图像处理英文文献翻译英语论文.doc

《图像处理英文文献翻译英语论文.doc》由会员分享，可在线阅读，更多相关《图像处理英文文献翻译英语论文.doc（6页珍藏版）》请在三一办公上搜索。

1、图像处理英文文献翻译-英语论文图像处理英文文献翻译 Divides the shape algorithm the analysis and the designHas studied the minute shape picture analysis process, proposed the reasonable minute shape picture withdrew the flow. to complete using Visual C+ based on the Windows platform minute shape dimension computational proce

2、dure (Fractal Dimension Calculation Program: FDCP) the design, and by the straight line, kantuo, 3 point array, the Koch curve division shape graph carried on to it has demarcated.Divides the shape theory 1 4 the widespread application relies on to is divided the shape picture a minute shape which w

3、ithdraws, processing and in the minute shape dimension computation. reality exists nearly all is the random minute shape, its dimension computation has a gauge minute shape the dimension to be much more complex than, moreover has many different computational methods, they all need to carry on the ma

4、ssive image analysis and the statistical operation. therefore the establishment computer program completes work and so on image analysis, processing and minute shape dimension computation is solution above question effective means. Visual C+ is at present most popular programming language, It has th

5、e object-oriented programming, supports Windows all characteristics, the probability is good and so on merit. this article uses Visual the C+ design based on Windows the platform FDCP procedure.1.minute shape image analysismaterials science domain many processes all have the massive random minute sh

6、apes, like the material break, the liquid metal coagulation, the thin film , picture, SEM and TEM appearance likes and so on) generally all is the simulated signal, must first complete the digitized process, then does other the information to remove, thus withdraws the minute shape graph. obviously,

7、 this is a very complex process. minute shape graph withdraws the flow like chart 1 to show. first obtains the material test specimen the appearance picture, As primitive picture; Using equipment and so on scanner primitive SEM (or TEM and so on) the appearance picture will input in the computer and

8、 by the certain gradation picture document format (for example: TIFF, BMP, PCX) preserves; After removes the noise signal, carries on careful processing to the picture enable the picture to have the appropriate contrast gradient; Carries on the minute shape characteristic under the best contrast gra

9、dient condition withdrawing, obtains the corresponding minute shape picture (or graph). because the Windows operating system is widely used, the BMP picture form obtains nearly possesses based on the Windows platform picture software support, therefore the minute shape image analysis procedure uses

10、the BMP form to carry on the data storage, moreover general picture processing software generally all can complete between the different picture form the transformation. BMP form picture document to be composed by 3 parts: Position chart article article (BITMAPFILEHEADER) construction of data, posit

11、ion chart information (BITMAPINFO) construction of data and position chart array information. analysis position chart article article after position chart information construction of data, can correctly use in the position chart array the information. minute shape image analysis to use Visual with t

12、he minute Uygur computational procedure the C+ language programming, take Windows as the working platform. position chart array information has recorded the picture each picture element value. when the production picture, Windows from the picture left under angle starts by scanline the picture, name

13、ly from to is left right, from under to on, 11 recorded. these bytes the picture picture element value to compose the position chart array information, its recording order like chart 2 showed. the minute shape picture primitive picture generally all is the gradation picture (colored picture may proc

14、ess gradation picture), divided the shape characteristic to withdraw mainly is judges. the different picture 1274图像处理英文文献翻译according to its gradation distribution situation to have the different gradation distribution, possibly appeared 1 or many gradation peak values. basis peak value integer as we

15、ll as the treating processes difference, was allowed to use different method to complete the gradation picture two values processes. picture two values is refers through establishes some threshold value, And take this threshold value as the threshold, has the multi- gradation level the picture inver

16、sion two gradation levels picture (namely black and white chart), namely all picture elements respectively set for black (picture element value are 0) and are white (picture element value are 1). the different threshold value hypothesis method have formed the different two values treating processes,

17、 commonly used two values processing had following several methods: (1) fixed threshold value law Regarding gradation picture collection F, is smaller than it was equal to (or is bigger than was equal to) some threshold value theta picture element F i j sets is 0; Is bigger than (or is smaller than)

