Neural filtering of colored noise based on Kalman filter structure

In this paper, adaptive filtering approaches of colored noise based on the Kalman filter structure using neural networks are proposed, which need not extend the dimensions of the filter. The colored measurement noise is first modeled from a Gaussian white noise through a shaping filter. The Kalman filtering model of colored noise is then built by adopting an equivalent observation equation, which can avoid the dimension extension and complicated computations. An observation correlation-based algorithm is suggested to estimate the variance of the measurement noise by use of a single layer neural network. The Kalman gain can be obtained when a perfect knowledge of the plant model and noise variances is given. However, in some cases, the difficulties of the correlative method and the Kalman filter equations are the amount of computations and memory requirements. A neural estimator based on the Kalman filter structure is also analyzed as an alternative in this paper. The Kalman gain is replaced by a feedforward neural network whose weight adjustment permits minimization of the estimation error. The estimator has the capability of estimating the states of the plant in a stochastic environment without knowledge of noise statistics. If the noise of the plant is white and Gaussian and its statistics are well known, the neural estimator and the Kalman filter produce equally good results. The neural filtering approaches of colored noise based on the Kalman filter structure are applied to restore the cephalometric images of stomatology. Several experimental results demonstrate the feasibility and good performances of the approaches.     

基于图像处理的药片包装视觉检测系统滤波算法

目的为了有效滤除药片包装视觉检测系统中的噪声,提升图像清晰度,保证后期图像分割、边缘处理顺利进行。方法针对药片视觉检测图像中存在大量不确定噪声,提出一种自适应模糊神经网络的图像滤波算法。在模糊神经网络结构中引入一个鲁棒性较强的隶属函数,并通过梯度下降法对模糊神经网络中的参数进行优化训练,利用优化后的网络结构对被噪声污染的图像进行滤波处理。结果仿真结果表明,该算法能够在保留较完整的图像边缘和重要细节的前提下,有效滤除药片中的噪声。结论该滤波算法有效提高了药片图像的清晰度,对于后期药片图像分割以及边缘化处理具有重要意义。     

A Method Based on Piecewise Linear Models for Accurate Restoration of Images Corrupted by Gaussian Noise

Piecewise linear (PWL) models are very attractive for image processing due to their simplicity and effectiveness. A new filtering architecture adopting multiparameter PWL functions is proposed for accurate restoration of images corrupted by Gaussian noise. The filtering performance is analyzed by taking into account the different behavior from the point of view of noise removal and detail preservation. The sensitivity to a change of the parameter settings is also investigated. In the new approach, the parameter values are automatically selected by resorting to a procedure that estimates the standard deviation of the Gaussian noise. Results dealing with different test images and noise variances show that the method yields a very accurate restoration of the image data     

基于神经网络的图像混合滤波及融合算法研究

当图像中同时存在高斯噪声和椒盐噪声时,单一的均值滤波或中值滤波很难达到最佳滤波效果。 分析了噪声特点和各种滤波方法的优势,提出了一种基于神经网络的图像混合滤波及融合算法:首先建立概率神经网络,检测椒盐噪声和高斯噪声点,并分别利用中值滤波和均值滤波去除噪声点,然后建立径向基函数神经网络,利用训练好的径向基函数神经网络融合 2 种不同滤波的图像,输出理想的融合图像。 Matlab 仿真实验结果表明,该算法有效去除混合噪声的同时,能很好地保护图像的边缘与细节,是一种有效的方法。     

周期间隙性排气噪声滤波消声器的试验研究

本文分析了周期间隙性排气噪声的声学特性，提出对该噪声滤波消声器的要求，试验研究了不同滤波单元的滤波特性，在此基础上建立了降低该类噪声滤波消声器总成结构，并对其进行了台架及人工实际使用验证，证明该滤波消声器具有较高的实际应用与推广价值。     

去除脉冲噪声的自适应开关中值滤波

为消除图像中的脉冲噪声,提出了自适应开关中值(ASM)滤波算法。该算法采用一种新的噪声检测方法将图像中的像素分为信号点和噪声点两类。对检测出的噪声点统计其个数并由此估算图像中的噪声密度,根据估计的噪声密度自适应确定滤波窗口尺寸,采用改进的中值滤波对检测出的噪声点进行处理;而信号点则保留其灰度值不予处理。对ASM滤波进行仿真实验,结果表明,它能在有效去除噪声的同时很好地保护图像细节,较传统中值滤波及其它改进中值滤波算法有更优的滤波性能。     

