粗糙表面亚像素级精度实时测量系统
详细信息    本馆镜像全文|  推荐本文 |  |   获取CNKI官网全文
摘要
课题的提出
    极片浆料涂敷技术是锂离子电池研制和生产中的关键技术之一。电池极片涂敷层的厚度及均匀性决定着电池质量的好坏。极片涂敷层厚度及均匀性测量是锂离子电池生产工艺中不可缺少的重要工序,也是保证产品质量提高生产效率的重要手段。极片涂敷层表面颗粒噪声较强。在颗粒粒度大于测量精度要求的情况下,实现在线实时高精度厚度及均匀性测量是本文研究的主要内容。
    二. 课题的难点
    课题的任务是实现对具有粗糙表面的极片涂敷层进行厚度及厚度均匀性在线实时测量。测量的精度要求为5。课题研究的难点主要在精度和实时性两个方面。这两个方面往往是相互矛盾的,要达到高的测量精度,图像处理算法会比较复杂,实时性很难得到保证。我们需要在保证高测量精度的前提下完成实时测量。研究的难点主要有以下几点:
    1. 被测物表面质量情况不好,表面的颗粒可以达到十几微米,物体表面颗粒造成的图像散斑噪声比较严重,完成对粗糙表面的高精度测量具有一定的难度。
    2. 在高测量精度要求的前提下实现实时测量,图像实时采集、处理系统实时性实现比较困难。高的测量精度必然会使图像处理算法比较复杂,如何实时完成处理算法是研究的一个难点。
    3. 由于是对粗糙表面进行高精度测量,如何评价测量结果,并通过试验优化测量系统的参数也是课题研究的一个难点。
     测量系统的总体设计
    测量系统的功能是完成对粗糙表面厚度和厚度均匀性的实时在线测量。
    我们设计的测量系统由图(一)表示。测量系统包括激光三角法测量装置、图像实时处理系统和数据显示部分。
    激光三角法测量装置包括:激光器、光学放大部分。
    图像实时处理系统主要有CCD、A/D转换、图像高速处理硬件和协调逻辑单元组成。
    数据显示部分主要完成对测量结果数据进行显示的功能。
    
    
    
    
    
    
    
    
    
    
    
    
    
    图一 测量系统框图
    四. 测量原理的研究
    根据课题要求和课题的实际情况,本文提出了线边缘激光三角法对具有粗糙表面的极片涂敷层进行厚度及厚度均匀性测量。课题要求的测量精度为5,像素级测量很难满足要求。本文使用线边缘激光三角法实现了对粗糙表面的亚像素级测量。
    通过对被测物表面情况的分析和光斑重心法产生误差机理的研究,发现光斑重心法存在着测量的局限性,难以对图像颗粒噪声严重的物体表面进行高精度测量,并且不能测量物体表面厚度均匀性情况。本文研究的线边缘法克服了光斑重心法的缺点,能够在颗粒噪声较大的情况下达到较高的测量精度。
    线边缘法的关键是准确求出激光边缘的位置。由于被测物表面粗糙,表面的颗粒较大,被测物表面图像有大量的随机散斑噪声。微分算子容易受到噪声的干扰,难以准确的检测到边缘线的真实位置,不适合粗糙表面的高精度测量。本文采用改进的阈值法边缘检测及形态滤波法对被测物表面图像进行处理,能够得到较满意的图像边缘。计算结果可以满足电池极片测量精度要求,实现粗糙表面高精度厚度测量。
    本文使用线边缘最小二乘法拟和直线的参数定量分析被测物表面情况。通过边缘拟和直线的斜率和总变异平方和能够较好的实现粗糙表面凸凹性及厚度不均匀性等的测量。
    本文研究的测量系统对粗糙表面进行高精度测量。由于图像颗粒噪声严
    
