09 chapter9 Digital Image Processing.pptx

第1页：第09章数字图像处理封面

数字图像处理：章节封面，标示本章主题与内容范围。

第2页：9.1 What Is Digital

• 9.1 What Is Digital Image Process ing 1
• An image may be defined as a two-dimensional function,, where xandyare spatial (plane) coo rdi nates, and the amplitude offat any pair of coordinate sis called the intensity or gray level of the image at that point. Whenx, y, and the amplitude values offare all finite, discrete quant ities

本页为英文原文段落，中文译文见本章「参考译文」部分。

第3页：9.1 图像处理概念

Vision is the most advanced one in our senses, so it isnot surprising 3 that images play the single most important role in human perception. However, unlike hum ans, who are limited to the visual band of the electro magnetic (EM) spectrum, imaging mach ines cover almost the entire EM spectrum, ranging from gamma toradio waves. They can operate on images gene rated bysources that humans are not accustomed to assoc iatingwith images. These include ultrasound, electron microscopy, and computer-gene rated images. Thus, digital image process ing encompass es a wide and varied field of applications

本页为英文原文段落，中文译文见本章「参考译文」部分。

第4页：9.1 图像处理概念

There is no general agreement among authors regard ing where image process ing stops and other related areas, such as image analysis and computer vision, start. Sometimes a distinction is madeby def in ing image process ing as a discipline in which both the in put 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 as an image process ing operation. On the other hand, there are fields such as computer vision whose ultimate goal is to use com put ers to emulate human vision, including learn ing and being able to make inference sand take actions based on visual inputs 4. 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 termsof development, with progress having been much slower than origin ally anticipated. The area of image analysis (also called image understanding) is in between image process ing and computer vision

本页为英文原文段落，中文译文见本章「参考译文」部分。

第5页：9.1 图像处理概念

There are no clear-cut bound aries in the continuum from image process ing at one end to computer vision at the other. However, one useful paradigm is to consider three types of computerize dprocesses in this continuum: low-, mid-, and high-level process es. Low-level process es involve primitive opera tions such as image pre process ing to reduce noise, contrast enhancement, and image sharpen ing. A low-level process is character ized by the fact that bothits inputs and outputs are images. Mid-level process ing on image sinvolves tasks such as segmentation (partition ing an image intoregions or objects), description of those objects to reduce them to aform suitable for computer process ing, and classification (recognition) of individual objects. A mid-level process is character ized by the fact that its inputs gene rally are images, but its outputs are attribute sextracted from those images (e. g., edges, con tours, and the identity of individual objects). Finally, higher-level process ing involves"making sense" of an ensemble of recognized objects, as in image analysis, and, at the far end of the continuum, perform ing the cognitive functions norm ally associated with vision

本页为英文原文段落，中文译文见本章「参考译文」部分。

第6页：9.1 图像处理概念

Based on the preceding comments 5, we see that a logical place of overlap between image process ing and image analysis is the areaof recognition of individual regions or objects in an image. Thus, what we call in this chapter digital image process ing encompass esprocesses whose inputs and outputs are images and, in addition, encompass es process es that extract attributes from images, up toand including the recognition of individual objects. As a simple illust ration to clarify these concepts, consider the area of automate dana lysis of text. The process es of acquiring an image of the area contain ing the text, pre process ing that image, extract ing(segment ing) the individual char act ers, describing the char act ers ina form suitable for computer process ing, and recogni zing those individual char act ers are in the scope of what we call digital image process ing in this chapter. Making sense of the content of the pagemay be viewed as being in the domain of image analysis and even computer vision, dep ending on the level of complexity implied by the statement "making sense". As will become evident shortly, digital image process ing, as we have defined it

