一种全方位多角度自适应形态滤波器及其算法

根据形态滤波器的结构元特性,结合形态开—闭(闭—开)组合运算及自适应算法,建立了一种能有效去除复杂纹理图像中的各种噪声,并较好的保持图像细节的方法。实验证明,这种全方位多角度自适应形态滤波器与现有的中值滤波器、全方位结构元形态闭—开最小、开—闭最大滤波器等非线性滤波器相比较有更好的降噪性能及细节保护能力,扩展了形态滤波器在滤除高斯噪声方面的应用,提高了形态滤波器的适应性。     

一种基于神经网络的形态滤波器优化设计方法

给出了一种利用神经网络实现灰度形态滤波器参数的优化设计方法。文中首先简述了形态滤波器的基本理论,并设计出了灰度形态学的两种最基本运算(膨胀和腐蚀)的神经网络结构,其网络权值即为形态运算的结构元素值。然后,按照d学习规则,自适应地对结构元素(形状和数值)进行学习训练,以获得最佳的灰度形态滤波器参数。该方法设计简便,实用性强且易于推广,对提高形态滤波性能效果显著。     

二值图象的多结构元数学形态滤波

与单结构元形态滤波相比，多结构元形态滤波能够保护或提取更多的图象几何特征．本文提出了二值图象的多结构元腐蚀、膨胀、开、闭四种多结构元形态滤波器，讨论了四种滤波器的基本性质和滤波性能，并将他们纳入完备网格空间一般形态变换的框架，最后讨论了多结构元形态滤波器的设计．     

图像滤波的形态学开、闭型神经网络算法

该文设计完成了一种具有实用意义的形态学开、闭滤波的神经网络模型及其滤波参数的优化训练算法。实验结果表明该方法设计简便,实用性强且易于推广,对提高形态滤波性能效果明显。分析表明,形态滤波器可分解为形态滤波运算和结构元素选择两个基本问题。形态滤波运算规则已由定义本身确定,于是形态滤波器的最终滤波性能就仅仅取决于结构元素的选择。进行自适应优化训练的目的正是使结构元素具有图像目标的形态结构特征,从而使形态滤波器对复杂变化的图像具有良好的滤波性能和稳健的适应能力。     

复杂背景小目标实时检测技术

针对ATR系统小目标识别实时性的需求,提出了一种快速的小目标检测算法,该方法在分析形态滤波的基本运算即膨胀、腐蚀后发现:形态滤波计算量大、实时性差的原因在于膨胀和腐蚀的计算过程中存在比较运算的冗余,即当新像素进入结构元中时,需要重新比较所有像素的灰度值,而提出的形态滤波算法则仅需把新像素的灰度值与前一次的结果值相比较而得到该次运算的结果,这样就大大减少了比较次数而提高了算法的速度.最后将算法移植到以TSM320C6203为核心的硬件平台上实现了目标检测的实时处理.实验表明:该快速算法较一般算法的计算时间减少70%.     

扫描隧道显微镜的新型实时图像处理系统

本文介绍了一种用于扫描隧道显微镜的实时图像处理系统。通过改变形态学运算，定义条件腐蚀、条件膨胀、开运算、闭运算、可滤除幅度大于特定值的噪声信息。选择适当的模板结构和门限阈值，这种改进的形态滤波器既保留了传统形态滤波器的优点，又克服了它的特点。     

基于FPGA和符号LMS算法的自适应滤波器设计

为了减少实现基于FPGA和LMS算法的自适应滤波器过多消耗硬件资源的问题,提出了符号LMS算法,通过降低乘法运算的次数来提高自适应滤波器的运行速度,并使用流水线技术进行优化。软件仿真验证了符号LMS算法的可行性,硬件仿真证实了采用该算法和流水线技术的自适应滤波器的优越性。     

多结构元受控形态滤波器

基于受控形态学中的形态开-闭和闭-开运算,采用多结构元素,构造了一类多结构元受控形态滤波器.该滤波器具有平移不变性、递增性、对偶性和幂等性等重要性质.多结构元受控形态滤波器不仅可以有效地抑制图像噪声,而且较好地保持了图像的几何结构特征,通过与传统形态滤波器的对比实验,进一步验证了该滤波器的优越性能.     