    
    重,对各种算法的测量结果需要进行评价。本文使用了非正态分布总体参数的置信区间对测量结果的可信度进行评价。
    五. 实时图像采集处理系统的研究
    测量系统有实时性的要求,就要对被测物体表面图像进行实时采集和处理。一般的图像采集处理系统采用PC机做为核心处理单元,由于图像处理需要大量的时间和内存,也有使用高性能的工作站和小型机来完成这一工作,前者构造的系统实时性不好,后者构造的系统造价高、系统复杂、体积庞大。本文提出的图像实时采集处理系统,以DSP(TMS320C6201)为核心处理器组成在线实时图像处理系统。
    系统A/D转换部分有SAA7111实现,控制部分由FPGA实现,DSP响应中断实现数据的转移、存储和处理。采用FPGA+DSP实现视频信号数据采集和处理,提高系统性能,同时具有适应性与灵活性强,设计、调试方便等优点。最终系统处理速度为每秒处理10帧图像。
    测量系统的试验研究
    本文研究的测量系统是对高粗糙度表面进行高精度实时测量。研究中提出了线边缘法的测量原理,并对依据实际情况提出适合粗糙表面高精度测量的图像算法改进方法和参数优化。研究中还有一些问题需要通过大量的试来进行分析。我们的试验研究部分主要有:
    1. 标定出单像素代表的位移,为亚像素级测量提供前提条件。
    2. 对边缘点坐标数据总体的分布进行了研究,找出适用于高粗糙度表面度测量的精度评价方法。
    3. 对影响测量精度因素进行了分析。对滤波算法提高测量精度的效果进行了研究,形态滤波能够较好的提高测量精度。
    通过课题研究,实现了厚度增量及厚度均匀性的在线实时测量。研究结果表明能够达到测量5的精度要求,实时性可以达到每秒动态测量10次,满足测量的实时性要求。本研究的测量系统具有一定的通用性。
1. INTRODUCTION TO THE PAPER
    The technique of paste coating on battery electrodes is key technique in the development and produce of the li-ion batterys used in the mobile telephone. Thickness and thickness uniformity of coating layer on battery electrodes are decisive factors to the quality of the batterys. The measurement of thickness and thickness uniformity of coating layer on battery electrodes is an important working procedure in the produce of the li-ion batterys used in the mobile telephone and the important means to guarantee the quality of products and improve efficiency. The superficial grain noise of coating layer on battery electrodes is bad. How to accomplish the on-line real time measurement of thickness and thickness uniformity of coating layer on battery electrodes which grain noise granularity is more than the measurement accuracy is main content of the research.
    2. DIFFICULTY TO THE PAPER
    The mission of the subject is to implement the on-line real time measurement of thickness and thickness uniformity of coating layer on battery electrodes with great grain noises. The demand of measurement accuracy is 5 microns. The difficulties of the research include the accomplishment of measurement accuracy and on-line real time measurement. Sometimes two respects are contradictory. It is difficult to accomplish on-line real time measurement, because the image processing algorithm is complex in order to attain the high measurement accuracy. We need to gurrantee the high measurement accuracy and accomplish on-line real time measurement. There are some difficulties in the reseach.
    1. The quality of the surface measured is bad, and the granularity of superficial grain is more than ten microns. The image speckle noise brought by the superficial grain is bad. It is difficult to accomplish on-line real time measurement and attain the high measurement accuracy.
    2. It is difficult to accomplish real time image dada acquisition and processing system, and attain the high measurement accuracy. In order to attain
    
    
    the high measurement accuracy the image processing algorithm is complex. It is difficult to accomplish real time image processing algorithm.
    3. The aim of measurement system in the paper is to accomplish high accuracy measurement of rough surface. It is difficult to evaluate the measurement result and optimize the measurement system parameter by means of the experiments.
    3. THE DESIGN TO MEASUREMENT SYSTEM
    The mission of the subject is to implement the on-line real time measurement of thickness and thickness uniformity of rough surface. Measuremnet system function structure is shown by figure (1). The measurement system consists in laser trianglulation measurement unit, real time image acquisition and processing unit and measurement result display unit.
    The laser trianglulation measurement unit consists in laser and optics enlargement unit.
    The real time image acquisition and processing unit consists in CCD, video signal analog-digital convert unit, image processing hardware unit and harmonization logic unit.
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    Figure(1): Measuremnet system function structure
    
