煮糖过程糖结晶显微图像处理方法研究

由于蔗糖晶体的镜面透明,煮糖显微观察镜采集出来的图像变为了伪三维的图像,失真,不利于制糖过程的自动监控;针对这类特殊图像探求一种较好的图像预处理方法,首先采用几种常用的图像预处理方法如：灰度调整,照度均匀,直方图均衡,对数变换及中值滤波的方法对伪三维图像进行预处理,再对处理结果运用分割效果较好的Canny算法分割,并将最终结果进行比较分析,经比较,采用灰度调整和中值滤波的预处理的图像,分割效果较其它方法理想,分割结果适用于煮糖过程的自动监控.     

基于图像处理的蔗糖结晶颗粒识别方法

针对煮糖过程蔗糖结晶图像的特点采用颜色空间转换的方法,将图像从RGB颜色空间转换为HIS颜色空间,利用HIS颜色空间各分量相对独立性以及结晶颗粒和糖浆溶液的色调差异通过改进的大津法对H分量进行阈值分割,再通过数学形态学,中值滤波,孔洞填充,去除噪声颗粒等方法进行后续处理,最后对处理后的图像提取结晶颗粒的个数、面积、周长、形状指数以及面积比等特征值并进行分析来完成煮糖过程中蔗糖结晶颗粒的识别。实验证明该方法识别率高,切实可行。     

基于HSI颜色空间的蔗糖结晶图像分割方法

针对煮糖过程蔗糖结晶图像的特点,采用颜色空间转换的方法,将图像从RGB颜色空间转换到HSI颜色空间,利用HSI颜色空间各分量相对独立性以及结晶颗粒和糖浆溶液的色调差异通过改进的大津法对H分量进行阈值分割,再通过数学形态学,中值滤波,孔洞填充,去除噪声颗粒,去除图像边界的非完整颗粒等方法进行后续处理,最终将结晶颗粒从复杂的图像信息中分割出来。实验证明该方法分割效果好,切实可行。     

图像识别技术在蔗糖颗粒检测中的应用

研究设计了蔗糖煮糖过程中蔗糖结晶颗粒的检测系统,利用图像识别的方法避免传统蔗糖结晶检测中的物理接触,减轻工人劳动强度,使检测结果更加精确,实现计算机控制煮糖结晶过程,有效地促进了蔗糖产业的长期发展.首先利用百万像素CCD和光学显微镜获取蔗糖结品的颗粒图像,用网像采集卡完成信号的计算机输入,利用形态学等算法预处理待测图像,然后运用基于遗传算法的otsu法分割图像得到二值图像,最后提取特征值计算出颗粒的个数及面积.试验结果表明处理后的图像分割效果很好,能够准确计算出颗粒特征值,实现了蔗糖结品颗粒的自动检测.     

面向彩色手术显微图像的去噪算法改进

对于要求高保真的彩色手术显微图像,去除采集过程中引入的脉冲噪声是一项非常重要的任务。将自适应矢量中值滤波方法应用于彩色图像去噪,其效果非常理想。该方法可根据噪声干扰的情况自动选择滤波窗口的大小。先对图像各区域进行噪声检测,如果为非噪声区域,则不进行滤波;如果为噪声区域,则根据各区域受噪声污染状况自动确定滤波窗口尺寸。实验显示,采用自适应矢量中值滤波能有效滤除彩色显微图像的噪声,并能较好地保护边缘和细节,其性能较一般的矢量中值滤波方法有明显优势。     

煮茧机温度压力简易调节系统在丝厂的应用

丝绸行业中应用的简易式温度压力自动调节系统,是为满足煮茧机温度压力指标而设计的。在丝绸生产中,煮茧质量好坏将直接影响生丝质量,人们希望通过煮茧,使不同大小、厚薄、不同丝胶含量,不同性质的原料茧,达到较好的一致性。真空渗透煮茧工艺过程为:蚕茧真空渗透→进煮茧机→预热→触蒸→低温吸水→动摇→静蒸→保护→出茧。在整个工艺     

基于双背景色阈值面积消去算法的B超图像轮廓提取

本文提出一种基于双背景色阈值面积消去算法的半自动单像素轮廓提取方法.通过给定的阈值,此算法可自动在图像中寻找面积小于阈值的斑点和噪声点(即空穴)并将其消除,有效解决了B超图像处理中由于斑点及噪声引起的完整单像素轮廓提取困难的问题.并基于此方法,以心房图像为例讨论了针对二维B超图像的轮廓提取方案.     

