外文翻译---基于模糊逻辑技术图像上边缘检测
自动化专业外文翻译---模糊逻辑控制机器人走迷宫
外文资料FUZZY LOGIC CONTROL FOR ROBOT MAZE TRA VERSAL: ANUNDERGRADUATE CASE STUDYJames Wolfer Chad A. GeorgeAbstractAs previously reported, Indiana University South Bend has deployed autonomous robots in their Computer Organization course to facilitate introducing computer science students to the basics of logic, embedded systems, and assembly language. The robots help to provide effective, real-time feedback on program operation and to make assembly language less abstract. As a part of their coursework students are required to program a sensor-based traversal of a maze. This paper details one solution to this problem employing a fuzzy logic controller to create linguistic rules.Key words:Fuzzy logic, pedagogy, robots, student projectsINTRODUCTIONAssembly language programming in a computer science environment is often taught using abstract exercises to illustrate concepts and encourage student proficiency.To augment this approach we have elected to provide hands-on, real-world experience to our students by introducing robots into our assembly language class.Observing the physical action of robots can generate valuable feedback and have real-world consequences – robots hitting walls make students instantly aware of program errors, for example.It also provides insight into the realities of physical machines such as motor control, sensor calibration, and noise. To help provide a meaningful experience for our computer organization students, we reviewed the course with the following objectives in mind:• Expand the experience of our students in a manner that enhances the student's insight, provides a hands-on, visual, environment for them to learn, and forms an integrated component for future classes.•Remove some of the abstraction inherent in the assembly language class. Specifically, to help enhance the error detection environment.• Provide a kinesthetic aspect to our pedagogy.• Build student expertise early in their program that could lead to research projects and advanced classroom activities later in their program. Specifically, in this case, to build expertise to support later coursework in intelligent systems and robotics.As one component in meeting these objectives we, in cooperation with the Computer Science department, the Intelligent Systems Laboratory, and the University Center for Excellence in Teaching, designed a robotics laboratory to support the assembly language portion of the computer organization class as described in [1].The balance of this report describes one example project resulting from this environment. Specifically, we describe the results of a student project developing an assembly language fuzzy engine, membership function creation, fuzzy controller, and resulting robot behavior in a Linux-based environment.We also describe subsequent software devlopment in C# under Windows, including graphical membership tuning, real-time display of sensor activation, and fuzzy controller system response. Collectively these tools allow for robust controller development, assemblylanguage support, and an environment suitable for effective classroom and publicdisplay.BACKGROUNDRobots have long been recognized for their potential educational utility, with examples ranging from abstract, simulated, robots, such as Karel[2] and Turtle[3] for teaching programming and geometry respectively, to competitive events such as robotic soccer tournaments[4].As the cost of robotics hardware has decreased their migration into the classroom has accelerated [5, 6]. Driven by the combined goals for this class and the future research objectives, as well as software availability, we chose to use off-the-shelf, Khepera II, robots from K-Team[7].SIMULATED ROBOT DIAGRAMThe K-Team Kephera II is a small, two-motor robot which uses differential wheel speed for steering. Figure 1 shows a functional diagram of the robot. In addition to thetwo motors it includes a series of eight infrared sensors, six along the “front” and two in the “back”of the robot. This robot also comes with an embedded system-call library, a variety of development tools, and the availability of several simulators. The embedded code in the Khepera robots includes a relatively simple, but adequate, command level interface which communicates with the host via a standard serial port. This allows students to write their programs using the host instruction set (Intel Pentium in this case), send commands, and receive responses such as sensor values, motor speed and relative wheel position.We also chose to provide a Linux-based programming environment to our students by adapting and remastering the Knoppix Linux distribution [9]. Our custom distribution supplemented Knoppix with modified simulators for the Khepera, the interface library (including source code),manuals, and assembler documentation. Collectively, this provides a complete development platform.The SIM Kheperasimulator[8] includes source code in C, and provides a workable subset of the native robot command language. It also has the ability to redirect input and output to the physical robot from the graphics display. Figure 2 shows the simulated Khepera robot in a maze environment and Figure 3 shows an actual Khepera in a physical maze. To provide a seamless interface to the simulator and robots we modified the original simulator to more effectively communicate through a pair of Linuxpipes, and we developed a small custom subroutine library callable from the student's assembly language programs.Assignments for the class range from initial C assignments to call the robot routines to assembly language assignments culminating in the robot traversing the maze. FUZZY CONTROLLEROne approach to robot control, fuzzy logic, attempts to encapsulate important aspects of human decision making. By forming a representation tolerant of vague, imprecise, ambiguous, and perhaps missing information fuzzy logic enhances the ability to deal with real-world problems. Furthermore, by empirically modeling a system engineering experience and intuition can be incorporated into a final design.Typical fuzzy controller design [10] consists of:• Defining the control objectives and criteria• Determining the input and output relationships• Creating fuzzy membership functions, along withsubsequent rules, to encapsulate a solution fromintput to output.• Apply necessary input/output conditioning• Test, evaluate, and tune the resulting system.Figure 4 illustrates the conversion from sensor input to a fuzzy-linguistic value. Given three fuzzy possibilities, …too close‟, …too far‟, and …just right‟, along with a sensor reading we can ascertain the degree to which the sensor reading belongs to each of these fuzzy terms. Note that while Figure 4 illustrates a triangular membership set, trapezoids and other shapes are also common.Once the inputs are mapped to their corresponding fuzzy sets the fuzzy attributes are used, expert system style, to trigger rules governing the consequent actions, in this case, of the robot.For example, a series of rules for a robot may include:• If left-sensor is too close and right sensor is too far then turn right.• If left sensor is just right and forward sensor is too far then drive straight.• If left sensor is too far and forward sensor is too far then turn left.