铁轨表面缺陷的视觉检测与识别算法

提出一种铁轨表面缺陷的视觉检测与识别算法。设计铁轨表面缺陷视觉检测与识别系统的总体结构,基于水平投影法提取铁轨表面区域,采用逻辑操作组合检测结果,使用 BP 神经网络进行缺陷分类。实验结果表明,该算法能准确地检测与识别铁轨表面的疤痕和波纹擦伤这2种缺陷,分类正确率分别达到99%和95%。     

钢板表面质量机器视觉检测系统设计

针对国内钢厂采用人工方法检查钢板表面缺陷存在可靠性差的问题,开发设计了基于机器视觉技术的带钢表面缺陷自动检测系统.系统通过摄像头采集带钢表面的图像,然后采用图像处理及模式识别算法对图像进行实时处理和分析,从而检测出钢扳表面缺陷,并对缺陷进行自动分类识别.实验结果表明,系统能够对带钢表面进行实时在线监测,并能正确识别常见的带钢表面缺陷.     

基于纹理提取和SVM技术的自动木材缺陷识别

要：支持向量机（SVM）是一种新的模式识别方法,有较好的泛化能力和推广能力。研究了基于纹理提取和支持向量机的自动木材表面缺陷的识别问题,借助LBP纹理特征提取技术实现对木材图像数据降维处理,并研究了木材表面不同类型缺陷的分布规律。利用支持向量机分类算法对木材表面有无缺陷进行了快速准确的自动识别,实现了木材表面缺陷的自动定位。多次交叉实验表明,SVM分类算法对木材表面缺陷具有较好的识别能力,识别率可达96％以上。     

带钢表面缺陷图像的可拓分类算法

针对带钢表面缺陷难以识别和分类的问题,将可拓集合理论应用于带钢表面缺陷图像的分类,由缺陷图像的灰度共生矩阵计算出的能量、熵、惯性矩和相关性作为分类的特征向量,通过对大量图像数据的统计分析,确定不同缺陷类别的各参数经典域和节域,计算出待分类缺陷相对于各缺陷类别的加权关联度值,由最大值确定待分类缺陷所属类别。选择实际图像进行分类实验,仿真结果显示能够取得较好的分类效果。     

基于支持向量机的钢板缺陷分类问题的研究

现实工业生产中,钢板表面存在不同类型的缺陷,为了研究这些缺陷类型,需要对大量钢板进行特征提取,从缺陷中提取有价值的属性或度量。随后对提取的特征进行选择,选择降低缺陷分类错误的特征信息。在钢板表面缺陷检测系统中,缺陷识别是关键步骤之一,属于多分类问题。采用主成分分析对初始数据进行降维处理,然后采用支持向量机作为分类器,对钢板表面缺陷进行分类,以研究钢板的缺陷类型。同时采用基于Keras的神经网络进行对比分类,并优化钢板缺陷分类。     

一种带钢表面缺陷识别与分类的研究——基于混合加权特征和RBF网络的方法

本论文中,采用灰度直方图特征、灰度共生矩阵特征和小波变换特征的提取方法,三种特征方法的结合能够很好的实现分类的目的。在提取特征向量的基础上,本研究基于MATLAB6.5环境下的神经网络工具箱,采用了兼顾识别速度与分类准确性的RBF神经网络分类器对带钢表面缺陷进行识别与分类,此算法可以作为高速生产线的带钢表面缺陷的实时检测优选方案。     

一种基于智能机器视觉识别的工业探伤技术

磁粉探伤广泛应用于铁磁性工件表面和近表面缺陷检测,但目前的缺陷识别仍依赖于人工判断。为实现检测自动化,将机器视觉技术应用于缺陷识别,提出一种智能的螺栓缺陷在线检测系统,针对螺栓在生产过程中产生的纵向、横向裂纹,使用基于统计的磁痕裂纹识别方法和基于改进LBP特征和SVM分类器识别方法进行级联完成螺栓零件表面裂纹的自动识别。实验结果表明,该方法与传统的识别方法对比具有很好的分类识别效果,漏检率减少5%,同时能够满足实时性的需求。     

冷轧带钢表面缺陷在线检测软件体系结构设计

为解决带钢表面缺陷在线检测系统的信息支持和后续处理的问题,遵循软件工程的设计思想,提出了一套完整的冷轧带钢表面缺陷在线检测系统的软件体系结构设计方案,建立了系统的层次结构模型、功能结构模型及其后台数据库,通过可视化程序设计和关系数据库实现了检测软件的系统功能.该系统不仅能够识别分类各种缺陷,而且可通过数据库保存各类数据信息,便于对各类信息数据进行查询和管理,并解决了系统算法更新的问题.该体系结构的设计为在线检测缺陷分类识别的进一步研究奠定了基础,具有一定的实用价值.     