噪声源识别的近场声全息方法和数值仿真分析

将近场声全息(NAH)用于噪声源的识别和定位，对不同类型噪声源的数值模拟结果和理论分析表明：对于复杂声源，采用近场声全息方法可以精确地定位噪声源，并且能很好地分辨出各噪声源振幅的强弱；在波数域加窗滤波后，声压测量的误差对声源识别结果的影响不大。并对重建结果误差的产生原因及近场声全息相关参数的选取原则做了详细的分析，对工程测量和噪声控制有一定的指导意义。     

A versatile analytical expression for the inverse Abel transform applied to experimental data with noise

A method for reconstruction of radially distributed plasma emission coefficients from projections with noise is proposed. The method represents the projections based on overlapping piecewise polynomial least squares fitting to take the inversion. Parameters that affect the inversion accuracy are analyzed and discussed in detail. Results for profiles with various shapes are presented and compared with those obtained with other methods. It is shown that for data with different numbers of points and different levels of noise, our method is more accurate and yields markedly better results for very sparse data. In addition, excellent results have been obtained from experimental intensities of an arc plasma without filtering of noise.     

气固两相流介质中声衰减测量方法研究

通过声波衰减谱反推气固两相流状态,是气固两相流声学监测研究的一种重要方法。在假设声能量的相对衰减量与声波传播距离成正比的前提下,推导出声压级随传播距离变化的表达式。设计搭建了测试两相流下声衰减的实验台,根据声压级随传播距离变化的关系,推导出实验计算气固两相流声衰减系数的表达式。实验条件下的噪声主要为风机的漩涡噪声和旋转噪声,以及机械噪声和气流的气动噪声。分析不同噪声的特点,提出了通过A计权网络和小波滤波相结合的滤除噪声的方法。     

基于双边滤波的自适应彩色图像去噪研究

目的为了克服彩色图像去噪后存在的特征模糊,研究基于双边滤波的自适应彩色噪声图像去噪方法。方法使用二维离散小波变换(DWT)对含噪声的彩图图像进行近似分量、水平细节分量、垂直细节分量和对角细节分量等4个方向的分解。根据DWT各方向分量归一化后的方差比例,利用RBF神经网络构造双边滤波系数模型确定不同方向的最佳去噪系数,提出彩色噪声图像自适应去噪方法(DWT-ABF),并将该方法与常规方法作对比。结果在不同噪声类型以及混合噪声失真情况下文中方法都能有效地去除噪声,并同时保留图像细节信息,且与其他方法相比,文中方法去噪后的图像都具有更高的PSNR值。结论文中方法克服了传统双边滤波无法自行确定最佳参数的缺陷,同时也良好地解决了去噪图像特征模糊的问题。     

用自联想网络处理带噪声的发动机数据

研究了用自联想网络(AANN)进行数字滤波的方法。自联想网络采用一种带有瓶颈层的特殊结构,且具有单位总增益。在经过大量带噪声样本的训练之后,各变量之间能够建立起内在联系。输入信息通过瓶颈层前的压缩及瓶颈层后的解压缩过程,信息中的精华将被提取,从而使人们能够利用冗余信息抑制其测量噪声,使发动机测量参数在最大程度上减少噪声对其带来的负面影响。     

噪声概率快速估计的自适应椒盐噪声消除算法

提出一种可识别噪声概率自动调节滤波窗口的自适应椒盐噪声消除算法。对非理想椒盐噪声污染图像随机区域进行变窗口中值滤波,将结果与滤波前比对获得噪声点数,滤波区域即按此点数排序。然后取每种滤波窗口下的中间三组数据,该数据平均加权获取图像噪声概率初估计,对初估计平均加权即得图像噪声概率。滤波前首先采用阈值法排除明显噪声点,剩余像素中再以离窗口中心像素距离平方的倒数为权值估计中心像素。最后由噪声概率按照T-S模糊规则对不同模型的输出估计值进行融合。实验证明,与传统中值滤波等算法相比,该算法具有噪声自动估计和自适应窗口调节能力,滤波后标准均方差可减少20%以上,速度可提高一倍多。     

Adaptive noise filtering using an error-backpropagation neuralnetwork

A neural network of the feedforward-error backpropagation type proposed by D.E. Rumelhart et al. (1986) was applied to filter noise from spectral data commonly encountered in infrared absorption of molecular transitions. The purpose was to gain insight into the way a neural network can be trained to remove noise from a noise-corrupted signal with implications for signal processing in general. The neural network simulation was implemented in Fortran and run on a VAX 8800. Training of the neural network occurred on a set of spectral data with random transitions and line shape parameters. Preliminary results of the performance of the adopted neural network are reported and discussed along with observed limitations. Future improvements on noise filtering using a neural network are proposed     