• If forward sensor is close then turn right sharply.The logical operators …and‟, …or‟, and …not‟ are calculated as follows: …and‟ represents set intersection and is calculated as the minimum value, …or‟ is calculated as the maximum value or the union of the sets, and …not‟ finds the inverse of the set, calculated as 1.0-fitness.Once inputs have been processed and rules applied, the resulting fuzzy actions must be mapped to real-world control outputs. Figure 5 illustrates this process. Here output is computed as the coordinate of the centroid of the aggregate area of the individual membership sets along the horizontal axis.ASSEMBLY LANGUAGE IMPLEMENTATIONTwo implementations of the fuzzy robot controller were produced. The first was written in assembly language for the Intel cpu architecture under the Linux operating system, the second in C# under Windows to provide a visually intuitive interface for membership set design and public demonstration.Figure 6 shows an excerpt of pseudo-assembly language program. The actual program consists of approximately eight hundred lines of hand-coded assembly language. In the assembly language program subroutine calls are structured with parameters pushed onto the stack. Note that the code for pushing parameters has been edited from this example to conserve space and to illustrate the overall role of the controller. In this code-fragment the …open_pipes‟ routine establishes contact with the simulator or robot. Once communication is established, a continous loop obtains sensor values, encodes them as fuzzy inputs, interprets them through the rule base to linguistic output members which are then converted to control outputs which are sent to the robot. The bulk of the remaining code implements the fuzzy engine itself.FUZZY CONTROLLER MAIN LOOPMembership sets were manually defined to allow the robot to detect and track walls, avoid barriers, and negotiate void spaces in it field of operation. Using this controller, both the simulated robot and the actual Khepera successfully traversed a variety of maze configurations.ASSEMBLY LANGUAGE OBSERV ATIONSWhile implementing the input fuzzification and output defuzzification in assembly language was tedious compared with the same task in a high level language, the logic engine proved to be well suited to description in assembly language.The logic rules were defined in a type of psuedo-code using …and‟, …or‟, …not‟ as operators and using the fuzzy input and output membership sets as parameters. With the addition of input, output and flow control operators, the assembly language logic engine simply had to evaluate these psuedo-code expressions in order to map fuzzy inputs memberships to fuzzy output memberships.Other than storing the current membership fitness values from the inputfuzzyfication, the only data structure needed for the logic engine is a stack to hold intermediate calculations. This is convenient under assembly language since the CPUs stack is immediately available as well as the nescesary stack operators.There were seven commands implemented by the logic rule interpreter: IN, OUT, AND, OR, NOT, DONE, and EXIT.•IN – reads the current fitness from an input membership set and places the value on the stack.•OUT – assigns the value on the top of the stack as the fitness value of an output membership set if it is greater than the existing fitness value for that set.•AND – performs the intersection operation by replacing the top two elements on the stack with the minimum element.•OR – performs the union operation by replace the top two elements on the stack with their maximum.•NOT – replaces the top value on the stack with its compliment.•DONE – pops the top value off the stack to prepare for the next rule•EXIT – signals the end of the logic rule definition and exits the interpreter.As an example the logic rule “If left-sensor is too close and right sensor is too far then turn right”, might be define d by the following fuzzy logic psuedo-code: IN, left_sensor[ TOO_CLOSE ]IN, right_sensor[ TOO_FAR ] ANDOUT, left_wheel[ FWD ]OUT, right_wheel[ STOP ]DONEEXITBy utilizing the existing CPU stack and implementing the logic engine as anpsuedo-code interpreter, the assembly language version is capable of handling arbitrarily complicated fuzzy rules composed of the simple logical operators provided. IMPLEMENTATIONWhile the assembly language programming was the original focus of the project, ultimately we felt that a more polished user interface was desirable for membership set design, fuzzy rule definition, and controller response monitoring. To provide these facilities the fuzzy controller was reimplemented in C# under Windows. through 10 illustrate the capabilities of the resulting software. Specifically, Figure 7 illustrates user interface for membership defination, in this case …near‟. Figure 8 illustrates theinterface for defining the actual fuzzy rules. Figure 9 profiles the output response with respect to a series of simulated inputs. Finally, real-time monitoring of the system is also implemented as illustrated in 10 which shows the robot sensor input values.Since the Khepera simulator was operating system specific, the C# program controls the robot directly. Again, the robot was successful at navigating the maze using a controller specified with this interface.SUMMARYTo summarize, we have developed a student-centric development environment for teaching assembly language programming. As one illustration of its potential we profiled a project implementing a fuzzy-logic engine and controller, along with a subsequent implementation in the C# programming language. Together these projects help to illustrate the viability of a robot-enhanced environment for assembly language programming.REFERENCES[1] Wolfer, J &Rababaah, H. R. A., “Creating a Hands-On Robot Environment for Teaching Assembly Language Programming”, Global Conference on Engineering and Technology Education, 2005[2] Pattic R.E., Karel the Robot: a gentle introduction to the art of programming, 2nd edition. Wiley, 1994[3] Abelson H. and diSessa A., Turtle geometry: the computer as a medium for exploring mathematics. MIT Press, 1996[4] Amirijoo M., Tesanovic A., and Nadjm-Tehrani S., “Raising motivation in real-time laboratories: the soccer scenario” in SIGCSE Technical Symposium on Computer Sciences Education, pp. 265-269, 2004.[5] Epp E.C., “Robot control and embedded systems on inexpensive linux platforms workshop,” in SIGCSE Technical Symposium on Computer Science Education, p. 505, 2004[6] Fagin B. and Merkle L., “Measuring the effectiveness of robots in teaching computer science,” in SIGCSE Technical Symposium on Computer Science Education, PP. 307-311, 2003.[7] K-Team Khepera Robots, , accessed 09/06/05.[8] Michel O., “Khepera Simulator package version 2.0: Freeware mobile robot simulator written at the university of nice Sophia-Antipolis by Olivier Michel. Downloadable from the world wide web. http://diwww.epfl.ch/lami/team/michel/khep-sim, accessed 09/06/05.[9] Knoppix Official Site, , accessed 09/06/05.[10] Earl Cox., The Fuzzy Systems Handbook, Academic Press, New York, 1999.模糊逻辑控制机器人走迷宫James Wolfer Chad A. George摘要美国印第安纳大学南本德已部署在他们的计算机组织课程自主机器人，以方便学生介绍计算机科学逻辑的基础知识，嵌入式系统和汇编语言。
模糊逻辑中英文对照外文翻译文献
模糊逻辑中英文对照外文翻译文献(文档含英文原文和中文翻译)译文:模糊逻辑欢迎进入模糊逻辑的精彩世界，你可以用新科学有力地实现一些东西。
在你的技术与管理技能的领域中，增加了基于模糊逻辑分析和控制的能力，你就可以实现除此之外的其他人与物无法做到的事情。
以下就是模糊逻辑的基础知识：随着系统复杂性的增加，对系统精确的阐述变得越来越难，最终变得无法阐述。
于是，终于到达了一个只有靠人类发明的模糊逻辑才能解决的复杂程度。
模糊逻辑用于系统的分析和控制设计，因为它可以缩短工程发展的时间；有时，在一些高度复杂的系统中，这是唯一可以解决问题的方法。
虽然，我们经常认为控制是和控制一个物理系统有关系的，但是，扎德博士最初设计这个概念的时候本意并非如此。
实际上，模糊逻辑适用于生物，经济，市场营销和其他大而复杂的系统。
模糊这个词最早出现在扎德博士于1962年在一个工程学权威刊物上发表论文中。
1963年，扎德博士成为加州大学伯克利分校电气工程学院院长。