基于SSCL的模糊C均值图像分类方法

针对传统模糊C均值聚类方法对噪声敏感和过分依赖于初始聚类中心的缺点,提出基于SSCL的模糊C均值图像分类的自适应算法。该算法首先通过SSCL获得初始类别数和类别中心,然后作为模糊C均值聚类的输入,自动对图像进行分割,并对图像分割结果利用空间信息进行后处理。实验结果表明该方法较好地解决了FCM算法中的初始化和噪声敏感问题,具有较好的分类结果。     

加入PSO算法在蔗糖结晶模糊控制中的应用

在煮糖过程中人为因素对出糖数量和质量都有很大的影响,而蔗糖结晶过程难以建立准确的数学模型使得传统的控制方法无法实现精准控制,在总结经典PID控制和已经实际应用的蔗糖结晶过程模糊控制理论的基础上,提出了采用模糊推理对PID控制器参数进行自整定,加入改进粒子群算法对控制器过程参数进行优化调整,完成融入PSO算法的蔗糖结晶过程模糊控制器的设计,利用每次控制输出反馈自整定来提高控制器性能;通过Matlab软件模拟仿真测试效果表明,该方法增强了系统鲁棒性,可以提高出糖的品质。     

一种改进的多天线信号能量检测方案

能量检测（ED）方法是最常用的信号检测方法之一,其具有实现复杂度低和不需要信号先验信息的优点。当噪声方差已知时,能量检测算法可以获得较好的检测性能。在大多数情况下,噪声的方差是需要预估的,因此噪声方差估计的不确定性会对能量检测算法性能造成较大的影响。为了减小由噪声方差不确定性造成的影响,提出了一种改进的多天线能量检测方案。方案将多天线采集的信号进行简单组合,从而构建出一种与噪声方差无关的判决统计量。理论公式表明,方案的检测概率和虚警概率是与噪声方差无关的。仿真结果显示,当天线的数量大于3时,方案的检测性能优于传统的能量检测性能。     

An application of Non-Parametric Predictive Inference on multi-class classification high-level-noise problems

This paper presents an application of the Non-parametric Predictive Inference model for multinomial data (NPIM) on multiclass classification noise tasks, i.e. classification tasks where the variable under study has 3 or more possible states or values; and the data sets have incorrect class labels in their training and/or test data sets. In an experimental study, we show that the combination or fusion of the information obtained from decision trees built using the NPIM in a Bagging scheme, can improve the process of classification in multi-class classification noise problems. Via a set of statistical tests, we compared this approach with other successful methods used in similar scheme, on a wide set of data sets. It must be remarked that the new approach has a notably performance, compared with the rest of models, when the level of noise is increased.     

横向联邦学习中PCA差分隐私数据发布算法

为了让不同组织在保护本地敏感数据和降维后发布数据隐私的前提下,联合使用PCA进行降维和数据发布,提出横向联邦PCA差分隐私数据发布算法。引入随机种子联合协商方案,在各站点之间以较少通信代价生成相同随机噪声矩阵。提出本地噪声均分方案,将均分噪声加在本地协方差矩阵上。一方面,保护本地数据隐私;另一方面,减少了噪声添加量,并且达到与中心化差分隐私PCA算法相同的噪声水平。理论分析表明,该算法满足差分隐私,保证了本地数据和发布数据的隐私性,较同类算法噪声添加量降低。实验从隐私性和可用性角度评估该算法,证明该算法与同类算法相比具有更高的可用性。     

An icosahedron-based method for even binning of globally distributed remote sensing data

A new scheme is presented for binning globally distributed measurements. The scheme is based on a network of evenly distributed grid points, defined by repeated subdivision of a spherical icosahedron. Delanuany triangulation is then used to obtain bin perimeters for each grid point, which results in a network of bins that are evenly distributed across the entire globe and have uniform area. A modified winding rule is used to determine which datapoints are in which bin. This binning method is especially suited to remote sensing applications involving datasets covering polar regions, where conventional rectangular latitude/longitude bins introduce distortion and streaking into the binned data if noise is present. It also has the property that adjacent bins overlap, providing Nyquist sampling and preventing spatial aliasing. Tests on synthetic data show that this icosahedral binning scheme preserves underlying data trends and is robust to noise.     