18、 theta the picture element value all to install in parts is 1; Among them, the different picture is different according to the gradation distribution peak value may choose the different theta value. (2) the double fixed threshold value law Has the double peak value regarding the gradation distributi

19、on picture collection F, may establish 2 different threshold values theta 1 and theta 2 (theta 12 .divides the shape dimension computational procedure to design FDCP the design to need to involve to the big memory allocation, the regression analysis, the multi- windows, data sharing and the position

20、 operation and so on many contents, therefore uses the object-oriented programming (OOP) Visual C+ is the relational schematic drawing. host which divides between the shape dimension computational procedure various functions module controls the module control entire procedure as the FDCP programming

21、 language. chart 4 the window frame to renovate, the news transmission and processing as well as each function sub- module linking up with the sub- window user area; The image display module responsibly completes in the sub- window user area demonstrated divides the shape picture as well as the comp

22、uted result; The computation module includes 4 different computational methods the the sub- module to halve the shape picture with different method the dimension to carry on the computation; The data exchange module mainly uses in with other Windows procedure or software carries on data exchange. th

23、is procedure to use 2 different method, namely uses the clipboard and the WMF picture document and the data file. use visible programming language Visual C+ basic kind of storehouse (Microsoft Foundation Class: MFC) carries on the programming, may fully use the C+ language object-oriented programmin

24、g the merit, can enhance the code to carry out the efficiency, guaranteed the procedure is easy to expand, the maintenance, the transplant, the code compact, the efficiency higher merit, simultaneously fully used the Windows platform to support the multi- duties the characteristic, was easy to reali

25、ze the different application procedure data exchange. other also to be allowed according to the need, to calculated the module to make the corresponding supplement, caused to increase the new computational method is easy. procedure to provide the extremely friendly user contact surface. chart 5 is F

26、DCP had demonstrated 5 level of Koch curve main window contact surface. it had File, View, Fractal, Window and Help and so on the menu as well as respective under pull the menu composition. among, the image display module uses MFC and the SDK mix programming, the procedure frame completes using MFC.

27、 but because MFC brings the CBitmap kind of demonstration speed is slower, for the acceleration image display process, uses MFC the frame to transfer the SDK function to realize the demonstration function. because this procedure needs to support the many pictures document simultaneous working, there

28、fore the FDCP procedure must use the MDI window form to carry on the design. for the confirmation procedure reliability, Separately used the straight line, the square, the Koch curve, kantuo 3 point array this 4 kinds to have the gauge minute shape graphthe computed result to carry on the confirmati

29、on to FDCP. among, straight line, square and kantuo 3 point array divided the shape dimension computed result and the theory value strict is equal. calculates the 5th level of Koch curve minute shape dimension with the FDCP box math module, result like chart 6 has shown. may obtain from chart 6, FDC

30、P the computation Koch tune line segment shape dimension was 1.2166. Koch curve theory minute shape dimension is DKoch = ln4/ln3 = 1.2619, but because the infinite level Koch curve was cannot result in, Therefore demarcates uses the 5th level of Koch curve, the computed result indicated FDCP used in

31、 to calculate divides Uygur is the reliable. literature 5 8 successfully the contact surface minute shape dimension which gradient function material existed to Mo/ beta -Sialon and Ta/ beta in the -Sialon has conducted the research with this article procedure. math has calculated chart with the FDCP

32、 procedure box 3 (c) the crystal dimension is 1.694; Chart 3 (d) a jinbian dimension is 1.388, result like chart 7 shows. thus it can be seen, the different research object constitutes the minute shape dimension is obviously different, therefore, must correctly select the research to divide the shap

33、e object and the choice appropriate computational method .3. conclusionunion image analysis and the processing technology, use visible programming language Visual C+ designed (FDCP). the procedure had easy based on the Windows platform minute shape dimension computational procedure to maintain, Migh