本页为英文原文段落，中文译文见本章「参考译文」部分。

第7页：生词与短语：image

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第8页：生词与短语：intensity

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第9页：课文注释（9.1 图像处理概念）

本页为课文注释（Notes），讲解本节难句的句式与翻译处理。

• Notes
• 1.本节主要介绍数字图像处理的概念
• 2. be com posed of由……组成，由……构成。此句可翻译为：值得提及的是数字图像是由有限的元素组成的，每一个元素都有一个特定的位置和幅值
• 3. it is not surprising that…显而易见的是，毫无疑问。此句可翻译为：视觉是人类最高级的感知器官，所以，毫无疑问图像在人类感知中扮演着最重要的角色
• 4.此句为一简单句，句型是there +be句型，whose引导的定语从句修饰fields。此句可翻译为：另一方面，有些领域(如计算机视觉)研究的最高目标是用计算机去模拟人类视觉，包括理解和推理并根据视觉输入采取行动等
• 5. based on the preceding comments根据上述讨论，根据上述分析。此句可翻译为：根据上述讨论可知，图像处理和图像分析两个领域合乎逻辑的重叠部分是图像中特定区域或物体的识别这块

第10页：9.2 The Origins of

• 9.2 The Origins of Digital Image Process ing 1
• One of the first applications of digital images was in the newspaper industry, when pictures were first sent by submarine cable between London and New York. Introduction of the Bart lane cable picture transmission system in the early 1920 sreduced the time required to transport a picture across the Atlantic from more than a week to less than three hours. Specialized printing equipment coded pictures for cable trans mission and then reconstructed them at the receiving end
• Some of the initial problems in improving the visual quality ofthese early digital pictures were related to the selection of printing procedures and the distribution of intensity levels. The printing method was abandoned toward the end of 1921 in favorof a technique based on photographic repro duct ion made fromtapes perforated at the telegraph receiving terminal. The improve ments over the printing method are evident, both in tonal quality and in resolution

本页为英文原文段落，中文译文见本章「参考译文」部分。

第11页：9.2 图像处理起源

• The early Bart lane systems were capable of codingimages in five distinct levels of gray. This capability wasincreased to 15 levels in 1929. During this period, introduction of a system for develop ing a film plate via light beams that were modulated by the coded picture tape improved the repro duct ion process considerably
• Although the examples just cited involve digital images, they are not considered as digital image process ingresults in the context of our definition because com put ers were not involved in their creation. Thus, the history of digital image process ing is inti mat ely tied to 2 the development of the digital computer. In fact

本页为英文原文段落，中文译文见本章「参考译文」部分。

第12页：9.2 图像处理起源

The idea of a computer goes back to the invention of the abacus in Asia Minor, more than 5000 years ago. More recent ly, there were develop ments in the past two centuries that are the foundation of what wecall a computer today 3. However, the basis for what we call a modern digital computer dates back to only the 1940 s with the introduction by John von Neumann of two key concepts: (1) a memory to hold a sto red program and data, and (2) conditional branch ing 4. These two ideas arethe foundation of a central process ing unit (CPU), which is at the heart of com put ers today. Star ting with von Neumann, there were a series of keyadvances that led to com put ers powerful enough to be used for digital image process ing. Briefly

本页为英文原文段落，中文译文见本章「参考译文」部分。

第13页：9.2 图像处理起源：Work

The first com put ers powerful enough to carry out meaningful image process ing tasks appear ed in the early 1960 s. The birth ofwhat we call digital image process ing today can be traced to thea vail ability of those mach ines and the onset of the space program during that period. It took the combination of those two develop ments to bring into focus the potential of digital image process ing concepts. Work on using computer techniques forimproving images from a space probe began at the Jet Propulsion Laboratory (Pasadena, California) in 1964 when pictures of the moon transmitted by Ranger 7 were processed bya computer to correct various types of image distortion inherent in the on-board television camera. Figure 9.1 shows the first image of the moon taken by Ranger 7 on July 31,1964 at 9:09 a. m. Eastern Daylight Time (EDT), about 17 minutes before imp acting the lunar surface (the markers, called reseau marks, are used for geometric cor recti ons). This also is the first image ofthe moon taken by a U. S. spacecraft. The imaging less ons learned with Ranger 7 served as the basis for improved method sused to enhance and restore images from the Surveyor mission sto the moon

本页为英文原文段落，中文译文见本章「参考译文」部分。

第14页：9.2 图像处理起源

• Figure 9.1 The first picture of the moon by a U. S. spacecraft
• In parallel with 5 space applications, digital image process ingtechni ques began in the late 1960 s and early 1970 s to be used in medical imaging, remote earth resources observations, and astronomy. The invention in the early 1970 s of comp uteri zed axial tomography (CAT), also called comp uteri zed tomography (CT) for short