基于小波变换与形态学运算的ECG自适应滤波算法

针对ECG信号常用滤波算法存在的缺陷,提出了基于小波变换与形态学运算的自适应滤波新算法。该算法利用形态学滤波器去除基线漂移信号,用小波滤波器去除高频干扰信号,并将这两部分所得到的心电噪声分量作为自适应滤波器的参考输入信号,利用自适应滤波器调整对含噪ECG信号进行滤波处理。最后,经实验验证了本文算法的有效性。     

变抽头长度LMS自适应滤波算法

该文将自适应滤波器抽头长度与权值调整问题归结为单一的权值调整问题,提出了抽头长度的一般更新公式及新的变抽头长度LMS算法,从理论上分析了其合理性与收敛性。新算法用长滤波器与短滤波器的时平均平方误差估计稳态均方误差,采用了自适应调整的抽头长度步长,可在滤波器权值未收敛时就快速更新抽头长度。论文还证明了目前文献中几种有效的变抽头长度算法也可看作或化为文中抽头长度一般更新公式的特例,理论分析与自适应系统辨识的仿真结果验证了新算法的有效性。     

Structuring element decomposition for efficient implementation ofmorphological filters

In order to implement morphological filters on image processing systems. The size of the structuring element must be small due to the architectural constraints of the systems. The decomposition of the structuring element is required for filters with large structuring elements. An algorithm for the decomposition of an arbitrary 4-connected binary structuring element is developed which enables sub-optimal implementation of binary morphological filters on a 3×3 pipeline machine. The given structuring element is first decomposed into the union of elements using a sequential search procedure, then each element is further decomposed optimally into 3×3 elements, resulting in final sub-optimal 3×3 hybrid decomposition. The proposed algorithm is applied to some structuring elements and results close to the optimum are obtained     

Computationally efficient,one-pass algorithm for morphological filters

Many useful morphological filters are built as long concatenations of erosions and dilations: openings, closings, size distributions, sequential filters, etc. This paper proposes a new algorithm implementing morphological dilation and erosion of functions. It supports rectangular structuring element, runs in linear time w.r.t. the image size and constant time w.r.t. the structuring element size, and has minimal memory usage.It has zero algorithm latency and processes data in stream. These properties are inherited by operators composed by concatenation, and allow their efficient implementation. We show how to compute in one pass an Alternate Sequential Filter (ASFⁿ) regardless the number of stages n.This algorithm opens the way to such time-critical applications where the complexity and memory requirements of serial morphological operators represented a bottleneck limiting their usability.     

Adaptive Fuzzy Morphological Filtering of Impulse Noise in Images

In this paper we first introduce a neural network implementationfor fuzzy morphological operators, and by means of a trainingmethod and differentiable equivalent representations for theoperators we then derive efficient adaptation algorithms to optimizethe structuring elements. Thus we are able to design fuzzy morphologicalfilters for processing multi-level or binary images. The convergencebehavior of basic structuring elements and its significance forother structuring elements of different shape is discussed. Besidesthe filter design, the localized structuring elements obtainedfrom the training method give a structural characterization ofthe image which is useful in many applications. The performanceof the fuzzy morphological filters in removing impulse noisein multi-level and binary images is illustrated and comparedwith existing procedures.     

