两种改进型中值滤波算法比较

重点介绍了综合型中值滤波算法以及一种自适应中值滤波算法．针对这两种改进型中值滤波算法,对含有高斯噪声,椒盐噪声,混合噪声以及高密度噪声图像进行去噪处理,比较综合型中值滤波算法（文中采用了十字型和交叉型）和自适应中值滤波算法对不同图像的去噪效果．     

基于量子分解的航拍图像中值滤波

针对无人机航拍图像在采集或者传输过程中容易产生噪声的问题,提出了一种基于量子图像分解的中值滤波算法。该算法首先将经典图像表示为量子图像,并通过量子理论对量子图像进行分解,得到分解子图,然后利用改进的快速中值滤波算法对分解子图进行去噪,最后将各子图去噪结果进行合成得到最终的去噪图像。实验证明,该方法能够有效去除无人机航拍图像中的椒盐噪声和高斯噪声,所提算法在去噪效果方面与传统中值滤波算法相比,信噪比提高了约17%,与递归中值滤波相比提高了约28%。并且算法的效率得到了一定程度的提高。     

基于不同窗口形状下的CT图像中值滤波算法

CT作为医学影像工具,对CT图像的去噪声处理是医学图像预处理中的重要环节,中值滤波是传统滤波算法中对滤除椒盐噪声有较好滤波性能的方法之一,滤波窗口形状有着多样性。文中对于不同形状窗口进行了程序编写,研究了不同形状的滤波窗口、不同窗口大小的中值滤波算法对CT图像的去噪效果,这些影响到中值滤波中心像素值的选取以及滤波处理的计算量。实验结果表明,中值滤波的效果有差别,为改进中值滤波算法及对CT图像的去噪声处理提供了参考,同时也有利于研究中值滤波与其他滤波方法结合进行去噪声处理,可以更好地保留图像的细节。     

基于极值检测的图像滤波算法

针对极值中值滤波法在去除椒盐噪声时存在的不足,提出了一种改进的图像滤波算法.首先使用极值法检测图像中的噪声点,然后采用窗口由小到大变化的自适应算法得到噪声像素的滤波值,最后通过计算噪声像素滤波前后灰度值的差值来修正被误判像素的灰度值.对不同类型、受不同强度噪声污染图像的去噪实验表明,该方法在不同噪声率下均优于传统的中值滤波法及其一些改进算法,当噪声率较大时其去噪和保边性能得到了显著提高.     

改进自适应中值滤波的图像去噪

传统自适应中值滤波的最大最小窗口尺寸固定,并且其最大最小窗口相差较大时,运算时间较长,去噪效果并小一定最佳。本文针对传统自适应中值滤波算法的不足,提出了改进自适应中值滤波算法,首先根据椒盐噪声的分布特点,从单幅含椒盐噪声图像中估算出椒盐噪声的浓度,并分析噪声浓度与自适应中值滤波窗口尺寸之间的关系,建立它们之间的函数关系一其次根据噪声浓度确定自适应中值滤波的最大最小窗口尺寸,最后对图像进行自适应中值滤波：实验结果表明本文算法运算时间随着噪声浓度的变化而变化,而且从PSNR角度来看本文去噪效果比传统自适应中值滤波效果较好。     

改进自适应中值滤波算法研究

重点介绍了自适应中值滤波算法以及两种改进的自适应中值滤波算法.针对这3种滤波算法,对含有不同密度椒盐噪声的图像进行去噪实验.结果表明:改进算法去噪效果明显、能有效保护图像细节,PSNR保持在25 dB以上;改进算法在高密度噪声时也能得到细节较为清晰的图像,PSNR比改进前的提高17 dB以上.     