本页为英文原文段落，中文译文见本章「参考译文」部分。

第15页：9.2 图像处理起源

As the source rotates, this procedure is repeated. Tomography consists of algor it hms that use the sense ddata to construct an image that represents a "slice"through the object. Motion of the object in a direction perpendicular to the ring of detect ors produces a set ofsuch slices, which constitute a three-dimensional (3 D) rendition of the inside of the object. Tomography wasinvented independent ly by Sir Godfrey N. Houns field and Professor Allan M. Cormack, who shared the 1979 Nobel Prize in Medicine for their invention. It is interesting tonote that X-rays were discovered in 1895 by Wilhelm Conrad Roentgen, for which he received the 1901 Nobel Prize for Physics. These two inventions, nearly 100 years apart, led to some of the most active application are as of image process ing today

本页为英文原文段落，中文译文见本章「参考译文」部分。

本页英文原文由图片提取，部分单词之间空格缺失，已按大小写与标点边界补回分隔，确切分词请以课件原图为准。

第16页：9.2 图像处理起源

From the 1960 s until the present, the field of image process ing has grown vigorous ly. In addition to applications in medicine and the space program, digital image process ing techniques now are usedin a broad range of applications. Computer procedures are used to enhance the contrast or code the intensity levels into color for easier interpretation of X-rays and other images used in industry, medicine, and the biological sciences. Geo graph ers use the same or similar techniques to study pollution patter ns from aerial and satellite imagery. Image enhancement and restoration procedures are usedto process de graded images of un recoverable objects or experimental results too expensive to duplicate. In arc he ology, image process ing methods have success fully restored blurred pictures that were the only available records of rare art if acts lost or damaged after being photog rap hed. In physics and related fields, computer techniques rout in ely enhance images of experiments inareas such as high-energy plasmas and electron microscopy. Similar ly successful applications of image process ing concepts canbe found in astronomy, biology, nuclear medicine, law enforcement, defense

本页为英文原文段落，中文译文见本章「参考译文」部分。

第17页：9.2 图像处理起源

These examples illustrate process ing results intended for human interpretation. The second major area of application of digital image process ing techniques mentioned at the beginning of this chapter is in solving problems dealing with machine perception. In this case, interest focuses on procedures for extract ing froman image information in a form suitable for computer process ing. Often, this information bears little resemb lance to visual feature sthat humans use in interpret ing the content of an image. Examples of the type of information used in machine perception are statistical moments, Fourier transform coefficients, and multi dimensional distance measures. Typical problems in machine perception that rout in ely utilize image process ingtechni ques are automatic character recognition, indus trial machine vision for product assembly and inspection, military recognizance, automatic process ing of finger prints, screening of X-rays and blood samples

本页为英文原文段落，中文译文见本章「参考译文」部分。

本页英文原文由图片提取，部分单词之间空格缺失，已按大小写与标点边界补回分隔，确切分词请以课件原图为准。

第18页：生词与短语：submarine

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第19页：生词与短语：lunar

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第20页：生词与短语：microscopy

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第21页：课文注释（9.2 图像处理起源）

本页为课文注释（Notes），讲解本节难句的句式与翻译处理。

• Notes
• 1.本节主要介绍数字图像处理的起源及发展
• 2. be tied to原意为“束缚于……”，“捆绑于……”，此处翻译为“与……有关”。此句可翻译为：数字图像处理的发展历史和数字计算机的发展密切相关
• 3.句中的that引导定语从句修饰develop ments；而在宾语从句中，what we call a computer today作介词of的宾语，直接译为“计算机”。此句可翻译为：近两个世纪以来算盘的一些发展为计算机的出现奠定了基础
• 4. date back to追溯到。此句可译为：然而，现代计算机的基础还要回溯到20世纪40年代由约翰·冯·诺依曼提出的两个重要概念
• (1)保存程序和数据的存储器
• (2)条件分支
• 5. In parallel with与……同时，与……平行。此句可译为：进行空间应用的同时，数字图像处理技术自20世纪60年代末和20世纪70年代初开始还被用于医学图像、地球遥感监测和天文学等领域