遗传算法在缺陷特征选择中的研究

由于冷轧带钢表面缺陷的类型多种多样,在所提取的特征集中,需要寻找出一组较优的特征集,使之可以更有效地表达缺陷的本质特征,从而提高缺陷识别的准确率。针对该问题,研究了遗传算法在缺陷特征选择中的应用,并在充分研究信息熵理论的基础上,以平均净分类信息为遗传算法的适应度函数,以弥补互信息熵作为适应度函数所导致的不足。实验表明,利用遗传算法得到的特征集,对现场的冷轧带钢表面缺陷进行分类时,能得到更高的分类准确率。     

基于复杂网络特性的带钢表面缺陷识别

针对带钢表面缺陷识别问题,提出一种基于动态演化复杂网络特性的特征描述方法, 这些特征同时具有位移、旋转不变性、大小不变性、较强的抗干扰能力和鲁棒性,为 缺陷识别提供良好的分类特征;为了提高分类器的效率,应用主成分分析法 (Principal component analysis, PCA) 对复杂网络特 征向量进行特征降维处理;采用最优有向无环图支持向量机 (Directed acyclic graph support vector machine, DAG-SVM)算法进行缺陷分类.结果表明该方法识别率高而且识别速度快.     

Realistic imitation of mosquito's proboscis: Electrochemically etched sharp and jagged needles and their cooperative inserting motion

Aiming at the use in low-invasive medical treatments, this paper proposes a realistic imitation of mosquito's proboscis. A silicon needle is electrochemically etched, making the three-dimensionally sharp tip with finely smooth surface. The jagged shank shape is machined by a deep reactive ion etching (DRIE). The combined needles comprising a central straight needle and two outer jagged needles are fabricated, imitating a labrum and two maxillas of the mosquito, respectively. The cooperative motion of the three needles imitating the mosquito's motion is realized by applying PZT actuators independently to all the needles. The effectiveness of inserting these needles cooperatively was experimentally confirmed. Considering practical medical application, a biodegradable polymer needle with three-dimensionally sharp tip is also developed. The fabrication process based on micromolding is as follows: a nickel negative cavity is made by electroplating on a silicon sharp needle, to which melted polymer is injected, and it is finally released using a lost molding technique. The effectiveness of sharp tip for easy insertion was experimentally proven.     

计算机视觉检测技术在吻合器检测中的应用研究

提出了一种利用计算视觉检测带手柄吻合器漏针新方法。通过采用正面光源投照的检测装置,综合运用图像处理和识别方法对利用CMOS相机采集的吻合器图像进行数学形态学,图像增强等方法处理,并提出模糊局部阈值分割算法实现吻合器漏针的自动识别,并根据面积特征判别漏针数。实验表明：该方法能够对多种型号颜色的吻合器进行稳定、准确的快速检测,检测正确率可达95.6％,平均每次检测时间为2.385s,本方法实用性强,适合现代化生产的要求。     

中医针刺手法用传感针的研制

为了能够科学地研究分析各种针刺手法,获得手法的客观参数,应用微小力传感技术和生物力学原理,研制了一套用于针刺手法提插和捻转操作的实时检测传感针,配合相应的计算机采集硬件和软件系统,能够对针刺手法过程进行检测记录和分析.通过对人体穴位进行临床均匀提插和捻转操作,实现了针刺过程中针体受力的定量化测量,为针刺手法的研究和针刺疗效的客观描述提供了实验手段和科学依据.     

避雷针保护范围三维可视化

提出了避雷针保护区域可视化设计系统要实现的基本功能：三维显示保护区域、生成三视图的工程图纸、智能化设计避雷针的高度、数量、位置参数等;设计了系统的整体框架和模块间的数据传递关系,采用VC＋＋并结合OpenGL技术,实现了单针和双针等高、不等高及任意三针避雷针保护曲面的三维可视化．     

Fabrication of a 2-D in-plane micro needle array integrated with microfluidic components using crystalline wet etching of (110) silicon

We propose a two-dimensional (2-D) in-plane micro-needle array with shaft sidewalls aligned parallel to the vertical (111) crystalline plane of (110) silicon. Six types of needle tips with various shapes and tapered angles were fabricated so as to maintain the tip sharpness. Two layers of micro needles (upper and bottom needle arrays) for the 2-D array were realized using simultaneous etching from the front and back sides of (110) silicon. In addition, microfluidic components were embedded in the micro-needle chip to inject or extract biochemical samples. The length of the micro needles was easily extended to 2200 μm, and the insertion forces of the single and arrayed micro needles were evaluated by pricking chicken breast flesh. In case of a micro needle having a tapered angle of 10° and tip end width of 1 μm, the insertion force per needle was as low as 15 mN, which is lower than those reported in previous studies.     