34、t the expansion good, the contact surface friendly, the time efficiency and the spatial efficiency high, data sharing is good and so on the merit. with the straight line, the square, the Koch curve, kantuo, 3 point array and so on has the gauge minute shape graph to carry on to FDCP has demarcated,

35、finally indicated FDCP used in to calculate the minute shape dimension is reliable. 图像处理英文文献翻译 分形算法的分析与设计研究了分形图像的分析过程，提出合理的分形图像提取流程利用Visual C+完成了基于Windows平台的分形维数计算程序(Fractal Dimension Calculation Program: FDCP)的设计，并以直线、康托尔三分点集、科赫曲线等分形图形对其进行了标定分形理论14的广泛应用依赖于分形图像的提取、处理与分形维数的计算现实中存在的分形几乎都是无规分形，其维数的计算比有

36、规分形的维数复杂得多，而且有许多不同的计算方法，它们都需要进行大量的图像分析与统计运算因此编制计算机程序来完成图像分析、处理与分形维数计算等工作是解决上述问题的有效办法Visual C+是目前最为流行的程序设计语言，它具有面向对象编程、支持Windows所有特性、可移植性好等优点本文采用Visual C+设计基于Windows平台的FDCP程序1分形图像分析材料科学领域的许多过程都存在着大量的无规分形，如材料的断裂，液态金属的凝固，薄膜的生长，材料的氧化、烧结与侵蚀等但材料的初始图像(金相照片、岩相照片、SEM与TEM的形貌像等)一般都是模拟信号，必须首先完成数字化过程，然后将其他干挠信息去除

37、，从而提取出分形图形显然，这是一个很复杂的过程分形图形提取流程如图1所示首先获得材料试样的形貌像，作为原始图像；利用扫描仪等设备将原始的SEM(或TEM等)形貌像输入计算机中并以一定的灰度图像文件格式(如：TIFF，BMP，PCX)保存；去除噪音信号后，对图像进行细致处理使图像具有合适的对比度；在最佳的对比度状态下进行分形特征的提取，得到相应的分形图像(或图形)由于Windows操作系统被广泛使用，BMP图像格式得到几乎所有基于Windows平台图像软件的支持，因此分形图像分析程序采用BMP格式进行数据存贮，而且通用的图像处理软件一般都能完成不同图像格式之间的转换BMP格式的图像文件由3个部分

38、组成：位图文件头(BITMAPFILEHEADER)数据结构、位图信息(BITMAPINFO)数据结构和位图阵列信息分析位图文件头与位图信息数据结构之后，才能正确地利用位图阵列中的信息分形图像分析与分维计算程序采用Visual C+语言编程，以Windows为工作平台位图阵列信息记录了图像的每一个像素值在生成图像时，Windows从图像的左下角开始逐行扫描图像，即从左到右、从下到上，将图像的像素值一一记录下来这些字节组成了位图阵列信息，其记录顺序如图2所示分形图像的原始图像一般都是灰度图像(彩色图像可处理成灰度图像)，分形特征提取主要是根据其灰度分布情况来判断不同的图像具有不同的灰度分布，可能

39、出现1个或多个灰度峰值根据峰值个数以及处理过程的不同，可以用不同的方法完成灰度图像的二值化过程图像的二值化是指通过设定某个阈值，并以该阈值为门限，把具有多灰度级的图像变换成二灰度级图像(即黑白图)，即所有的像素分别置为黑(像素值为0)和白(像素值为1)不同的阈值设定方法形成了不同的二值化处理过程，常用的二值化处理有以下几种方法：(1)固定阈值法. 对于灰度图像集F，将其小于等于(或大于等于)某阈值的像素Fij置为0；大于(或小于)的像素值全部置为1；其中，不同的图像根据灰度分布峰值的不同可以选择不同的值(2)双固定阈值法. 对于灰度分布存在双峰值的图像集F，可以设定2个不同的阈值1与2(12)