那就意味着达到了电气工程领域的顶尖。
扎德博士认为模糊控制是那时的热点，不是以后的热点，更不应该受到轻视。
目前已经有了成千上万基于模糊逻辑的产品，从聚焦照相机到可以根据衣服脏度自我控制洗涤方式的洗衣机等。
如果你在美国，你会很容易找到基于模糊的系统。
想一想，当通用汽车告诉大众，她生产的汽车其反刹车是根据模糊逻辑而造成的时候，那会对其销售造成多么大的影响。
以下的章节包括：1)介绍处于商业等各个领域的人们他们如果从模糊逻辑演变而来的利益中得到好处，以及帮助大家理解模糊逻辑是怎么工作的。
2)提供模糊逻辑是怎么工作的一种指导，只有人们知道了这一点，才能运用它用于做一些对自己有利的事情。
这本书就是一个指导，因此尽管你不是电气领域的专家，你也可以运用模糊逻辑。
需要指出的是有一些针对模糊逻辑的相反观点和批评。
一个人应该学会观察反面的各个观点，从而得出自己的观点。
我个人认为，身为被表扬以及因写关于模糊逻辑论文而受到赞赏的作者，他会认为，在这个领域中的这种批评有点过激。
edge_detection_边缘检测
边缘检测-edge detection1.问题描述边缘检测是图像处理和计算机视觉中的基本问题，边缘检测的目的是标识数字图像中亮度变化明显的点。
图像属性中的显著变化通常反映了属性的重要事件和变化。
这些包括（i）深度上的不连续、（ii）表面方向不连续、（iii）物质属性变化（iv）场景照明变化。
边缘检测是图像处理和计算机视觉中，尤其是特征提取中的一个研究领域。
边缘检测的评价是指对边缘检测结果或者边缘检测算法的评价。
诚然，不同的实际应用对边缘检测结果的要求存在差异，但大多数因满足以下要求：1）正确检测出边缘2）准确定位边缘3）边缘连续4）单边响应，即检测出的边缘是但像素的2.应用场合图像边缘检测大幅度地减少了数据量，并且剔除了可以认为不相关的信息，保留了图像重要的结构属性。
有许多方法用于边缘检测，它们的绝大部分可以划分为两类：基于查找一类和基于零穿越的一类。
基于查找的方法通过寻找图像一阶导数中的最大和最小值来检测边界，通常是将边界定位在梯度最大的方向。
基于零穿越的方法通过寻找图像二阶导数零穿越来寻找边界，通常是Laplacian过零点或者非线性差分表示的过零点。
3.研究历史和现状边缘检测作为图像处理的一个底层技术，是一个古老又年轻的课题，有着悠久的历史。
早在1959年，B.Julez就提到过边缘检测，随后，L.G.Robert于1965年对边缘检测进行系统的研究。
3.1一阶微分算子一阶微分算子是最原始，最基本的边缘检测方法，它的理论依据是边缘是图像中灰度发生急剧变化的地方，而图像的提督刻画了灰度的变化速率。
因此，通过一阶微分算子可以增强图像中的灰度变化区域，然后对增强的区域进一步判断边缘。
在点（x,y）的梯度为一个矢量，定义为：梯度模值为：梯度方向为：根据以上理论，人们提出了许多算法，经典的有：Robert算子，Sobel算子等等，这些一阶微分算子的区别在于算子梯度的方向，以及在这些方向上用离散化数值逼近连续导数的方式和将这些近似值合成梯度的方式不同。
边缘检测中英文翻译
Digital Image Processing and Edge DetectionDigital Image ProcessingInterest in digital image processing methods stems from two principal application areas: improvement of pictorial information for human interpretation; and processing of image data for storage, transmission, and representation for autonomous machine perception.An image may be defined as a two-dimensional function, f(x,y), where x and y are spatial (plane) coordinates, and the amplitude of f at any pair of coordinates (x, y) is called the intensity or gray level of the image at that point. When x, y, and the amplitude values of f are all finite, discrete quantities, we call the image a digital image. The field of digital image processing refers to processing digital images by means of a digital computer. Note that a digital image is composed of a finite number of elements, each of which has a particular location and value. These elements are referred to as picture elements, image elements, pels, and pixels. Pixel is the term most widely used to denote the elements of a digital image.Vision is the most advanced of our senses, so it is not surprising that images play the single most important role in human perception. However, unlike humans, who are limited to the visual band of the electromagnetic (EM) spectrum, imaging machines cover almost the entire EM spectrum, ranging from gamma to radio waves. They can operate on images generated by sources that humans are not accustomed to associating with images. These include ultrasound, electron microscopy, and computer generated images. Thus, digital image processing encompasses a wide and varied field of applications.There is no general agreement among authors regarding where image processing stops and other related areas, such as image analysis and computer vision, start. Sometimes a distinction is made by defining image processing as a discipline in which both the input and output of a process are images. We believe this to be a limiting and somewhat artificial boundary. For example, under this definition,even the trivial task of computing the average intensity of an image (which yields a single number) would not be considered an image processing operation. On the other hand, there are fields such as computer vision whose ultimate goal is to use computers to emulate human vision, including learning and being able to make inferences and take actions based on visual inputs. This area itself is a branch of artificial intelligence(AI) whose objective is to emulate human intelligence. The field of AI is in its earliest stages of infancy in terms of development, with progress having been much slower than originally anticipated. The area of image analysis (also called image understanding) is in between image processing and computer vision.There are no clearcut boundaries in the continuum from image processing at one end to computer vision at the other. However, one useful paradigm is to consider three types of computerized processes in this continuum: low, mid, and highlevel processes. Low-level processes involve primitive operations such as image preprocessing to reduce noise, contrast enhancement, and image sharpening. A low-level process is characterized by the fact that both its inputs and outputs are images. Mid-level processing on images involves tasks such as segmentation (partitioning an image into regions or objects), description of those objects to reduce them to a form suitable for computer processing, and classification (recognition) of individual objects. A midlevel process is characterized by the fact that its inputs generally are images, but its outputs are attributes extracted from those images (e.g., edges, contours, and the identity of individual objects). Finally, higherlevel processing involves “making sense” of an ensemble of recognize d objects, as in image analysis, and, at the far end of the continuum, performing the cognitive functions normally associated with vision.Based on the preceding comments, we see that a logical place of overlap between image processing and image analysis is the area of recognition of individual regions or objects in an image. Thus, what we call in this book digital image processing encompasses processes whose inputs and outputs are images and, in addition, encompasses processes that extract attributes from images, up to and including the recognition of individual objects. As a simple illustration to clarify these concepts, consider the area of automated analysis of text. The processes of acquiring an image of the area containing the text, preprocessing that image, extracting (segmenting) the individual characters, describing the characters in a form suitable for computer processing, and recognizing those individual characters are in the scope of what we call digital image processing in this book. Making sense of the content of the page may be viewed as being in the domain of image analysis and even computer vision, depending on the level of complexity implied by the statement “making sense.” As will become evident shortly, digital image processing, as we have defined it, is used successfully in a broad range of areas of exceptional social and economic value.The areas of application of digital image processing are so varied that some formof organization is desirable in attempting to capture the breadth of this field. One of the simplest ways to develop a basic understanding of the extent of image processing applications is to categorize images according to their source (e.g., visual, X-ray, and so on). The principal energy source for images in use today is the electromagnetic energy spectrum. Other important sources of energy include acoustic, ultrasonic, and electronic (in the form of electron beams used in electron microscopy). Synthetic images, used for modeling and visualization, are generated by computer. In this section we discuss briefly how images are generated in these various categories and the areas in which they are applied.Images based on radiation from the EM spectrum are the most familiar, especially images in the X-ray and visual bands of the spectrum. Electromagnetic waves can be conceptualized as