基于三维轴距的图像去噪算法

图像去噪是图像处理邻域的重要课题,该文在3维直方图的基础上定义图像的3维轴距,分析无噪声及椒盐噪声影响下图像3维轴距的分布特点,提出基于3维轴距的去噪算法。算法首先对图像进行边沿扩展,再利用3维轴距来检测噪声,最后用非噪声中值滤波来消除噪声。仿真比较了3维轴距去噪算法与中值滤波、自适应开关中值、快速自适应均值滤波、修正方向加权中值滤波和自适应模糊开关中值滤波算法的去噪性能,结果证明该文算法的有效性。     

一种改进的加权均值滤波算法

当图像中同时存在脉冲噪声和高斯噪声时,传统的中值滤波算法和均值滤波算法均不能达到较好的去噪效果。针对这一问题,提出了一种改进的加权均值滤波算法。算法采用局部阈值优化的方法计算各像素点的权值,将滤波窗口各像素点的灰度值与对应的权值进行加权运算,结果作为窗口中心点的滤波输出。仿真实验结果证明,该算法对脉冲噪声和高斯噪声具有较强的去噪能力,且较好地保持了图像的细节,效果均优于传统中值、均值滤波算法和改进的中值滤波算法(IMF)。     

基于噪声检测的自适应中值滤波算法

针对现有中值滤波算法对于高密度噪声图像以及纹理细腻图像的边缘处理能力欠佳的缺陷,提出一种基于噪声检测的自适应中值滤波算法.新算法根据噪声点与周围信息的关联程度将噪声点滤波值进行调整,从而更好的处理图像的细节部份.新算法中的自适应策略加强了滤波算法的去噪性能,使其对于含有任意噪声密度的图像也能很好的进行噪声滤除.通过仿真分析,新算法对于细节丰富的图像以及高密度噪声的图像滤波效果良好,有效的提高图像的峰值信噪比,其去噪效果相比其他方法更加优秀.     

基于Matlab的椒盐噪声滤波算法研究与实现

介绍了图像去噪流程,研究了图像椒盐噪声处理中的两种算法,均值滤波算法和中值滤波算法,详细阐述了两种算法的基本原理和实现方法,在Matlab环境下利用两种算法对图像进行去噪处理,并对去噪结果进行比较、分析,实验结果表明两种算法都能有效滤除图像中的椒盐噪声,中值滤波算法在保护图像细节方面要优于均值滤波算法。     

波导带通滤波器的精确设计和快速优化

介绍了采用三维微波电磁场仿真软件快速提取直接耦合波导带通滤波器耦合窗口参数的方法,并提出了优化最终模型的快速扫描技术.全波仿真及实验结果表明,该方法能满足波导带通滤波器快速精确设计的要求.     

Inverse Ultraspherical Filters

The inverse ultraspherical filter is derived and its properties analyzed. It is shown that the inverse ultraspherical filter has smaller transition band than the inverse Chebyshev filter under certain circumstances while still maintaining the maximally flat passband characteristic. Filter pole and zero calculations are described and typical magnitude and delay responses generated. Nomographs of inverse ultraspherical filters are also provided for determining filter order and for possible magnitude response optimization.Celestino A. Corral was born in Havana, Cuba. He earned his Bachelors, Masters and Ph.D. degrees in electrical engineering from the University of Miami, Florida, in 1985, 1987, and 1993, respectively. He completed his Ph.D. program under a Patricia Roberts Harris fellowship. From 1988 to 1990 Dr. Corral was a Member of the Technical Staff at Sandia National Laboratories, Albuquerque, New Mexico, where he was involved in the analysis, modeling, and design of custom circuits for piezoelectric sensor components. In 1994, he joined Scientific-Atlanta in Atlanta, Georgia, as an Associate Staff RF Engineer and worked on communication systems and specialized microwave software. In 1997 he was Senior Design Engineer at Origin Data Systems in Boca Raton, Florida, where he had multiple roles in system, hardware, and software development for a novel asset management/RFID system. Currently, Dr. Corral is a Distinguished Member of the Technical Staff in Motorola Labs, Plantation, Florida, and is involved in ultra-wideband, wideband, and propagation research. His interests include circuit and filter theory, communication systems, numerical analysis, signal processing, and constrained optimization. He has published over twenty-five journal and conference papers, has six patents filed, and two trade secrets.Dr. Corral was principal recipient of the Myril B. Reed Best Paper Award for the 40th Midwest Symposium on Circuits and Systems in 1997. Dr. Corral is also the principal recipient of the Ambrose Fleming Premium for IEE Proceedings on Circuits, Devices, and Systems, 2000. Dr. Corral is a Senior Member of IEEE and a member of Omicron Delta Kappa, Tau Beta Pi, and Eta Kappa Nu honor societies.     