    
    4. RESEARCH ON MEASUREMENT PRINCIPLE
    On the basis of the actual situation of the subject and mesuremnet demand, the measurement method of thickness and thickness uniformity of coating layer on battery electrodes using line edge structured laser trianglulation measurement equipment is presented in the paper. The measurement accuracy requireed is 5 microns. Results of experiments show that the measurement system attained subpixel measurement.
    A detailed analysis is made on quality of the surface and a research is made on error mechanism result from the measurement using the center of gravity method. We find that the measurement method of thickness and thickness uniformity of rough surface using the center of gravity is limited and unfit for the high precision measurement of rough surface. The center of gravity method is unfit for the measurement of thickness uniformity of rough surface. The line ed
引文
[1].K.R.Castleman,数字图像处理,电子工业出版社,1998
    [2].章毓晋,图像工程上册:图像处理和分析,清华大学出版社,1999
    [3].章毓晋,图像工程上册:图像理解与计算机视觉,清华大学出版社,1999
    [4].崔岂,图像处理与分析数学形态学方法及应用,科学出版社,2000
    [5].TMS320C62x/C67x CPU and Instruction Set User’s Guide,Texas Instuments Incorporated,1998
    [6]. TMS320C62x/C67x Peripherals Reference Guide,Texas Instuments Incorporated,1998
    [7].李方慧等,TMS320C6000系列DSPs原理与应用,电子工业出版社,2003
    [8].章毓晋,图像分割,科学出版社,1999
    [9].何立民,I2C总线应用系统设计,北京航空航天大学出版社,1995
    [10].闫德勤,视频多媒体硬件技术与软件编程,学苑出版社,1995
    [11].R.J.奥芬,图像的并行处理技术,科学出版社,1989
    [12].杨枝灵,Visual C++ 数字图像获取 处理及实践应用,人民邮电出版社,2003
    [13].Friedrich M.Wahl,数字图象信号处理,上海远东出版社,1993
    [14].刘榴娣,实用数字图像处理,北京理工大学出版社,1998
    [15].李忠范,数理统计与随机过程,吉林大学出版社,2000
    [16].容观澳,计算机图像处理,清华大学出版社,2000
    [17].王法,C语言图像处理程序设计, 中国科技大学出版社,1993
    [18].黄战华、蔡怀宇,三角法激光测量系统的误差分析及消除方法, 光电工程,VOL.29,No.3,2002
    [19].石成英,CCD激光微位移测量系统的测量头设计,激光杂志,VOL.29,No.3,2002
    [20].王彦,一种提高CCD探测灵敏度的方法,光电工程,VOL.27,No.6,2000
    [21].姚新程,一种实现CCD亚像元位移分辨率的新方法,仪器仪表学报,Vol.23,No.1,2002
    [22].王学忠,对软材料厚度测量的探究,第八界全国光电技术与系统学术会议论文专辑
    [23].张文伟,激光三角位移传感器光斑自检法研究,光电子·激光,Vol.10,
    