propagating sinusoidal waves of varying wavelengths, or they can be thought of as a stream of massless particles, each traveling in a wavelike pattern and moving at the speed of light. Each massless particle contains a certain amount (or bundle) of energy. Each bundle of energy is called a photon. If spectral bands are grouped according to energy per photon, we obtain the spectrum shown in fig. below, ranging from gamma rays (highest energy) at one end to radio waves (lowest energy) at the other. The bands are shown shaded to convey the fact that bands of the EM spectrum are not distinct but rather transition smoothly from one to the other.Fig1Image acquisition is the first process. Note that acquisition could be as simple as being given an image that is already in digital form. Generally, the image acquisition stage involves preprocessing, such as scaling.Image enhancement is among the simplest and most appealing areas of digital image processing. Basically, the idea behind enhancement techniques is to bring out detail that is obscured, or simply to highlight certain features of interest in an image.A familiar example of enhancement is when we increase the contrast of an imagebecause “it looks better.” It is important to keep in mind that enhancement is a very subjective area of image processing. Image restoration is an area that also deals with improving the appearance of an image. However, unlike enhancement, which is subjective, image restoration is objective, in the sense that restoration techniques tend to be based on mathematical or probabilistic models of image degradation. Enhancement, on the other hand, is based on human subjective preferences regarding what constitutes a “good” en hancement result.Color image processing is an area that has been gaining in importance because of the significant increase in the use of digital images over the Internet. It covers a number of fundamental concepts in color models and basic color processing in a digital domain. Color is used also in later chapters as the basis for extracting features of interest in an image.Wavelets are the foundation for representing images in various degrees of resolution. In particular, this material is used in this book for image data compression and for pyramidal representation, in which images are subdivided successively into smaller regions.F ig2Compression, as the name implies, deals with techniques for reducing the storage required to save an image, or the bandwidth required to transmi it.Although storagetechnology has improved significantly over the past decade, the same cannot be said for transmission capacity. This is true particularly in uses of the Internet, which are characterized by significant pictorial content. Image compression is familiar (perhaps inadvertently) to most users of computers in the form of image file extensions, such as the jpg file extension used in the JPEG (Joint Photographic Experts Group) image compression standard.Morphological processing deals with tools for extracting image components that are useful in the representation and description of shape. The material in this chapter begins a transition from processes that output images to processes that output image attributes.Segmentation procedures partition an image into its constituent parts or objects. In general, autonomous segmentation is one of the most difficult tasks in digital image processing. A rugged segmentation procedure brings the process a long way toward successful solution of imaging problems that require objects to be identified individually. On the other hand, weak or erratic segmentation algorithms almost always guarantee eventual failure. In general, the more accurate the segmentation, the more likely recognition is to succeed.Representation and description almost always follow the output of a segmentation stage, which usually is raw pixel data, constituting either the boundary of a region (i.e., the set of pixels separating one image region from another) or all the points in the region itself. In either case, converting the data to a form suitable for computer processing is necessary. The first decision that must be made is whether the data should be represented as a boundary or as a complete region. Boundary representation is appropriate when the focus is on external shape characteristics, such as corners and inflections. Regional representation is appropriate when the focus is on internal properties, such as texture or skeletal shape. In some applications, these representations complement each other. Choosing a representation is only part of the solution for transforming raw data into a form suitable for subsequent computer processing. A method must also be specified for describing the data so that features of interest are highlighted. Description, also called feature selection, deals with extracting attributes that result in some quantitative information of interest or are basic for differentiating one class of objects from another.Recognition is the pro cess that assigns a label (e.g., “vehicle”) to an object based on its descriptors. As detailed before, we conclude our coverage of digital imageprocessing with the development of methods for recognition of individual objects.So far we have said nothing about the need for prior knowledge or about the interaction between the knowledge base and the processing modules in Fig2 above. Knowledge about a problem domain is coded into an image processing system in the form of a knowledge database. This knowledge may be as simple as detailing regions of an image where the information of interest is known to be located, thus limiting the search that has to be conducted in seeking that information. The knowledge base also can be quite complex, such as an interrelated list of all major possible defects in a materials inspection problem or an image database containing high-resolution satellite images of a region in connection with change-detection applications. In addition to guiding the operation of each processing module, the knowledge base also controls the interaction between modules. This distinction is made in Fig2 above by the use of double-headed arrows between the processing modules and the knowledge base, as opposed to single-headed arrows linking the processing modules.Edge detectionEdge detection is a terminology in image processing and computer vision, particularly in the areas of feature detection and feature extraction, to refer to algorithms which aim at identifying points in a digital image at which the image brightness changes sharply or more formally has discontinuities.Although point and line detection certainly are important in any discussion on segmentation,edge dectection is by far the most common approach for detecting meaningful discounties in gray level.Although certain literature has considered the detection of ideal step edges, the edges obtained from natural images are usually not at all ideal step edges. Instead they are normally affected by one or several of the following effects:1.focal b lur caused by a finite depth-of-field and finite point spread function; 2.penumbral blur caused by shadows created by light sources of non-zero radius; 3.shading at a smooth object edge; 4.local specularities or interreflections in the vicinity of object edges.A typical edge might for instance be the border between a block of red color and a block of yellow. In contrast a line (as can be extracted by a ridge detector) can be a small number of