Design of two-dimensional directional filters

A directional filter has a passband extending fully along a straight line passing through the origin. In its original form the transformation technique, a powerful tool for designing two-dimensional FIR filters, is not useful to design such a filter. In this communication we suggest a few analytical and optimisation-based methods to design directional filters using the transformation technique. The superiority of the proposed methods over existing techniques is demonstrated.     

一种基于CIC滤波器的有效锐化方法研究

介绍了对积分梳状滤波器（CIC滤波器）的有效锐化。所提出的锐化滤波器的结构由两个主要部分组成：一个梳状滤波器的级联部分和一个锐化滤波器部分。所提出的方案使得滤波器中锐化部分的工作速率比输入速率大为降低，其频谱响应特性比传统的也有所改进。通过MATLAB仿真，可看出改进锐化后的滤波器性能更优。     

Current-mode continuous-time filters using complementary current mirror pairs

A design of current-mode continuous-time filters for low voltage and high frequency applications using complementary bipolar current mirror pairs is presented. The proposed current-mode filters consist of simple bipolar current mirrors and capacitors and are quite suitable for monolithic integration. Since the filters are based on the integrator type of realization, the proposed method can be used for a wide range of applications. The frequency of the filters can easily be changed by the DC controlling current. A fifth-order Butterworth and a third-order leapfrog filter with tunable cutoff frequencies from 20 MHz to 100 MHz are designed as examples and simulated by SPICE using standard bipolar parameters.     

DDC中的抽取滤波器设计及FPGA实现

本文介绍了在数字下变频（DDC）中的抽取滤波器系统设计方法和具体实现方案。采用CIC滤波器、HB滤波器、FIR滤波器三级级联的方式来降低采样率。通过实际验证，证明了设计的可行性。     

Miniaturized dielectric waveguide filters

Design techniques for a new class of integrated monolithic high-permittivity ceramic waveguide filters are presented. These filters enable a size reduction of 50% compared to air-filled transverse electromagnetic filters with the same unloaded Q-factor. Designs for Chebyshev and asymmetric generalised Chebyshev filter and a diplexer are presented with experimental results for an 1800 MHz Chebyshev filter and a 1700 MHz generalised Chebyshev filter showing excellent agreement with theory.     

大功率谐波吸收式滤波器设计

介绍了大功率谐波吸收式滤波器的实现电路，这些电路满足了信号分流的技术要求。文中给出了详细的工作原理和测试数据。     

一种数字下变频抽取滤波器的设计

对一个针对数字下变频应用的抽取滤波器从设计指标到版图实现的设计过程进行了详细介绍.该抽取滤波器实现了20倍的降采样,由CIC滤波器、CIC补偿滤波器和半带滤波器三级依次串联而成.通过利用抽取滤波器的等价变换和多项分解性质,各滤波器级的硬件电路开销和运行功耗都得到了降低.     

Single-stage line-coupled ac/dc converter with high power factor and ripple-free input current

Single-stage line-coupled ac/dc converter with high power factor and ripple-free input current is proposed. The proposed power factor correction circuit can achieve high power factor and ripple-free input current using a coupled inductor. Experimental results for a 400?W converter at a constant switching frequency of 100?kHz are obtained to show the performance of the proposed converter.