Direct implementation of open-closing in morphological filtering

An algorithm for the direct implementation of commonly used 1D morphological filters is presented. Open-closing and close-opening operations are performed with a single-pass procedure resulting in a substantial increase in speed, proportional to the length of the structuring element, over alternative implementations     

Rotated texture classification by improved iterative morphologicaldecomposition

An improved algorithm based on iterative morphological decomposition (IMD) proposed by Wang et al. (1993) is described. The proposed algorithm requires less computation than the original IMD algorithm. The improved iterative morphological decomposition (IIMD) is compared with granulometric moments, multiresolution rotation-invariant SAR (MRRISAR) models and multichannel Gabor filters. It is found that the IIMD is superior to granulometric moments and MRRISAR in rotated texture classification. The IIMD may also perform better than multichannel Gabor filters by employing many different kinds of structuring elements. In the study, three kinds of pseudo rotation-invariant structuring elements, namely the disc, octagon and square, as well as a line structuring element are tested. Since the line structuring element is rotation-variant in nature, the image is rotated to different orientations of equal angular separation to find a set of primitive features. A Fourier transform is then applied to convert these features to rotation-invariant ones. An accuracy rate as high as 96% is achieved in classifying 30 classes of textured images in the experiment. It is also demonstrated that using both the normalised variance and the mean can give a better classification accuracy rate than using both the variance and the mean when classified by a simplified Bayes or Mahalanobis distance measure     

Spatially variant morphological restoration and skeleton representation.

The theory of spatially variant (SV) mathematical morphology is used to extend and analyze two important image processing applications: morphological image restoration and skeleton representation of binary images. For morphological image restoration, we propose the SV alternating sequential filters and SV median filters. We establish the relation of SV median filters to the basic SV morphological operators (i.e., SV erosions and SV dilations). For skeleton representation, we present a general framework for the SV morphological skeleton representation of binary images. We study the properties of the SV morphological skeleton representation and derive conditions for its invertibility. We also develop an algorithm for the implementation of the SV morphological skeleton representation of binary images. The latter algorithm is based on the optimal construction of the SV structuring element mapping designed to minimize the cardinality of the SV morphological skeleton representation. Experimental results show the dramatic improvement in the performance of the SV morphological restoration and SV morphological skeleton representation algorithms in comparison to their translation-invariant counterparts.     

A study of the generalized morphological filter

A new class of morphological filters is proposed for image enhancement. The filter, known as the generalized morphological filter (GMF), uses multiple structuring elements and combines linear and morphological operations. The GMF can be designed to suppress various types of noise yet preserve geometrical structure in an image. A study of several aspects of the performance of the filter is presented. The study includes geometrical feature preservation, noise suppression, structuring element selection, and the root signal structure. For the sake of comparison, averaging and median filters are also used in the experiments and corresponding figures of merit of the performance of the filter. The empirical study shows that the generalized morphological filter possesses effective noise suppression with reduced geometrical feature blurring.This work was supported by the National Science Foundation, under Grant No. CDR-8803017 to the Engineering Research Center for Intelligent Manufacturing Systems.     

Statistical evaluation of sequential morphological operations

In order to properly apply sequential morphological operations to random signals in applications concerned with noise suppression, the authors have examined their statistical properties using different structuring elements. The performance of flat and triangular structuring elements has been evaluated for signals with uniform, Gaussian, and Rayleigh density functions. In particular, the statistical properties of sequential morphological operations (i.e,, dilation, closing, clos-erosion, and clos-opening) are examined as a function of the parameters of the structuring element through Monte Carlo simulation, which overcomes the statistical dependency problem arising in the processed signal at different stages of morphological operations. The simulated results and their statistics (mean, variance, and skewness) present an interpretation of the signal root, biasing effects, and noise suppression capability of morphological filters     

Gray-scale structuring element decomposition

Efficient implementation of morphological operations requires the decomposition of structuring elements into the dilation of smaller structuring elements. Zhuang and Haralick (1986) presented a search algorithm to find optimal decompositions of structuring elements in binary morphology. We use the concepts of Top of a set and Umbra of a surface to extend this algorithm to find an optimal decomposition of any arbitrary gray-scale structuring element.