pixels of a different color on an otherwise unchanging background. For a line, there may therefore usually be one edge on each side of the line.To illustrate why edge detection is not a trivial task, let us consider the problemof detecting edges in the following one-dimensional signal. Here, we may intuitively say that there should be an edge between the 4th and 5th pixels.If the intensity difference were smaller between the 4th and the 5th pixels and if the intensity differences between the adjacent neighbouring pixels were higher, it would not be as easy to say that there should be an edge in the corresponding region. Moreover, one could argue that this case is one in which there are several edges.Hence, to firmly state a specific threshold on how large the intensity change between two neighbouring pixels must be for us to say that there should be an edge between these pixels is not always a simple problem. Indeed, this is one of the reasons why edge detection may be a non-trivial problem unless the objects in the scene are particularly simple and the illumination conditions can be well controlled.There are many methods for edge detection, but most of them can be grouped into two categories,search-based and zero-crossing based. The search-based methods detect edges by first computing a measure of edge strength, usually a first-order derivative expression such as the gradient magnitude, and then searching for local directional maxima of the gradient magnitude using a computed estimate of the local orientation of the edge, usually the gradient direction. The zero-crossing based methods search for zero crossings in a second-order derivative expression computed from the image in order to find edges, usually the zero-crossings of the Laplacian or the zero-crossings of a non-linear differential expression, as will be described in the section on differential edge detection following below. As a pre-processing step to edge detection, a smoothing stage, typically Gaussian smoothing, is almost always applied (see also noise reduction).The edge detection methods that have been published mainly differ in the types of smoothing filters that are applied and the way the measures of edge strength are computed. As many edge detection methods rely on the computation of image gradients, they also differ in the types of filters used for computing gradient estimates in the x- and y-directions.Once we have computed a measure of edge strength (typically the gradient magnitude), the next stage is to apply a threshold, to decide whether edges are present or not at an image point. The lower the threshold, the more edges will be detected, and the result will be increasingly susceptible to noise, and also to picking outirrelevant features from the image. Conversely a high threshold may miss subtle edges, or result in fragmented edges.If the edge thresholding is applied to just the gradient magnitude image, the resulting edges will in general be thick and some type of edge thinning post-processing is necessary. For edges detected with non-maximum suppression however, the edge curves are thin by definition and the edge pixels can be linked into edge polygon by an edge linking (edge tracking) procedure. On a discrete grid, the non-maximum suppression stage can be implemented by estimating the gradient direction using first-order derivatives, then rounding off the gradient direction to multiples of 45 degrees, and finally comparing the values of the gradient magnitude in the estimated gradient direction.A commonly used approach to handle the problem of appropriate thresholds for thresholding is by using thresholding with hysteresis. This method uses multiple thresholds to find edges. We begin by using the upper threshold to find the start of an edge. Once we have a start point, we then trace the path of the edge through the image pixel by pixel, marking an edge whenever we are above the lower threshold. We stop marking our edge only when the value falls below our lower threshold. This approach makes the assumption that edges are likely to be in continuous curves, and allows us to follow a faint section of an edge we have previously seen, without meaning that every noisy pixel in the image is marked down as an edge. Still, however, we have the problem of choosing appropriate thresholding parameters, and suitable thresholding values may vary over the image.Some edge-detection operators are instead based upon second-order derivatives of the intensity. This essentially captures the rate of change in the intensity gradient. Thus, in the ideal continuous case, detection of zero-crossings in the second derivative captures local maxima in the gradient.We can come to a conclusion that,to be classified as a meaningful edge point,the transition in gray level associated with that point has to be significantly stronger than the background at that point.Since we are dealing with local computations,the method of choice to determine whether a value is “significant” or not id to use a threshold.Thus we define a point in an image as being as being an edge point if its two-dimensional first-order derivative is greater than a specified criterion of connectedness is by definition an edge.The term edge segment generally is used if the edge is short in relation to the dimensions of the image.A key problem insegmentation is to assemble edge segments into longer edges.An alternate definition if we elect to use the second-derivative is simply to define the edge ponits in an image as the zero crossings of its second derivative.The definition of an edge in this case is the same as above.It is important to note that these definitions do not guarantee success in finding edge in an image.They simply give us a formalism to look for them.First-order derivatives in an image are computed using the gradient.Second-order derivatives are obtained using the Laplacian.数字图像处理与边缘检测数字图像处理数字图像处理方法的研究源于两个主要应用领域：其一是改进图像信息以便于人们分析；其二是为使机器自动理解而对图像数据进行存储、传输及显示。
sobel算子边缘检测原理
sobel算子边缘检测原理Sobel算子是一种常用于边缘检测的算子。
它采用了离散微分算子的方法，通过计算像素点与其周围像素点的灰度差异来检测边缘。
边缘是图像中明显的灰度变化的地方，是图像中物体间的分界线。
边缘检测的目的就是找到图像中的这些边缘。
Sobel算子是基于图像的灰度梯度的计算来实现边缘检测的。
在图像中，像素点处的灰度值代表了其周围像素的强度值。
梯度是指一个函数在其中一点的变化率。
在图像处理中，梯度可以指的是图像灰度值的变化率。
Sobel算子通过计算像素点的灰度梯度来检测边缘。
Sobel算子的原理是通过对图像进行两次卷积操作来计算梯度。
一次卷积操作用于在水平方向上计算梯度，另一次卷积操作用于在垂直方向上计算梯度。
对于一个图像中的像素点A，它的灰度梯度可以通过以下公式计算得到：G = abs(Gx) + abs(Gy)其中，G是像素点A的灰度梯度，Gx是像素点A在水平方向上的梯度，Gy是像素点A在垂直方向上的梯度。
Sobel算子采用了以下两个3×3模板来进行卷积操作：水平方向上的Sobel算子模板：[-101-202-101]垂直方向上的Sobel算子模板：[-1-2-1000121]在进行卷积操作时，将模板分别与图像中的像素点进行对应位置上的乘法运算，并将结果相加得到像素点的梯度值。
这样就可以得到整个图像的灰度梯度图像。
通过计算像素点的灰度梯度，我们可以找到图像中的边缘。
边缘通常具有较大的梯度值，因为边缘上存在明显的灰度变化。
因此，我们可以通过设定一个阈值来筛选出图像中的边缘。
Sobel算子在实际应用中有一些优点。
首先，它是一种简单而高效的边缘检测方法。
其次，Sobel算子可以用来检测水平和垂直方向上的边缘，因此可以检测到更多的边缘信息。
此外，Sobel算子还可以通过调整模板的尺寸来适应不同大小图像的边缘检测需求。
然而，Sobel算子也存在一些缺点。
首先，Sobel算子对噪声比较敏感，可能会在噪声处产生较大的边缘响应。
基于Sober算子的图像边缘检测
1 引言图像边缘是一种重要的视觉信息，图像边缘检测是图像处理、图像分析、模式识别、计算机视觉以及人类视觉的基本步骤。
其结果的正确性和可靠性将直接影响到机器视觉系统对客观世界的理解。
实现边缘检测有很多不同的方法，也一直是图像处理中的研究热点，人们期望找到一种抗噪强、定位准、不漏检、不误检的检测算法。
经典的算法中主要用梯度算子，最简单的梯度算子是Roberts算子，比较常用的有Prewitt算子和Sober算子，其中Sober算子效果较好，但是经典Sober算子也存在不足，其边缘具有很强的方向性，只对垂直与水平方向敏感，其他方向不敏感，这就使得那些边缘检测不到j。
对后续的图像处理有很的影响。
本文在此基础提出了一种新的算法，该算子该算法提高了传统Sober检测算子的性能，具有良好的检测精度。
对数字图像{f(x,y)}的每个像素，考察它上下左右邻点灰度的加权差，与之接近的邻点的权大。
据此，定义Sober算子如下：其卷积算子：适当取门限TH，作如下判断：s(i,j)>TH，(i,j)为阶跃状边缘点，{s(i,j)}为边缘图像。
Sober算子很容易在空问上实现，Sober边缘检测器不但产生较好的边缘检测效果，而且受噪声影响也比较小。
当使用大的邻域时，抗噪性能会更好，但这样会增加计算量，并且得出的边缘也会相应变粗。
Sober算子利用像素点上下、左右邻点的灰度加权算法，根据在边缘点处达到极值这一现象进行边缘的检测。