40、当某个像素Fij的像素值小于等于1或大于等于2时，将其置为0；当1Fij2时，将其置为1，反之亦然(3)各自像素阈值法. 对于灰度图像集F，设置图像数组同样大小的阈值数组，用阈值ij对像素Fij进行二值化，若Fijij，则Fij0；若Fijij，则Fij=1图3(a)是单晶Si在一定条件下，其氧化膜生成的枝状晶的SEM形貌像；3(b)是其灰度分布曲线；3(c)是二值化后的黑白图像；3(d)是在此基础上建立的边界不同的研究领域对分形特征的要求是不一样的如果研究的对象是枝状晶，则图3(c)可以作为分形图像；如果研究的对象是枝晶的边界或其他类海岸线，则需要建立其边界，即图3(d)将二值化图像的各种滤

41、波(空间邻域处理方法)结合模糊数学，以保证精确建立图像的边界，如颗粒材料的边界、晶粒边界、氧化界面、扩散界面等，然后可以对建立的边界进行各种测量，得到需要的信息2分形维数计算程序设计 FDCP的设计需要涉及到大内存分配、回归分析、多窗口、数据共享与位运算等诸多内容，因此采用面向对象编程(OOP)的Visual C+作为FDCP的程序设计语言图4是分形维数计算程序各功能模块之间的关系示意图主控模块控制整个程序的窗口框架与子窗口的用户区的刷新、消息传递与处理以及各个功能子模块的衔接；图像显示模块负责完成在子窗口用户区内显示分形图像以及计算结果；计算模块又包括4个不同计算方法的功能子模块，它们分别是

42、周长面积法、盒计数法、半径法与变量法，这些计算子模块用不同的方法对分形图像的维数进行计算；数据交换模块主要用于与其他Windows程序或软件进行数据交换本程序采用2种不同的方法，即利用剪贴板和WMF图片文件与数据文件利用可视化编程语言Visual C+的基本类库(Microsoft Foundation Class: MFC)进行编程，可以充分利用C+语言面向对象编程的优点，能够提高代码的执行效率，保证程序易于扩充、维护、移植，代码紧凑，效率高等优点，同时又充分利用了Windows平台支持多任务的特点，容易实现不同应用程序间的数据交换另外还可以根据需要，对计算模块作相应的补充，使得增加新的计算

43、方法简单易行程序提供非常友好的用户界面图5是FDCP显示了5级科赫曲线的主窗口界面它有“File”,“View”,“Fractal”,“Window”与“Help”等菜单以及各自的下拉菜单组成其中，图像显示模块采用MFC与SDK混合编程，程序框架利用MFC完成但由于MFC所带的CBitmap类的显示速度较慢，为加速图像显示过程，采用MFC的框架调用SDK函数来实现显示功能由于该程序需要支持多个图像文件同时处理，因此FDCP程序必须采用MDI窗口格式进行设计为验证程序的可靠性，分别用直线、正方形、科赫曲线、康托尔三分点集这4种有规分形图形对FDCP的计算结果进行了验证其中，直线、正方形与康托尔三

44、分点集的分形维数计算结果与理论值严格相等用FDCP的盒计数法模块计算了第5级科赫曲线的分形维数，结果如图6所示从图6可以得出，FDCP计算的科赫曲线分形维数为1.2166科赫曲线的理论分形维数为DKochln4/ln31.2619，而由于无穷级的科赫曲线是不可得的，因此标定采用第5级科赫曲线，计算结果表明FDCP用于计算分维是可靠的文献58用本文程序成功地对Mo/-Sialon与Ta/-Sialon梯度功能材料中存在的界面分形维数进行了研究用FDCP程序的盒计数法计算了图3(c)的枝晶的维数为1.694；图3(d)的枝晶边界维数为1.388，结果如图7所示由此可见，不同的研究对象所构成的分形维数是显然不同的，因此，必须正确选取研究的分形对象与选择合适的计算方法3结论结合图像分析与处理技术，利用可视化编程语言Visual C+设计了基于Windows平台的分形维数计算程序(FDCP)程序具有易于维护，可扩充性好、界面友好、时间效率与空间效率高、数据共享性好等优点用直线、正方形、科赫曲线、康托尔三分点集等有规分形图形对FDCP进行了标定，结果表明FDCP用于计算分形维数是可靠的