment observed pulses with synchronous oscillation of a neuron, research and development of new type neural network.
This article describes filtering algorithm based on PCNN model.PCNN neurons are synchronized ignition properties,it can by the neurons of different ignition ways to find where the noise in an image,and then take the appropriate method to remove noise,This way, it can denoising while keeping image details to get a better image .
Due to PCNN network parameters can only be manually adjusted, it is difficult to determine the optimum parameters.This article is based on a simplified PCNN model made improvements.Solving the problem that between different pixel and the adjacent pixels have different coupling relationships,optimized the connection coefficient of internal activities.Tests show that,Improved model for image processing capabilities is better than no improvement model.
Key words PCNN Image filtering Connection coefficients
目 录
摘 要 I
Abstract II
1 绪 论 1
1.1 研究背景及意义 1
1.2 脉冲耦合神经网络的国内外研究现状 2
1.3 本文章节安排 3
2 脉冲耦合神经网络的基本原理 4
2.1 脉冲耦合神经元基本模型 4
2.2 简化PCNN模型结构 6
2.3 脉冲耦合神经网络 9
2.3.1 PCNN运行机制 9
2.3.2 PCNN图像处理原理 10
2.4 PCNN特性分析 11
2.5 本章小结 13
3 基于PCNN模型的图像滤波方法 14
3.1 图像滤波 14
3.2 基于PCNN的滤波算法 16
3.2.1 基本思想 16
3.2.2 算法介绍 17
3.2.3 仿真测试 19
3.3 PCNN模型改进 22
3.3.1 改进思想 22
3.3.2 改进方法 23
3.3.3 模型行为分析 25
3.3.4 仿真测试 27 脉冲耦合神经网络的图像滤波+源程序:http://www.751com.cn/zidonghua/lunwen_32095.html