Sober算子对噪声具有平滑作用，提供较为精确的边缘方向信息，是一种较为常用的边缘检测方法。
2 算法设计针对经典Sober算子对边缘具有很强的方向性特点，提出了一种在Sober算子上改进的算法，其主要思想是先对图像进行全局阈值的分割处理，因为分割后的图像是二值图像，此时进行边缘提取，这就可以使各个方向的边缘都可以检测到。
但也可能会丢失原本可直接用Sober算子检测到的边缘。
因此，用处理后所得的图像与用Sober算子直接对原始图像进行边缘检测的图像相加，这一步就得尤为重要。
机械毕业设计英文外文翻译235基于模糊逻辑的AMT离合器结合控制研究
附录A 外文文献Study of Controlling Clutch Engagement forAMT Based on Fuzzy LogicTAN G Xia-qing , HOU Chao-zhen , CHEN Yun-chuangAbstract: The control of the clutch engagement for an automatic mechanical transmission in the process of a tracklayer getting to start is studied. The dynamic model of power transmission and automatic clutch system is developed. Using tools of Simulink , the transient characteristics during the vehicle starting , including the jerk and the clutch slip time , are provided here. Based on the analyses of the simulation results and driver’s experiences , a fuzzy controller is designed to control the clutch engagement . Simulation results verify its value.Key words : clutch ; automatic transmission ; fuzzy controlThe automatic mechanical transmission (AMT) has several advantages , such as simpleness , higher efficiency and lower costs. But these benefits come from settling a series of challenging control problem. For example , it is difficult andcomplex for an AMT to properly control the clutch engagement while the vehicle starting , because different drivers have different intentions ( for example smooth start and fast start) , the second cause is that the control goals of lengthening the clutch life and smoothly starting vehicle are contradictory. So it is an important research field for AMT , and some researchers are studying this problem too.The focus of this paper is to study the control of clutch engagement for AMT while the tracklayer starts. In most cases , experiment methods are used to improve starting quality , however , they require much effort and time to develop a new control algorithm and to investigate the effect of this design. On the contrary , a simulation method has the merit of saving money and time , and overcomes the restrictions of experimental conditions. The fuzzy controller is designed based on the analysis of the simulation results and dr iver’s experience.The organization of the paper is as follows. First , the system model is described , along with some simulation results. Secondly , the fuzzy control strategy of clutch engagement is developed. Finally ,conclusions from this work , as well as recommendation for future work , are also outlined.The control goal of the clutch control system is to ensure the vehicle starting according to driver’s intention and make the clutch engage smoothly and the jerk as small as possible. Based on the analysis of the simulation results and driver’s experience , we have the following conclusions.①The accelerator pedal βindicates driver’s intention and his judgement on the environment and vehicle’s states. The larger βis , the higher the engaging speed vshould be.com②The engine rotational speed ωe indicates its carrying capacity. The larger ωe is , the stronger the carrying capacity is.③The r is the speed ratio between the passive and the active departments of the clutch which can be expressed by r =ωc/ωe. It indicates the slip state of the clutch. The larger r is ,should be. Consequently , the higher the engaging speed vcomthe control strategy is expressed as follows :①Regulate the engine rotational speed ωes according to the signal of the accelerator pedal βbefore engaging the clutch. The larger βis , the higher ωes should be.is ②While engaging the clutch , the engaging speed vcom decided by the accelerator pedal β, the engine rotationalspeed ωe and the speed ratio r by using fuzzy logic.③Regulate the throttle opening as the driver regulates the accelerator pedal β.The fuzzy logic approach is used here to control the engaging speed of the clutch .The engine rotational speed ωe and engaging speed of the clutch vcom are normalized. Following this method , the i-th control rule can be= Dj . written as Ri : If β= Aj and r = Bj and ωe = Ci then vcomHere , A j is the fuzzy set of the accelerator pedal β, Bj is the fuzzy set of the engine rotational speed ωe , Cj is the fuzzy set of the speed ratio r and Dj is the fuzzy set of the clutch.engaging speed vcomConclusionThe key point of the vehicle start is to accomplish the driver’s intention and ensure the vehicle starting smoothly. The model results can be used to study the clutch engagement . To overcome the difficulty of clutch control , a fuzzy control strategy is proposed based on the states of accelerator pedal β, engine rational speed ωand speed ratio r. Simulationeindicates it is valuable. The future work is to optimize the parameters of the membership function in experiment and test its effects.附录B 外文文献翻译基于模糊逻辑的AMT离合器结合控制研究汤霞清, 侯朝桢, 陈云窗摘要:研究具有机械式自动变速器的履带式车辆起步时离合器结合控制问题。
机器视觉英文词汇
机器视觉英文词汇机器视觉英文词汇Aaberration 像差accessory shoes 附件插座、热靴accessory 附件achromatic 消色差的active 主动的、有源的acutance 锐度acute-matte 磨砂毛玻璃adapter 适配器advance system 输片系统ae lock(ael) 自动曝光锁定af illuminatoraf 照明器af spotbeam projectoraf 照明器af(auto focus) 自动聚焦algebraic operation 代数运算一种图像处理运算，包括两幅图像对应像素的和、差、积、商。
aliasing 走样（混叠）当图像象素间距和图像细节相比太大时产生的一种人工痕迹。
alkaline 碱性ambient light 环境光amplification factor 放大倍率analog input/output boards 模拟输入输出板卡analog-to-digital converters 模数转换器ancillary devices 辅助产品angle finder 弯角取景器angle of view 视角anti-red-eye 防红眼aperture priority(ap) 光圈优先aperture 光圈apo(apochromat) 复消色差application-development software 应用开发软件application-specific software 应用软件apz(advanced program zoom) 高级程序变焦arc 弧图的一部分；表示一曲线一段的相连的像素集合。
area ccd solid-state sensors 区域ccd 固体传感器area cmos sensors 区域cmos传感器area-array cameras 面阵相机arrays 阵列asa(american standards association) 美国标准协会asics 专用集成电路astigmatism 像散attached coprocessrs 附加协处理器auto bracket 自动包围auto composition 自动构图auto exposure bracketing 自动包围曝光auto exposure 自动曝光auto film advance 自动进片auto flash 自动闪光auto loading 自动装片auto multi-program 自动多程序auto rewind 自动退片auto wind 自动卷片auto zoom 自动变焦autofocus optics 自动聚焦光学元件automatic exposure(ae) 自动曝光automation/robotics 自动化/机器人技术automation 自动化auxiliary 辅助的Bback light compensation 逆光补偿back light 逆光、背光back 机背background 背景backlighting devices 背光源backplanes 底板balance contrast 反差平衡bar code system 条形码系统barcode scanners 条形码扫描仪barrel distortion 桶形畸变base-stored image sensor (basis) 基存储影像传感器battery check 电池检测battery holder 电池手柄bayonet 卡口beam profilers 电子束仿形器beam splitters 光分路器bellows 皮腔binary image 二值图像只有两级灰度的数字图像（通常为0和1，黑和白）biometrics systems 生物测量系统blue filter 蓝色滤光镜blur 模糊由于散焦、低通滤波、摄像机运动等引起的图像清晰度的下降。
基于模糊神经网络的边缘检测方法
时， 其他节 点受 到抑制 ， 而对 该类模 式不敏 感而难 从
以获胜 。当有 其他 类 模 式输 入 时 ， 这些 节 点 再参 与
方法检 测效 果也 不 很 好 ， 主要 因为 这 些 方 法缺
方法及 Ｐｅ ｉ 方 法 在 好 图像 和 被 噪 声 污 染 过 的 图 ｒｗｔ ｔ
像 中进行 了 比较 ， 实验结果 表 明 ， 们 的边 缘 检测效 我
果 要 优 于 用 Ｃ ｎ ｙ方 法 和 Ｐ ｅ ｉ 方 法 获 得 的边 缘 ａｎ ｒｗｔ ｔ
检测效 果 。
作 者 简 介 张建岭（ ９ ５ ） 男 ， １６ 一 ， 河北人 ， 讲师 。
・
５１ ・
张 建岭 王 慧 张 民 ：基 于模 糊神 经 网络 的 边 缘 检 测 方 法
１ 像 素 边 缘 分 类 及 神 经 网 络
对 于输人 图像 中的 每 一个 非边 界 像 素来 说 ， 我 们定 义其 在 ３×３邻 域 上 的 ４维 特 征 矢 量 为 ＝
少 二 维 结 构 知 识 ， 用 单 一 结 构 处 理 各 种 不 同 的 边 仅 缘 类 型 。 由 于 噪 声 及 其 他 因 素 的 影 响 ， 于 许 多 复 对
图像 的 边缘 含 有 图像 中最 重要 的信 息 ， 用 以 如
杂 的实际 图像 ， 有 上 述 边 缘 检 测方 法 都 会 产 生一 所 些 不尽 人 意的结 果 ， 如假边 缘 或漏检 边缘 等 。
所 示 ， 中类 型 １在 １方 向 上 的 灰 度 差 值 和 低 ， 其 在
说， 中心像 素 Ｐ 与其邻 域间 的双 向灰 度差 值 和分 别 用 ｄ ， ｄ ， 示 ， 图 １ 。ｄ ， ｄ表 如 所示 ， 其计算 公式 如下 ，
Sobel边缘检测
Sobel边缘检测Sobel算⼦：[-1 0 1-2 0 2-1 0 1]⽤此算⼦与原图像做卷积，可以检测出垂直⽅向的边缘。
算⼦作⽤在图像的第⼆列，结果是：200,200,200；作⽤在第三列，结果是：200,200,200；边缘 – 是像素值发⽣跃迁的地⽅（变化率最⼤处，导数最⼤处），是图像的显著特征之⼀，在图像特征提取、对象检测、模式识别等⽅⾯都有重要的作⽤。
Sobel算⼦和Scharr算⼦（1）Sobel算⼦：是离散微分算⼦（discrete differentiation operator），⽤来计算图像灰度的近似梯度，梯度越⼤越有可能是边缘。