Spectral identification of ozone‐damaged pine needles

Needles were collected from ponderosa and Jeffrey pine trees at three sites in the Sierra Nevada, and were assembled into 504 samples and grouped according to five dominant live needle conditions – green, winter fleck, sucking insect damage, scale insect damage, and ozone damage – and a random mixture. Reflectance and transmittance measurements of abaxial and adaxial surfaces were obtained at ca 0.3 nm spectral resolution from 400–800 nm, and binned to simulate Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) data. There were no significant differences in optical properties between the two surfaces. Ozone‐damaged needles were collected from Jeffrey pine trees at one site, and exhibited significantly different (family‐wise α = 0.01) reflectance and transmittance signatures – and significantly different signature slopes – at both spectral resolutions, from green and winter fleck needles from the same site. Ozone‐damaged needles had significantly different (family‐wise α = 0.01) abaxial surface reflectance and reflectance slope signatures from all other groups of needles, at both spectral resolutions. In comparison with three chlorophyll reflectance indices, a new red fall index (RFI) provides high classification accuracies for ozone‐damaged and non‐ozone‐damaged pine needles (overall acc. = 94%; κ = 59%). Thus, ozone‐damaged Jeffrey pine needles have a unique spectral signature in relation to dominant needle conditions of ponderosa and Jeffrey pine trees.     

The Path-of-Probability Algorithm for Steering and Feedback Control of Flexible Needles

In this paper we develop a new framework for path planning of flexible needles with bevel tips. Based on a stochastic model of needle steering, the probability density function for the needle tip pose is approximated as a Gaussian. The means and covariances are estimated using an error propagation algorithm which has second order accuracy. Then we adapt the path-of-probability (POP) algorithm to path planning of flexible needles with bevel tips. We demonstrate how our planning algorithm can be used for feedback control of flexible needles. We also derive a closed-form solution for the port placement problem for finding good insertion locations for flexible needles in the case when there are no obstacles. Furthermore, we propose a new method using reference splines with the POP algorithm to solve the path planning problem for flexible needles in more general cases that include obstacles.     

基于KPCA和灰色模型的凝汽器故障预测

汽轮机凝汽器的故障预测为其故障自愈的研究提供了理论依据.提出一种基于核主元分析和灰色预测模型的汽轮机凝汽器故障预测方法,首次将灰色预测理论应用于凝汽器的故障预测.采用核主元分析法对故障特征数据进行分析和处理,提取反映故障的主要特征量,以降低特征变量之间的非线性相关性,同时减少灰色预测模型的预测参数的数目.然后应用灰色预测理论建立故障特征的预测模型,对每一个主要特征量的趋势值进行预测,重构故障特征向量,用于汽轮机凝汽器故障的预测分析.     

Image Guidance of Flexible Tip-Steerable Needles

Image guidance promises to improve targeting accuracy and broaden the scope of medical procedures performed with needles. This paper takes a step toward automating the guidance of a flexible tip-steerable needle as it is inserted into the human tissue. We build upon a previously proposed nonholonomic model of needles that derive steering from asymmetric bevel forces at the tip. The bevel-tip needle is inserted and rotated at its base in order to steer it in 6 DOF. As a first step for control, we show that the needle tip can be automatically guided to a planar slice of the tissue as it is inserted. Our approach keeps the physician in the loop to control insertion speed. The distance of the needle tip position from the plane of interest is used to drive an observer-based feedback controller that we prove is locally asymptotically stable. Numerical simulations demonstrate a large domain of attraction and robustness of the controller in the face of parametric uncertainty and measurement noise. Physical experiments with tip-steerable nitinol needles inserted into a transparent plastisol tissue phantom under stereo image guidance validate the effectiveness of our approach.      

基于可视化异类特征优选融合的滚动轴承故障诊断

针对故障诊断过程中基于简单的多类故障特征联合决策存在特征集维数多、数据冗余、故障识别率不高的缺点,提出了一种基于异类特征优选融合的故障诊断方法。该方法根据多类特征数据的轮廓图,分析各维特征数据的聚类特性,去除聚类性弱、对故障区分无益的冗余特征维度,仅保留聚类性强的特征维度用于故障识别。在轴承故障诊断实验中,选用故障信号时域统计量和小波包能量两类多维特征进行优选融合,并采用反向传播（BP）神经网络进行故障模式识别。故障识别率达到100%,显著高于无特征优选的故障诊断方法。实验结果表明所提出的方法简便易行,可以显著提高故障识别率。