几种超声NDE信号降噪方法的性能比较

为提高超声无损检测的准确性,需要对超声NDE信号中因随机分布于媒质中的大量散射微粒所引起的结构噪声进行降噪。由于信号和噪声的频谱范围基本重叠,传统的线性滤波方法不能提供理想的降噪结果。介绍了几种对超声NDE信号进行降噪的新方法:Wigner-Ville分布法、小波变换法和基于时间延迟的神经网络法,并从信噪比(SNR)、检测概率(PD)和估测深度(ED)等三个重要参数对它们的降噪性能进行计算机仿真实验的比较。结果表明:小波变换法和神经网络法的降噪效果较Wigner-Ville分布法要好。对实际信号的测试还表明,小波变换由于不像神经网络那样需要训练,是一种更为理想的超声NDE信号降噪方法。     

Nonlinear filtering for recognition of phase-encoded images

We investigate the use of Fourier plane nonlinear filtering for phase-encoded images. We investigate the performance of the nonlinear joint transform correlator and the nonlinearly transformed matched filter for phase-encoded images with different types of input noise. We use the peak-to-output-energy ratio, peak-to-sidelobe ratio, and discrimination ratio as the metrics for measuring the performances. We mathematically analyze the peak-to-output-energy ratio of the nonlinearly transformed matched filter for phase-encoded images with spatially nonoverlapping white noise. Computer simulations are provided to show the performance improvements of the nonlinear filtering techniques for the phase-encoded images. In comparison with linear filtering techniques, we find that the nonlinear filtering techniques substantially improve the performance metrics. From the computer-simulation results it can be seen that the nonlinear joint transform correlator performs better than the nonlinearly transformed matched filter in detecting phase-encoded targets in the presence of different types of noise, such as additive overlapping white noise, spatially nonoverlapping white background noise, spatially nonoverlapping colored background noise, and nontarget objects.     

Denoising in digital speckle pattern interferometry using fast discrete curvelet transform

In digital speckle pattern interferometry, the denoising of speckle fringe patterns is of vital importance for quantitative extraction of phase distribution. A filtering method of fast discrete curvelet transform based on weighted average thresholding technique is proposed in this paper for noise removal in speckle fringe patterns. Both computer-simulated and experimental digital speckle pattern interferometry fringe patterns are adopted to evaluate the performance of the proposed filtering method. In addition, a widely used and representative filtering method, windowed Fourier filter, is introduced for making a comparison and validation in the image processing effect, and the parameter of peak signal noise ratio is also used for assessment of denoising effect. It is shown from the filtered results that the filtering method of fast discrete curvelet transform is effecitve to remove speckle noises and simultaneously preserve fringe structure information.     

微机械陀螺信号处理方法的比较研究

介绍了谱消除法、小波变换和数学形态学滤波三种信号处理方法进行消噪的基本原理,并分别采用这三种方法对微机械陀螺的静态漂移信号和动态仿真信号进行了去噪声处理,对处理前后的信号误差大小及其分布情况以及处理过程的运算量和实时性进行了比较.结果表明,对于静态漂移信号的处理,三种方法的滤波效果大致相当.对于动态信号,数学形态学滤波器的滤波效果比其它两种滤波器的滤波效果要好,并且,数学形态学滤波器具有运算量小,实时性较好的优点,十分易于硬件实现.     

基于WOSS的NS-Miracle水声信道模拟方法扩展

将世界海洋仿真系统(World Ocean Simulation System,WOSS)与声射线模型Bellhop结合并引入海洋噪声经验公式对NS-Miracle仿真系统的水声传输信道模拟方法进行扩展.为了验证扩展后的仿真系统,实验采用了基于水声信道特征经验模型、Bellhop射线模型及WOSS结合Bellhop射线...     

Impulse noise reduction using hybrid neuro-fuzzy filter with improved firefly algorithm from X-ray bio-images

Noise filtering performance in medical images is improved using a neuro-fuzy network developed with the combination of a post processor and two neuro-fuzzy (NF) filters. By the fact, the Sugeno-type is found to be less accurate during impulse noise reduction process. In this paper, we propose an improved firefly algorithm based hybrid neuro-fuzzy filter in both the NF filters to improve noise reduction performance. The proposed noise reduction system combines the advantages of the neural, fuzzy and firefly algorithms. In addition, an improved version of firefly algorithm called searching diversity based particle swarm firefly algorithm is used to reduce the local trapping problem as well as to determine the optimal shape of membership function in fuzzy system. Experimental results show that the proposed filter has proved its effectiveness on reducing the impulse noise in medical images against different impulse noise density levels.