Soble算⼦的功能集合了⾼斯平滑和微分求导，⼜被称为⼀阶微分算⼦，求导算⼦，在⽔平和垂直两个⽅向上求导，得到的是图像在X⽅法与Y⽅向梯度图像。
缺点：⽐较敏感，容易受影响，要通过⾼斯模糊（平滑）来降噪。
算⼦是通过权重不同来扩⼤差异。
梯度计算：（在两个⽅向求导，假设被作⽤图像为 I）⽔平变化: 将 I 与⼀个奇数⼤⼩的内核 Gx进⾏卷积。
⽐如，当内核⼤⼩为3时, Gx的计算结果为:垂直变化: 将 I 与⼀个奇数⼤⼩的内核 Gy进⾏卷积。
⽐如，当内核⼤⼩为3时, Gy的计算结果为:在图像的每⼀点，结合以上两个结果求出近似梯度：有时也⽤下⾯更简单公式代替，计算速度快：（最终图像梯度）。
（2）Scharr：当内核⼤⼩为3时, 以上Sobel内核可能产⽣⽐较明显的误差(毕竟，Sobel算⼦只是求取了导数的近似值)。
为解决这⼀问题，OpenCV提供了 Scharr 函数，但该函数仅作⽤于⼤⼩为3的内核。
该函数的运算与Sobel函数⼀样快，但结果却更加精确，不怕⼲扰，其内核为:（3）Sobel/Scharr提取边缘（求导）步骤：1）⾼斯模糊平滑降噪：GaussianBlur( src, dst, Size(3,3), 0, 0, BORDER_DEFAULT ); 2）转灰度：cvtColor( src, gray, COLOR_RGB2GRAY );3）求X和Y⽅向的梯度（求导）：Sobel(gray_src, xgrad, CV_16S, 1, 0, 3);Sobel(gray_src, ygrad, CV_16S, 0, 1, 3);Scharr(gray_src, xgrad, CV_16S, 1, 0);Scharr(gray_src, ygrad, CV_16S, 0, 1);4）像素取绝对值：convertScaleAbs(A, B); //计算图像A的像素绝对值，输出到图像B5）相加X和Y，得到综合梯度，称为振幅图像：addWeighted( A, 0.5,B, 0.5, 0, AB); //混合权重相加，效果较差或者循环获取像素，每个点直接相加，效果更好。
基于模糊元胞自动机的图像边缘检测方法
Ｅｄ ｅ Ｄｅ ｅ ｔｏ ｆＩ ａ ｅ ｓ ｄ ｏ ｚ ｙ Ｃｅｌ ｌｒ Ａｕ ｏ ａ ａ ｇ ｔｃ ｉｎ ｏ ｍ ｇ ｓ Ｂａ ｅ ｎ Ｆｕ ｚ ｌｕ ａ ｔ ｍ ｔ
ＺＨＡＮＧ Ｋｅ， ａ ｉ －ｈ ＹＡＮＧ ｅ ｍｉｇ Ｙｕ ｎ Ｊｎ ｓ ａ， ＸｕБайду номын сангаас— ｎ
维普资讯
２０ ０ ８年 ９月
第２ ２卷 第 ３期 总 ７ 期 ３
北 京 联合 大 学 学 报 （ 自然 科 学 版 ） Ｊｕｎｌ ｆ ｅ ｉ ｎ ｎＵ ｉ ｒ ｔ Ｎ ｔｒ ｃ ｎ ｅ） ｏ ｒａ ｏ Ｂ ｉｎ Ｕ ｉ ｎ ｅ ｉ （ ａ ａ Ｓｉ ｃ ｓ ｊｇ ｏ ｖ ｓｙ ｕｌ ｅ
Ｓ ｐ．２ ０８ ｅ ０ Ｖ０ ．２ Ｎｏ． ｕ Ｎｏ． ３ １２ ３ Ｓ ｍ ７
基 于模 糊 元 胞 自动 机 的 图像 边缘 检 测方 法
张 珂 ， 津 莎 ， 薛 明 苑 杨
０ １０ ） ７０３ （ 北 电力 大 学 电子 与通 信 工 程 系 ， 北 保 定 华 河
［ 摘 要 ］ 针 对基 于元胞 自动 机 图像 边缘 检 测 的 原 有 算 法 ， 出 了新 的 改进 算 法 。该 算 法采 用 提 基 于 方向信 息测 度与 边缘 有序 性度 量 的 多信 息 融合 方 法 ， 用模 糊 逻 辑 对 特征 信 息进 行 模 糊 推 利
［ 键词 ］ 边缘检 测 ； 胞 自动机 ； 向信 息测度 ； 缘有 序性 度量 ； 糊度 量 ； 模糊 化 关 元 方 边 模 反 ［ 中图 分类 号］ Ｔ ９ ．１ Ｐ３ １４ ［ 献标识 码 ］ Ａ 文 ［ 章编 号 ］ １０ ．３０ ２ ０ ）３０ ４ —６ 文 ０ ５０ １ （０ ８ ０ ．０ ９０
Ａｂｓｒ ｃ ：Ａ ｅ ｉ ｒｖ ｄ ｅ ｇ ｅｅ ｔｏ ｌｏｉｈ ｏ ｍａ ｅ ａ ｅ ｎ ｃ ｌ ｌｒａ ｔｍａａ ｉ ｐ ｅ ｅ ｔｄ．Ｔ ｉ ｔｏ Ｓ ｔａ ｔ ｎ ｗ ｍｐ ｏ ｅ ｄ ｅｄ ｔｃｉｎ ａ ｇ ｒｔ ｍ ｆｉ ｇ ｓｂ ｓ ｄ ｏ ｅｌ ａ ｕ ｏ ｔ ｓ ｒ ｓ ｎｅ ｕ ｈ ｓｍｅｈ ｄ Ｕ － ｅ ｉｅ ｔｏ ｎ ｏｍａｉｎ ｍｅ ｓ ｒ ｎ ｄ ｅ ｏｄ ｒｍｅ ｓ ｒ ｓ ｅ ｇ ｈ ｒｃｅ ｓｉ ｎ ｏ ａｉｎ， ｕ ｅ ｕ ｚ ｏ ｉ ｏ ｉｆｒ ｓｄｒ ｃｉｎ ｉｆｒ ｔ ａ ｕ ｅ ａ ｄ ｅ ｇ ｒ ｅ ａｕ ｅ ａ ｄ ｅ ｃ ａ ａ ｔ ｒ ｔｃ ｉ ｆｒ ｔｏ ｏ ｉ ｍ ｓ ｓ ｆ ｚｙ ｌｇｃ ｔ ｎｅ
sobel边缘检测原理
sobel边缘检测原理Sobel边缘检测原理Sobel边缘检测是一种常用的图像处理技术，它可以用来检测图像中的边缘。
Sobel算子是一种离散微分算子，它可以将图像中的每个像素点与其周围的像素点进行卷积运算，从而得到该像素点的梯度值。
Sobel算子可以分为水平和垂直两个方向，分别用于检测图像中的水平和垂直边缘。
Sobel算子的原理是基于图像中的灰度变化来检测边缘。
在图像中，边缘处的灰度值会发生明显的变化，而非边缘处的灰度值则相对平滑。
因此，通过计算像素点周围的灰度值差异，可以得到该像素点的梯度值，从而判断该点是否为边缘点。
Sobel算子的计算公式如下：Gx = [-1 0 1; -2 0 2; -1 0 1] * AGy = [-1 -2 -1; 0 0 0; 1 2 1] * A其中，Gx和Gy分别表示水平和垂直方向的梯度值，A表示原始图像的像素矩阵。
在计算过程中，先将原始图像进行灰度化处理，然后对每个像素点进行卷积运算，得到该点的梯度值。
最后，将水平和垂直方向的梯度值进行平方和开方运算，得到该像素点的总梯度值。
Sobel算子的优点是计算简单、速度快，可以有效地检测图像中的边缘。
但是，它也存在一些缺点，比如对噪声比较敏感，容易产生误检测。
因此，在实际应用中，需要结合其他的图像处理技术来进行优化和改进。
总之，Sobel边缘检测是一种简单而有效的图像处理技术，可以用来检测图像中的边缘。
它的原理是基于图像中的灰度变化来进行计算，可以通过卷积运算得到每个像素点的梯度值。
虽然Sobel算子存在一些缺点，但是在实际应用中仍然具有广泛的应用价值。
基于sobel、canny的边缘检测实现参考模板
基于sobel 、canny 的边缘检测实现一．实验原理Sobel 的原理：索贝尔算子（Sobel operator ）是图像处理中的算子之一，主要用作边缘检测。
在技术上，它是一离散性差分算子，用来运算图像亮度函数的梯度之近似值。
在图像的任何一点使用此算子，将会产生对应的梯度矢量或是其法矢量.该算子包含两组3x3的矩阵，分别为横向及纵向，将之与图像作平面卷积，即可分别得出横向及纵向的亮度差分近似值。
如果以A 代表原始图像，Gx 及Gy 分别代表经横向及纵向边缘检测的图像，其公式如下:101202*101x G A -+⎛⎫ ⎪=-+ ⎪ ⎪-+⎝⎭121000*121y G A +++⎛⎫ ⎪= ⎪ ⎪---⎝⎭图像的每一个像素的横向及纵向梯度近似值可用以下的公式结合，来计算梯度的大小。
在以上例子中，如果以上的角度Θ等于零，即代表图像该处拥有纵向边缘，左方较右方暗。
在边沿检测中，常用的一种模板是Sobel 算子。
Sobel 算子有两个，一个是检测水平边沿的 ；另一个是检测垂直平边沿的 。
与 和 相比，Sobel 算子对于象素的位置的影响做了加权，因此效果更好。
Sobel 算子另一种形式是各向同性Sobel(Isotropic Sobel)算子，也有两个，一个是检测水平边沿的 ，另一个是检测垂直平边沿的 。
各向同性Sobel 算子和普通Sobel 算子相比，它的位置加权系数更为准确，在检测不同方向的边沿时梯度的幅度一致。
由于建筑物图像的特殊性，我们可以发现，处理该类型图像轮廓时，并不需要对梯度方向进行运算，所以程序并没有给出各向同性Sobel 算子的处理方法。
由于Sobel 算子是滤波算子的形式，用于提取边缘，可以利用快速卷积函数， 简单有效，因此应用广泛。
美中不足的是，Sobel 算子并没有将图像的主体与背景严格地区分开来，换言之就是Sobel 算子没有基于图像灰度进行处理，由于Sobel 算子没有严格地模拟人的视觉生理特征，所以提取的图像轮廓有时并不能令人满意。
图像边缘检测算法英文文献翻译中英文翻译
image edge examination algorithmAbstractDigital image processing took a relative quite young discipline, is following the computer technology rapid development, day by day obtains the widespread edge took the image one kind of basic characteristic, in the pattern recognition, the image division, the image intensification as well as the image compression and so on in the domain has a more widespread edge detection method many and varied, in which based on brightness algorithm, is studies the time to be most long, the theory develops the maturest method, it mainly is through some difference operator, calculates its gradient based on image brightness the change, thus examines the edge, mainly has Robert, Laplacian, Sobel, Canny, operators and so on LOG。
First as a whole introduced digital image processing and the edge detection survey, has enumerated several kind of at present commonly used edge detection technology and the algorithm, and selects two kinds to use Visual the C language programming realization, through withdraws the image result to two algorithms the comparison, the research discusses their good and bad points.ForewordIn image processing, as a basic characteristic, the edge of the image, which is widely used in the recognition, segmentation,intensification and compress of the image, is often applied to high-level are many kinds of ways to detect the edge. Anyway, there are two main techniques: one is classic method based on the gray grade of every pixel; the other one is based on wavelet and its multi-scale characteristic. The first method, which is got the longest research,get the edge according to the variety of the pixel gray. The main techniques are Robert, Laplace, Sobel, Canny and LOG algorithm.The second method, which is based on wavelet transform, utilizes the Lipschitz exponent characterization of the noise and singular signal and then achieve the goal of removing noise and distilling the real edge lines. In recent years, a new kind of detection method, which based on the phase information of the pixel, is developed. We need hypothesize nothing about images in advance. The edge is easy to find in frequency domain. It’s a reliable method.In chapter one, we give an overview of the image edge. And in chapter two, some classic detection algorithms are introduced. The cause of positional error is analyzed, and then discussed a more precision method in edge orientation. In chapter three, wavelet theory is introduced. The detection methods based on sampling wavelet transform, which can extract maim edge of the image effectively, and non-sampling wavelet transform, which can remain the optimum spatial information, are recommended respectively. In the last chapter of this thesis, the algorithm based on phase information is introduced. Using the log Gabor wavelet, two-dimension filter is constructed, many kinds of edges are detected, including Mach Band, which indicates it is a outstanding and bio-simulation method。
毕业设计(论文)外文翻译-基于数据挖掘的直销电子商务平台会员奖励管理系统开发-洪维坤【范本模板】
毕业设计(论文）外文资料翻译系部：计算机科学与技术系专业：计算机科学与技术姓名：洪维坤学号: 0807012215外文出处：Proceeding of Workshop on the （用外文写)of Artificial,Hualien,TaiWan,2005 指导老师评语:签名：年月日不确定性数据挖掘：一种新的研究方向Michael Chau1， Reynold Cheng2， and Ben Kao31：商学院，香港大学,薄扶林，香港2:计算机系，香港理工大学九龙湖校区,香港3：计算机科学系，香港大学，薄扶林，香港摘要由于不精确测量、过时的来源或抽样误差等原因，数据不确定性常常出现在真实世界应用中。
目前，在数据库数据不确定性处理领域中，很多研究结果已经被发表。
我们认为，当不确定性数据被执行数据挖掘时，数据不确定性不得不被考虑在内，才能获得高质量的数据挖掘结果.我们称之为“不确定性数据挖掘”问题。
在本文中，我们为这个领域可能的研究方向提出一个框架.同时,我们以UK-means聚类算法为例来阐明传统K—means算法怎么被改进来处理数据挖掘中的数据不确定性。
1.引言由于测量不精确、抽样误差、过时数据来源或其他等原因，数据往往带有不确定性性质。
特别在需要与物理环境交互的应用中，如:移动定位服务［15]和传感器监测［3]。
例如：在追踪移动目标(如车辆或人）的情境中，数据库是不可能完全追踪到所有目标在所有瞬间的准确位置.因此,每个目标的位置的变化过程是伴有不确定性的。
为了提供准确地查询和挖掘结果，这些导致数据不确定性的多方面来源不得不被考虑。
在最近几年里，已有在数据库中不确定性数据管理方面的大量研究，如：数据库中不确定性的表现和不确定性数据查询。
然而，很少有研究成果能够解决不确定性数据挖掘的问题。
我们注意到,不确定性使数据值不再具有原子性。
对于使用传统数据挖掘技术,不确定性数据不得不被归纳为原子性数值。
洗衣机设计中英文对照外文翻译文献