第22页：9.3 Fundamental

• 9.3 Fundamental Steps in Digital Image Process ing 1
• It is helpful to divide the material covered in the following sections into the two broad categories defined in Section 9.1: methods whose input and output are images, andmethods whose inputs may be images, but whose outputs are attributes extract ed from those images 2. This organization is sum mari zed in Figure 9.2. The diagram does not imply that every process is applied toan image. Rather

本页为英文原文段落，中文译文见本章「参考译文」部分。

第23页：图9.2 Fundamental

图9.2 Fundamental

本页为图示页，图中给出9.3 图像处理步骤的结构或流程，具体图形见课件原图。

第24页：9.2 gave some hints

• Image acquisition is the first process shown in Figuere 9.2. The discussion in Section 9.2 gave some hints regard ing the origin of digital images. This topic is considered in much moredetail in the following, where we also introduce a number ofbasic digital image concepts that are used throughout the chapter. Note that acquisition could be as simple as being given an image that is already in digital form. Gene rally, the image acquisition stage involves pre process ing, such ass cal ing
• Image enhancement is among the simple st and most appealing areas of digital image process ing. Basically, theidea behind enhancement techniques is to bring out detail thatis obs cured, or simply to highlight certain features of interest in an image 3. A familiar example of enhancement is when we incr ease the contrast of an image because "it looks better". Itis important to keep in mind that enhancement is a very subjective area of image process ing

本页为英文原文段落，中文译文见本章「参考译文」部分。

第25页：9.3 图像处理步骤

• 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 rest oration techniques tend to be based on mathematical or probabilistic models ofimage deg rad at ion 4. Enhancement, on the other hand, is based onhuman subjective prefer ences regard ing what cons titutes a "good"enhancement result
• Color image process ing is an area that has been gaining in importance because of the significant incr ease in the use of digital images over the Internet
• Compression, as the name implies, deals with techniques for reducingthe storage required to save an image, or the bandwidth required to transmit it 5. Although storage technology has improved significant lyover the past decade, the same cannot be said for trans mission capacity. This is true particular ly in uses of the Internet

本页为英文原文段落，中文译文见本章「参考译文」部分。

第26页：9.3 图像处理步骤

• Morphological process ing deals with tools for extract ing image compo nents that are useful in the represent at ion and description of shape. The material in this chapter beg ins a transition from process es that output images toprocesses that output image attributes, as indicated in Section 9.1
• Segmentation procedures partition an image into its constituent parts or objects. In general 6, autonomous segmentation is one of the most difficult tasks in digital image process ing. A rugged segmentation procedure brings the process a long way toward successful solutionof imaging problems that require objects to be identified individual ly. On the other hand, weak or erratic segmentation algor it hms almost always guarantee eventual failure. In general, the more accurate the segmentation

本页为英文原文段落，中文译文见本章「参考译文」部分。

本页英文原文由图片提取，部分单词之间空格缺失，已按大小写与标点边界补回分隔，确切分词请以课件原图为准。

第27页：9.3 图像处理步骤

• Represent at ion and description almost always follow the output of aseg mentation stage, which usually is raw pixel data, cons tituting either the boundary of a region (i. e., the set of pixels sep ara ting one image region from another) or all the points in the region itself 7. In either case
• Recognition is the process that assigns a label (e. g., "vehicle") to an object based on its descriptors. As detailed in Section 9.1, we conclude our coverage of digital image process ing with the development of methods for recognition of individual objects

本页为英文原文段落，中文译文见本章「参考译文」部分。

第28页：9.3 图像处理步骤

So far we have said nothing about the need for prior knowledge or about the interaction between the knowledge base and the process ing modules in Figure 9.2. Knowledge about a problem domain is coded into an image process ing system in the form of a knowledge database. This knowledge may be as simple as de tailing regions of animage where the information of interest is known to belocated, thus limiting the search that has to be conducted ins eek ing that information. The knowledge base also can bequite complex, such as an int err elated list of all major possible defects in a materials inspection problem or animage database contain ing high-resolution satellite image sof a region in connection with change-detection applications. In addition to guiding the operation of each process ing module, the knowledge base also controls the inter action between modules. This distinction is made in Figure 9.2 bythe use of double-headed arrows between the process ingmodules and the knowledge base, as opposed to single-headed arrows linking the process ing modules