    
    No.3,1999
    [24].陈晓东,利用重心法求光斑信号位置的误差分析,光学技术,Vol.26,No.1,Jan,2001
    [25].吴晓波,图像测量系统中的误差分析及提高测量精度的途径,光学精密工程,Vol.5,No.1,Feb,1997
    [26].周会成,用激光线光源实现快速测量,计量技术,No.7,1998
    [27].黄战华,三角法激光测量系统的误差分析及消除方法,光电工程,Vol.29,No.3,June,2002
    [28].徐盛,TMS320C6201 数字信号处理器在图像处理中的应用,集成电路应用,No.1,1999
    [29].朱雷,基于TMS320C6201EVM板开发图像处理系统,重庆大学学报,Vol.26,No.8,Aug,2003
    [30].屈玉福,视觉检测系统中亚像素边缘检测技术的对比研究, 仪器仪表 学报,Vol.24,No.4,2003
    [31].侯涛,基于矩匹配算法的CT 图像亚像素级精度测量方法的研究,计算机测量与控制,No.7,11,2003
    [32].刘志敏,基于数学形态学的图像形态滤波,红外与激光工程,Vol.28,
    No.4,1999
    [33].史伟,图像处理在物体表面洁净度检测中的应用,实用测试技术,No.
    5,2002
    [34].Hecht-Nielsen R,Theory of the Backpropagation Neural Network, International Joint Conference on Neural Networks,New York:IEEE TAB Neural Network Committee,1989,vol.1:593-606
    [35].Yang Daosheng, Cheng Jihong, Zhou Huichen, etal. New Algorithm to Calculate the Center of Laser Reflectios. Proc. San Jose:SPIE,1998, 3306:54-58
    [36].X.Li and M.T.Ochard ,Edge-directed prediction for lossless compression of natural images, IEEE TRANSACTION ON IMAGE PROCESSING, Vol.10,No.6, PP.813-817,June,2001
    [37].M.Gharavi and Alkhansari, A Fast Globally Optimal Algorithm for Template Mathing Using Low-Resolution Prunning,IEEE TRANSACTION ON IMAGE PROCESSING VOL.10,NO.4,PP.526-533,April 2001
    [38].Tian Q.Hulms M N.Algorithms for Subpixel Registration.Computer Vision .Graphics .and Image Processing. 1986.35:222-223.
    
    [39]. Robert. M. Simmons, Jeffrey T.Finer, Steven Chu, Spudich. Quantitative Measurements of Force and Displacement using an Optical Trap. Biophysical journal, 1996,1(4):1813-1822.
    [40].D.Ozdemir and L.Akarun, Fuzzy algorithms for combined quantization and dithering, IEEE TRANSACTION ON IMAGE PROCESSING, Vol.10,No.6, PP.923-931,June,2001
    [41]. West G A.Clack T A. A survey and Examination of Subpixel Measurement Techniques. Close-Range Photogrammetry Meets Machine Vision. USA: SPIE . 1990 . 456-462.
    [42]. Fillard J P.M timet H. Lussert J M .etal . Computer simulation of super resolution point source image detection .Opt Eng. 1993.32(11): 2936-2944
    [43]. Peters W H.Ranson W F. Digital imaging technique in experimental stress analysis. Opt Eng. 1982.21(3): 427-431.
    [44]. Francois Blais. Practical Consideratin for a Design of a High precision 3-D Laser Scanner System. Proc SPIE,1988, 959:225-245
    [45]. E JL. Design of Optical triangulation derices. Optics &laser technology, 1989,21:(5):335-338.
    [46]. Am ir M odjarrad. Non-contact Measurement Using a Laser Scanning Probe.Proc SPIE,1988,1012:229-239.
    [47]. A Ishijima,II Tanaka,T Yanagida. Simultaneous Measurement of Individual ATPase and Mechanical Reactions by a Single Myosin Molecule at Work. Optical Review,1999,6(1):16-23.
    [48]. Fan YY,Huynh V M.Investigation of light scattering from rough periodic surfaces-Numerical solutions . Optics & Laser Technology ,1992,24(3):145-150.
    [49]. Brian F.Alexander,etal.Elimination of Systenatic Error in Subpixel Accuracy Centroid Estimation.Opt.Eng,1991,130(9):1320-1331.
    [50]. DataSheet of SAA7111-Video Input Processor(VIP).Philips Company,1999.
    [51].C6201 Data Sheet ,TEXAS Instruments.
    [52].Zhou P.Goodson K E.Subpixel dispalacement and deformation gradient measuremnet using digital image speckle correlation .Opt Eng.2001. 40(8) :1613-1620.

© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号

地址:北京市海淀区学院路29号 邮编:100083

电话:办公室:(+86 10)66554848;文献借阅、咨询服务、科技查新:66554700