10中英文对照外文翻译洗衣机设计方案1. 1. 引言引言本报告讨论了提高标准容量的家用洗衣机的能源效率的可行的设计方案。
在本设计方案分析中所使用的程序是基于由能源部（本设计方案分析中所使用的程序是基于由能源部（DOE DOE DOE）解释的规定，该规定概）解释的规定，该规定概述了工艺改进措施。
述了工艺改进措施。
根据合理的规则，根据合理的规则，根据合理的规则，不同的电器排名后，不同的电器排名后，不同的电器排名后，确定了潜在的设计方确定了潜在的设计方案，这在电器效率标准颁布的过程中又进了一步。
然后用筛选因子来确定是否应进一步考虑排除某一设计方案。
筛选设计方案中使用的很多输入程序都是由洗衣机制造商提供。
其他来源于贸易刊物和家用电器制造商协会（机制造商提供。
其他来源于贸易刊物和家用电器制造商协会（AHAM AHAM AHAM）。
）。
虽然（在新工艺中）现有分析的设计方案阶段在规章制定预告（虽然（在新工艺中）现有分析的设计方案阶段在规章制定预告（ANOPR ANOPR ANOPR）之）之前，与原ANOPR[2]ANOPR[2]相对应，但是以前从厂家收集到的数据，现有的相对应，但是以前从厂家收集到的数据，现有的AHAM 的输入和其他利益相关者在这份报告中仍然会被考虑。
美国能源部打算在未来发出补充的法规制定提案预告。
2. 2. 产品分类产品分类对消费者有使用价值的家电被纳入产品类的分析的行列。
类别是家电类型的一个子集，例如，洗衣机是一种家电，但是紧凑型洗衣机就是一种产品类别，家电产品根据不同的能源利用率被分成不同的类型。
能源部划分类别根据产品的容量或其他相关的性能特点，量或其他相关的性能特点，如产品的实用性和工作效率。
如产品的实用性和工作效率。
如产品的实用性和工作效率。
一般情况下，一般情况下，一般情况下，类的定义类的定义是通过从家电制造商，贸易协会和其他的相关的会议研讨中搜集的数据来做出划分。
那些已被指定的没有通过能源部测试程序的类的不再进一步分析。
sobel、prewitt、roberts边缘检测方法的原理
sobel、prewitt、roberts边缘检测方法的原理边缘检测是一种图像处理技术，它可以识别图像中的结构和边界，为后续图像处理操作提供依据。
边缘检测技术主要有Sobel、Prewitt和Roberts三种。
本文将介绍这三种边缘检测方法的原理以及它们之间的区别。
Sobel边缘检测是由Ivan E.Sobel于1960年研发的一种边缘检测技术，它是根据图像中的灰度值变化来计算出一个像素的梯度，从而检测出图像的边缘。
Sobel算子是一种以一阶微分运算为基础的滤波算子，它采用一种双线性结构，可以检测图像中横向、竖向、水平和垂直等多种边缘。
Sobel算子能够有效地检测出图像中的轮廓线，并降低噪声的影响。
Prewitt边缘检测也是基于一阶微分运算，它是由JohnG.Prewitt于1970年研发的一种滤波算子。
它可以植入到一个3×3的矩阵中，将每个像素点处的灰度值变化量进行累加，从而检测出图像中的边缘。
Prewitt边缘检测的优点是能够获得图像中的更多细节，而且对噪声具有较强的抗干扰能力。
Roberts边缘检测也是由一阶微分运算为基础，是由Larry Roberts于1966年研发的一种边缘检测技术。
它采用3×3的矩阵，把相邻的像素点的灰度值变化量进行累加，以检测出图像的边缘，它同样也能够获得更多的细节，并且对噪声也有较强的抗干扰能力。
总结起来，Sobel、Prewitt和Roberts三种边缘检测方法都是基于一阶微分运算，它们的算法类似，从某种程度上来说，它们都是拿某一个像素点处的灰度值变化量与其周围像素点的灰度值变化量进行累加比较，来检测出图像中的边缘。
但是它们在具体运用算子上还是略有不同，Sobel算子采用双线性结构，能够检测图像中横向、竖向、水平和垂直等多种边缘；而Prewitt和Roberts边缘检测方法的算法都是采用一个3×3的矩阵，将相邻的像素点的灰度值变化量累加，从而检测出边缘。
边缘检测的名词解释
边缘检测的名词解释边缘检测是计算机视觉领域中一项重要的图像处理技术，其目的是识别和提取图像中各个物体或场景的边缘信息。
边缘是指图像中颜色或亮度发生明显变化的地方，它标志着物体之间的分界线或者物体与背景之间的过渡区域。
边缘检测能够帮助我们理解图像中的结构，更好地分析图像内容并进行后续的图像处理和分析。
在计算机视觉应用中，边缘检测有着广泛的应用。
例如在目标识别中，边缘检测可以帮助我们找到物体的轮廓，从而进行物体的识别和分类。
在图像分割方面，边缘检测可以用来分割图像中的不同区域，提取感兴趣的物体。
此外，边缘检测还可以用于图像增强、图像压缩等领域。
常用的边缘检测算法包括Sobel算子、Laplacian算子、Canny算子等。
这些算法基于图像的灰度值和亮度变化来检测边缘。
Sobel算子通过计算图像中每个像素点的梯度幅值来确定边缘的位置和方向。
Laplacian算子则通过计算像素值的二阶导数来检测边缘。
而Canny算子则是一种综合性的边缘检测算法，它综合了Sobel 算子和Laplacian算子的优点，在性能上更加稳定和准确。
边缘检测并不是一项简单的任务，它受到噪声、光照变化、图像分辨率等因素的影响。
因此，在进行边缘检测前，通常需要进行预处理，比如图像平滑、灰度化等步骤，以减少这些干扰因素对边缘检测结果的影响。
边缘检测并非完美，它仍然存在一些问题和挑战。
例如，边缘检测往往会产生一些不连续和不完整的边缘，这需要通过进一步的处理和分析来解决。
此外，在图像中存在复杂的背景和纹理时，边缘检测的准确性也会受到影响。
因此，为了获得更好的边缘检测效果，我们需要结合其他的图像处理和分析技术，如图像分割、特征提取等。
总结起来，边缘检测是计算机视觉中一项重要的图像处理技术，其通过识别和提取图像中的边缘信息来帮助我们理解图像结构、进行目标识别和图像分割等应用。
虽然边缘检测还存在一些问题和挑战，但随着技术的不断进步和研究的不断深入，相信边缘检测在图像处理领域将发挥更大的作用。
laplacian边缘检测应用场景
laplacian边缘检测应用场景
Laplacian边缘检测是一种常用的图像处理技术，它可以检测图像中的边缘并将其突显出来。
以下是一些Laplacian边缘检测的应用场景：
1. 图像识别和目标检测：Laplacian边缘检测可用于检测图像中的物体边缘，从而帮助计算机进行图像识别和目标检测。
2. 医学图像处理：Laplacian边缘检测可用于医学图像处理，例如在X射线和CT扫描中高亮显示骨骼结构。
3. 安全监控：Laplacian边缘检测可用于安全监控系统中，检测人员或车辆的运动轮廓，并进行监控。
4. 视频编辑：Laplacian边缘检测可以在视频编辑中用于改善视频质量，例如去除视频中不需要的背景信息。
5. 特效处理：Laplacian边缘检测可用于图像或视频特效处理，例如增强图像或视频中的边缘轮廓，使其更加明显。
- 1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。
译文二：1基于模糊逻辑技术图像上边缘检测[2]摘要：模糊技术是经营者为了模拟在数学水平的代偿行为过程的决策或主观评价而引入的。
下面介绍经营商已经完成了的计算机视觉应用。
本文提出了一种基于模糊逻辑推理战略为基础的新方法，它被建议使用在没有确定阈值的数字图像边缘检测上。
这种方法首先将用3⨯3的浮点二进制矩阵将图像分割成几个区域。
边缘像素被映射到一个属性值与彼此不同的范围。
该方法的鲁棒性所得到的不同拍摄图像将与线性Sobel运算所得到的图像相比较。
并且该方法给出了直线的线条平滑度、平直度和弧形线条的良好弧度这些永久的效果。
同时角位可以更清晰并且可以更容易的定义。
关键词：模糊逻辑，边缘检测，图像处理，电脑视觉，机械的部位，测量1.引言在过去的几十年里，对计算机视觉系统的兴趣，研究和发展已经增长了不少。
如今，它们出现在各个生活领域，从停车场，街道和商场各角落的监测系统到主要食品生产的分类和质量控制系统。
因此，引进自动化的视觉检测和测量系统是有必要的，特别是二维机械对象[1,8]。
部分原因是由于那些每天产生的数字图像大幅度的增加(比如，从X光片到卫星影像)，并且对于这样图片的自动处理有增加的需求[9,10,11]。
因此，现在的许多应用例如对医学图像进行计算机辅助诊断，将遥感图像分割和分类成土地类别(比如，对麦田，非法大麻种植园的鉴定，以及对作物生长的估计判断)，光学字符识别，闭环控制，基于目录检索的多媒体应用，电影产业上的图像处理，汽车车牌的详细记录的鉴定，以及许多工业检测任务(比如，纺织品，钢材，平板玻璃等的缺陷检测)。
历史上的许多数据已经被生成图像，以帮助人们分析(相比较于数字表之类的，图像显然容易理解多了)[12]。
所以这鼓励了数字分析技术在数据处理方面的使用。
此外，由于人类善于理解图像，基于图像的分析法在算法发展上提供了一些帮助(比如，它鼓励几何分析)，并且也有助于非正式确认的结果。
虽然计算机视觉可以被总结为一个自动(或半自动)分析图像的系统，一些变化也是可能的[9,13]。
这些图像可以来自超出正常灰度和色彩的照片，例如红外光，X射线，以及新一代的高光谱[1]Abdallah A. Alshennawy, A yman A. Aly. Edge Detection in Digital Images Using Fuzzy LogicTechnique[]J. World Academy of Science, Engineering and Technology, 2009, 51:178-186.卫星数据集。
其次，许多不同的计算技术在计算机视觉系统上已经被采用，如标准的优化方法，人工智能的搜索策略，退火模拟，遗传算法[14,15,16]。
具体的线性时间不变过滤器的使用是最常见的应用到边缘检测问题的程序，并且计算量最小。
在第一阶滤波器的情况下，边缘被解释为两个相邻像素灰度值的突然变化。
在这种情况下的目标是确定图像中第一个派生的灰度点，它将起到很大的重要性。
通过应用新输出图像的门槛，边缘在任意方向都可以被检测到。
在其他方面，边缘检测滤波器的输出是多边形逼近技术的输入，多边形逼近技术提取将要被检测特征[1]。
被确认为具有特殊属性的特征点以及像素在图像分析中扮演了一个非常重要的角色。
特征点包括由著名最优秀的边缘发现者PreWitt, Sobel, Marr, and Canny测定的边缘像素[17:21]。
最近许多人对由一些运算子产生的特征点恢复了兴趣，比如Plessey“角”运算子，或者是Motavec介绍的兴趣点运算子[24,25]。
古典算子定义了一个像素作为特征点的特殊类。
古典算子在图像研究区域是高对比情况下运作良好。
事实上，在一些可以将图像通过简单阀值的转换成二进制图像的区域上，古典算子运作十分良好，如图一所示。
同时要明确古典算子的不足：古典边缘检测对标记边缘像素效果不佳，就算是明确的边缘，也只代表小小的灰度跳跃。
但往往这样的边缘人眼都能清晰可辨。
总之，特征点是通过窗口上的像素值之间的关系来特征化的。
阈值灰度图二进制图图1 测量物体的灰度和二进制图像图二模糊图像处理的一般结构最近的研究发现可以使用神经模糊功能开发边缘检测，在一个相对较小的原型边缘设置上训练之后，再用古典边缘检测在样品图像中进行分类。
这项工作是Bezdek 等人开创的，他们训练一个神经网络去像正规的Sobel算子一样给出相同的模糊的结果。
然而，通过笔者和合作者的工作已表明训练神经网络去分类脆值是个更有效的Bezdek计划变体。
神经模糊边缘检测的优势甚至超过传统的边缘检测。
在[27, 28]系统中的描述，模糊推理系统的输入是通过向原始图像申请高通滤波器，Sobel算子和低通滤波器而获得的。
然后整个结构转换成对比增强滤波器起作用，另一个问题是，转换成输入类里指定数量的部分图像。
为数字图像上的边缘检测，一种新型的基于模糊逻辑推理策略的FIS方法已经准备好了，而不是确定阀值或需要训练算法。
这种提出的方法通过将图像分割成浮动3*3的二进制区域开始。
一个直接的模糊推理系统测绘一个从浮动矩阵上得到不同值的范围来检测边缘。
A 模糊图像处理模糊图像处理是所有方法的集合，包括理解，描绘或处理作为模糊集的段或特征的图像。
描绘或处理取决于选定的模糊技术或需要解决的问题。
模糊图像处理有三个主要阶段：图像模糊化，隶属度值的修改，以及必要是的图像去模糊化，如图2所示。
模糊化和去模糊化的步骤归咎于我们不具备硬件模糊。
因此，图像数据的编码(模糊)和对结果的解码(去模糊化)使通过模糊技术处理图像变为可能。
模糊图像处理的主要力量是中间步骤(隶属度值的修改)。
当图像数据从灰度平面转换成隶属度平面(模糊)，通过模糊技术修改隶属度。
这可以是一个模糊聚类，一个基于规则的模糊方法，模糊积分方法等[29]。
B 模糊集与模糊隶属函数系统的实施考虑了一些情况，模糊化后得到的输入图像和输出图像都是8位量化；这种方法中，它们的灰度水平总是处于0到255之间。
模糊集被创建出来代表每个变量的强度；这些集相当于语言学变量“黑”，边和“白”。
模糊集对应输入，三角形对应输出，隶属函数因此起作用，如图3.图3函数的模糊隶属度与输入和输出有关图五模糊图像处理步骤该功能被采用去实施“与”和“或”的运作，分别是最小和最大的功能。
推理方法被选择作为去模糊程序，这意味着模糊集是通过应用输入数据的每个推理规则得到的，而输入数据是通过添加功能加入的；系统的输出是作为结果隶属函数的部分计算出来的。
输出的三个隶属函数的值被设计去分离图像中的黑，白和边缘值。
C 推理规则的定义推理规则取决于八个相邻灰度像素的权重，是否相邻权重是黑色或白色。
这些规则的强大之处是能直接提取处理图像中的所有边缘。
本研究是通过研究每个像素的每个邻居来测定处理图像中的所有像素。
每个像素的条件取决于使用浮点3 3伪装时可以扫描所有的灰度。
在这个位置，一些需要的规则被解释。
如果灰色在一行中代表黑色，并且残余的灰色是白，然后检测过的像素是边(图4-a)，那么前四个规则处理伪装中被选中或中心像素的垂直和水平方向的灰度值。
第二个四规则处理八邻居也取决于灰度权重的值，如果四个连续像素的权重是黑并且剩下的四邻居是白，那么中心像素代表边缘(图4-b)。
介绍了的规则和另一组规则是用来检测边缘，白和黑像素的。
剩下的图像有助于轮廓，黑色和白色区域。
从模糊结构的边得知，输入灰度是介于0-255灰度强度之间，并且根据所需的规则，灰度被转换为隶属函数的值，如图5所示。
根据去模糊化得到的FIS的输出被呈现介于0-255之间。
然后黑，白和边缘被检测出来。
从这次研究中的检测图像的经验中，被得到的最好结果是黑介于0-82，并且白权重介于80-255。
三．实验该系统通过不同的图像来进行测试，它的性能被拿来与Sobel算子和FIS方法进行比较。
当图像中的提取边如图6所示，我们使用这幅图像作为古典Sobel算子和FIS 方法的对比模型，我们调整与启动顺序相关的模糊规则来取得更好的结果。
原始图像显示在图6的部分上。
基于Sobel算子的边缘检测使用MATLAB上的处理工具箱，如b部分所示。
图片中的白色像素表明边缘，从而将保持平滑。
显然，边缘图像上还留有一些噪音，并且有一些边缘已损坏。
运用图像上新的FIS来检测它的边，结果发现，修改后的边缘图片有更少的噪音和更少的损坏，如图6-c所示。
为了分割任务，一个薄边缘更好，因为我们只想要保持边缘而不是附近的细节。
边缘图的值是正常的间隔0或1来代表急躁隶属度值。
原始捕捉的图像显示在图6-a。
我们观察，在b部分，在图像的二进制值里阀值自动估计的Sobel算子不允许边缘在低对比度区域被检测。
在两边产生的结果被发现(双边缘)在b部分的左边。
反过来，FIS系统甚至允许在低对比度区域检测边缘，如c 部分说明。
这是由于在不同对比度区域，模糊规则所给出的不同待遇，并且为了避免包括不属于连续线上的输出图像像素，该规则被制定。
图6-a 测试三角边的原始捕获图像图6-b使用经典算子检测出来的边缘图6-c用模糊推理规则检测提取出来的边在图7中，测量对象的一个合成图像和它的边缘分离成黑色，如a部分所示。
当Sobel算子被应用于这一图像，一个断开的边缘出现在左边。
模糊规则的采用被专门设立用来避免获取单边图像的双边缘结果，然而FIS系统被应用于相同的图像(c)。
它给出了线上的永久流畅度和直线性。
为了表现出边缘检测上的性能优化，图8通过齿轮的不同灰度图像来显示。
我们模糊技术产生的结果图像似乎在平坦地区更加平滑并且噪音更低，并且紧张地区比用传统的Sobel算子得来的更加尖锐。
图7-a为检测圆形和矩形图7-b使用经典算子得出的边边而拍摄的原始图像图7-c：通过FIS系统得到的边缘图8-c：为测试齿轮边捕获的原始图像图8-b：使用Sobel算子的出的边图8-c：FIS系统得到的边四．结论由于不确定性存在于图像处理的许多方面，模糊处理是可取的。
这些不确定性包括低水平图像处理上的附加和非附加噪音，假设算法的不精确性，以及在高水平图像处理上的含糊不清的解释。
为了边缘检测的同步处理，通常把边作为强度脊的模型。
然而在实践中，这一假设只有近似，导致了这些算法的一些缺陷。
模糊图像处理是一个专家知识边缘上的强大的工具形式表述，和不同来源上得来的不精确信息。
设计模糊规则是一个有吸引力的方案，他能尽可能的提高边缘质量。
这种算法的缺陷是它们需要大量的计算。
这些结果是我们得出这样的结论：1.FIS系统的实施呈现了对比度和采光变化上更大的鲁棒性，除了避免获取双边缘。
2.它在线条流畅性，直线直线性和曲线圆度性上给出了永久影响。
同时，边角更加锐化，并且可以很容易的定义。