本页为英文原文段落，中文译文见本章「参考译文」部分。

第29页：9.3 图像处理步骤

Although we do not discuss image display explicit lyat this point, it is important to keep in mind that viewing the results of image process ing can take place at the output of any stage in Figure 9.2. Wealso note that not all image process ing application sre quire the complexity of inter actions implied by Figure 9.2. In fact, not even all those modules are needed in some cases. For example, image enhancement for human visual interpretation seldom requires use of any of the other stages in Figure 9.2. In general, however, as the complexity of an image process ing task increases, so does the number ofprocesses required to solve the problem

本页为英文原文段落，中文译文见本章「参考译文」部分。

第30页：生词与短语：diagram

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第31页：生词与短语：morphological

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第32页：课文注释（9.3 图像处理步骤）

本页为课文注释（Notes），讲解本节难句的句式与翻译处理。

• Notes
• 1.本节主要介绍数字图像处理所包含的内容以及基本步骤
• 2.句中it是形式主语，真正的主语是helpful后的动词不定式to divide。此句可翻译为：把后续各部分涉及的材料划分为在9.1节定义的两个主要类别是有帮助的。一类的输入和输出都是图像；另一类的输入可能是图像，但是输出是从图像中提取的特征属性
• 3.句中连词or连接动词不定式to bring out和to highlight作表语。此句可翻译为：从本质上讲，增强技术的思想是显现那些被模糊了的细节，或简单地突出一幅图像中用户感兴趣的特征
• 4.句中that引导的同位语从句作sense的同位语。此句可翻译为：然而，不像增强，图像增强是主观的，而图像复原是客观的。从某种意义上说，复原技术倾向于以图像退化的数学或概率模型为基础
• 5.句中as为连词，译为“正如”、“如同……那样”。此句可翻译为：压缩，顾名思义，就是减少图像的存储量，或者在传输图像时降低频带
• 6. in general通常，总之，一般而言。此句可翻译为：通常，自主分割是数字图像处理中最为困难的任务之一
• 7. which引导的定语从句修饰the output，定语从句中的现在分词短语的逻辑主语是raw pixel data。翻译时可采用分译法和顺译法。此句可翻译为：表示与描述几乎总是在分割步骤的输出后边，这一输出通常是未加工的数据，其构成不是区域的边缘(区分一个图像区域和另一个区域的像素集)就是其区域本身的所有点

第33页：9.4 Compo nents of

• 9.4 Compo nents of an Image Process ing System 1
• As recent ly as the mid-1980 s, numerous models ofimage process ing systems being sold throughout the world were rather substantial peripheral devices that attached to equally substantial host com put ers. Late inthe 1980 s and early in the 1990 s, the market shifted toimage process ing hardware in the form of single boa rds designed to be compatible with industry standard busesand to fit into engineering workstation cabin ets and personal com put ers. In addition to lowering costs

本页为英文原文段落，中文译文见本章「参考译文」部分。

第34页：9.4 系统部件

• Although large-scale image process ing systems still are being sold for massive imaging applications, such as process ing of satellite images, the trend continues toward mini aturizing and blend ing of general-purpose 2 small com put ers with specialized image process ing hardware. Figure 9.3 shows the basic compo nents comp rising a typical general-purpose system used for digital image process ing. The function of each component is discussed in the following paragraphs
• With reference to sensing, two elements are required to acquire digital images. The first is a physical device that is sensitive to the energy radiated by the object we wish toimage 3. The second, called a digitizer, is a device for convert ing the output of the physical sensing device in to digital form. For instance, in a digital video camera, these nsors produce an electrical output proportional to light intensity. The digitizer converts these outputs to digital data