Identification of linear relations from noisy data: Geometrical aspects

The problem of determining linear models from data corrupted by noise has been revisited by several authors in recent years. With reference to the Frisch scheme, this work presents some geometrical aspects regarding the sets of solutions obtained in the identification procedure. An analysis is made on how these sets vary as the amount of noise on the data varies; the existence of particular solutions for data linked by a number of linear relations greater than one, is also shown.     

Effect of noise on deterministic joint remote preparation of an arbitrary two-qubit state

Quantum communication has attracted much attention in recent years. Deterministic joint remote state preparation (DJRSP) is an important branch of quantum secure communication which could securely transmit a quantum state with 100% success probability. In this paper, we study DJRSP of an arbitrary two-qubit state in noisy environment. Taking a GHZ based DJRSP scheme of a two-qubit state as an example, we study how the scheme is influenced by all types of noise usually encountered in real-world implementations of quantum communication protocols, i.e., the bit-flip, phase-flip (phase-damping), depolarizing, and amplitude-damping noise. We demonstrate that there are four different output states in the amplitude-damping noise, while there is the same output state in each of the other three types of noise. The state-independent average fidelity is presented to measure the effect of noise, and it is shown that the depolarizing noise has the worst effect on the DJRSP scheme, while the amplitude-damping noise or the phase-flip has the slightest effect depending on the noise rate. Our results are also suitable for JRSP and RSP.     

结合零水印的小波包图像水印方案

论文提出了一种结合零水印的小波包图像水印方案。它利用零水印算法对信号处理具有很强的鲁棒性的优点,及小波域图像水印技术所具有的较好的直观性、较强的说服力的特点,将两种方法结合在一个模型中,利用两种策略的互补性,有效获得了水印直观、形象和说服力,同时保证了水印的鲁棒性。实验证明,该水印方案对剪切、加乘性噪声、滤波、JPEG压缩等图像处理具有优异的鲁棒性。     

可穿戴装置个性化本地差分隐私保护方案

本地差分隐私(local differential privacy,LDP)可以对可穿戴装置(wearable devices)采集到的数据进行隐私保护,每个用户都会在本地扰乱自己的数据,并且将扰动后的数据发送给数据汇聚服务器,以保护用户免受私人信息泄漏的影响.可穿戴装置采集到的数据是多维的,但是现有的针对可穿戴装置多...     

Certain models from uncertain data: The algebraic case

The problem of deriving possible linear relations from data affected by additive noise has received remarkable attention in recent years particularly regarding the assumptions (‘prejudices’) behind the procedures leading to unique models. Unlike traditional approaches leading to unique solutions (least squares, maximum likelihood, etc.) the Frisch scheme, belonging to the family of Errors-in-Variables (EV) schemes, leads to a whole family of models compatible with a set of noisy data and is considered as mildly affected by prejudices. This paper shows how also under the assumptions of the Frisch scheme it is possible to obtain a unique model from uncertain data and also to derive the actual amount of noise when the noiseless data are linked by a single linear relation. The more general EV case of non-independent additive noises is then considered and it is shown how also under these assumptions it is possible to obtain the unique set of parameters linking the noiseless data and how the family of compatible noise covariance matrices is defined, in this case, by the infinite elements of a linear variety.     

数字通信系统中的符号时钟恢复技术

在数字通信系统中,由于有高斯噪声和多径的影响,接收信号产生损失,从而导致时钟信号的提取更加困难,而时钟信号的不准确性会降低整个系统的性能。本文我们给出一种改进的时钟恢复算法原理,算法主要包含简单有效的插值滤波模块,改进的Gardener算法和快速收敛的PLL。该算法可以适用于宽带无线通信系统中的数字接收机中,采用该算法的数字接收系统已经用FPGA验证通过。     

基于系统调用顺序和频度特性的入侵检测模型

提出了一种将系统调用的顺序特性和频度特性相接合来构建入侵检测模型（COFIDS模型）的新方法，该模型采用kNN(k-Nearest Neighbor Classifier)算法实现入侵检测，并利用一种改进的相似因子，来增加系统调用序列间相似度的差别，减少了识别误差，提高了检测率，降低了入侵检测的误报率。实验表明，COFIDS还具有较强的抗噪声干扰的能力。抗噪声干扰的能力。