本页为英文原文段落，中文译文见本章「参考译文」部分。

第35页：9.4 系统部件

Specialized image process ing hardware usually consist sof 4 the digitizer just mentioned, plus hardware that performs other primitive opera tions, such as an arithmetic logic unit (ALU), which performs arithmetic and logical opera tions in parallel on entire images. One example of how an ALU is used is in ave raging image sas quickly as they are digitized, for the purpose of noise reduction. This type of hardware sometimes is called afront-end subsystem, and its most distinguish ing characteristic is speed. In other words, this unit perform sfunctions that require fast data through puts (e. g., digit izing and ave raging video images at 30 frames/s) that the typical main computer cannot handle

本页为英文原文段落，中文译文见本章「参考译文」部分。

本页英文原文由图片提取，部分单词之间空格缺失，已按大小写与标点边界补回分隔，确切分词请以课件原图为准。

第36页：9.4 系统部件

• The computer in an image process ing system is age neral-purpose computer and can range from a PC to asuper computer. In dedicated applications, sometime sspecially designed com put ers are used to achieve arequired level of performance, but our interest here is on general-purpose image process ing systems. In these syst ems, almost any well-equipped PC-type machine is suitable for offline image process ing tasks
• Software for image process ing consists of specialize dmodules that perform specific tasks. A well-designed package also includes the capability for the user to write code that, as a minimum, utilizes the specialize dmodules. More sophisticated software packages allow the integration of those modules and general-purpose software comma nds from at least one computer language

本页为英文原文段落，中文译文见本章「参考译文」部分。

第37页：9.4 系统部件

Mass storage capability is a must in image process ingapplications. An image of size 1024×1024 pixels, in which the intensity of each pixel is an 8-bit quantity, requires one megabyte of storage space if the image is not compressed. When dealing with thousands, or even millions, of images, providing adequate storage in an image process ing system can be a challenge. Digital storage for image process ingapplications falls into three principal categories: (1) short-term storage for use during process ing, (2) on-line storage for relative ly fast recall, and (3) archival storage, character ized by infrequent access. Storage is measured inbytes (eight bits), Kbytes (one thousand bytes), Mbytes (one million bytes), Gbytes (meaning giga, or one billion, bytes), and Tbytes (meaning tera, or one trillion, bytes)

本页为英文原文段落，中文译文见本章「参考译文」部分。

本页英文原文由图片提取，部分单词之间空格缺失，已按大小写与标点边界补回分隔，确切分词请以课件原图为准。

第38页：9.4 系统部件

One method of providing short-term storage is computer memory. Another is by specialized boards, called frame buff ers, that store one or more images and can beaccessed rapidly, usually at video rates (e. g., at 30 complete images per second)6. The latter method allow svirtually instantaneous image zoom, as well as scroll(vertical shifts) and pan (horizontal shifts). Frame buff ers usu ally are housed in the specialized image process ing hardware unit shown in Figure 9.3. Online storage gene rally takes the form of magnetic disks or optical-media storage. The key factor character izing on-line storage is frequent access to the stored data. Finally, archival storage is character ized by massive storage require ments but infrequent need for access. Magnet ictapes and optical disks housed in "jukebox es" are the usual media for archival applications

本页为英文原文段落，中文译文见本章「参考译文」部分。

第39页：图9.3 Compo nents

图9.3 Compo nents

本页为图示页，图中给出9.4 系统部件的结构或流程，具体图形见课件原图。

第40页：9.4 系统部件

• Image displays in use today are mainly color (preferably flat screen) TV monitors. Monitors are driven by the outputs of image and graph ics display cards that are an integral part of the computer system. Seldom are there require ments for image dis playa pplications that cannot be met by display cards available com mercia lly as part of the computer system. In some cases, it is necessary to have stereo displays
• Hard copy devices for recording images include laser print ers, film came ras, heat-sensitive devices, inkjet units, and digital units, suchas optical and CD-ROM disks. Film prov ides the highest possible resolution, but paper is the obvious medium of choice for written material. For present ations, images are dis played on film transp arencies or in a digital medium if image projection equipment is used. The latter approach is gaining acceptance as the standard for image present ations

本页为英文原文段落，中文译文见本章「参考译文」部分。

第41页：9.4 系统部件

Networking is almost a default function in any computer system in use today. Because of the large amount ofdata inherent in image process ing applications, the key consideration in image transmission is bandwidth. Indedicated net works, this typical ly is not a problem, but comm unications with remote sites via the Internet arenot always as efficient. For tun at ely, this situation isimproving quickly as a result of optical fiber and other broadband techno log ies

本页为英文原文段落，中文译文见本章「参考译文」部分。

第42页：生词与短语：substantial

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第43页：生词与短语：supercomputer

本页为生词与短语（New words and phrases）列表，收录本节出现的专业词汇。

词条按原文「中文释义—音标—英文词条」的顺序识别，如与原件排版不符以课件原图为准。

第44页：课文注释（9.4 系统部件）

本页为课文注释（Notes），讲解本节难句的句式与翻译处理。

• Notes
• 1.本节主要介绍数字图像处理系统的部件
• 2. general-purpose通用的。此句可翻译为：虽然针对大规模图像应用(如处理卫星图像)的图像处理系统一直在出售，但朝着小型化和通用化的小型机并带有专用图像处理软件的混合型系统方向发展
• 3.句中that引导定语从句修饰device，而在该定语从句中，省略that的定语从句we wish to image修饰the object。此句可翻译为：第一个是物理设备，该设备对人们希望成像的物体发射的能量很敏感
• 4. consist of包含，由……构成。此句可翻译为：特殊的图像处理硬件通常由刚刚谈到的数字化器与执行其他原始操作[如算术逻辑单元(ALU)]的硬件组成，算术逻辑单元对图像执行并行操作
• 5.句中must为名词，意思是“绝对必要的事物”。此句可翻译为：大规模存储能力在图像处理中是必需的
• 6. that引导的定语从句修饰boards，过去分词短语called frame buff ers将被修饰词boards和that引导的定语从句分隔开来。此句可翻译为：另一种方法采用专用板，该存储板称为帧缓存(缓冲存储器)，它们可以存储一帧或多帧图像并可快速访问，通常以视频速率(30帧/秒)访问

第45页：译文 9.1 图像处理概念

• 参考译文
• 第9章数字图像处理
• 数字图像处理方法的研究源于两个主要应用领域：其一是为了便于人们分析而对图像信息进行改进；其二是为了实现机器自动感知而对图像数据进行存储、传输及显示。本章主要包括以下内容：(1)界定图像处理的范围；(2)回顾一下图像处理领域的历史起源；(3)简要讨论一下在数字图像处理中进行的主要研究；(4)概述一下典型的、通用的图像处理系统的组成
• 9.1数字图像处理的概念
• 一幅图像可定义为一个二维函数，这里x和y是空间坐标，而在任何一对空间坐标上的幅值f称为该点图像的强度或灰度。当x，y和幅值f为有限的、离散的数值时，称该图像为数字图像。数字图像处理是指借用数字计算机处理数字图像，值得提及的是数字图像是由有限的元素组成的，每个元素都有一个特定的位置和幅值，这些元素称为图像元素、画面元素或像素。像素是被广泛使用的表示数字图像元素的词汇

第46页：译文 9.1 图像处理概念

• 视觉是人类最高级的感知器官，所以，毫无疑问图像在人类感知中扮演着最重要的角色。然而，人类感知只限于电磁波谱的视觉波段，成像机器则可覆盖几乎全部电磁波谱，从伽马射线到无线电波。它们可以对人类不太习惯的那些图像源进行加工，包括超声波、电子显微镜及计算机产生的图像。因此，数字图像处理涉及各种各样的应用领域
• 图像处理涉及的范畴或其他相关领域(例如，图像分析和计算机视觉)的界定在创始人之间并没有达成一致。有时用处理的输入和输出内容都是图像这一特点来界定图像处理的范围。我们认为这一定义仅是人为界定和限制。例如，在这个定义下，甚至最普通的计算一幅图像灰度平均值的工作都不能算是图像处理。另一方面，有些领域(如计算机视觉)研究的最高目标是用计算机去模拟人类视觉，包括理解和推理并根据视觉输入采取行动等。这一领域本身是人工智能的分支，其目的是模仿人类智能。人工智能领域处在其发展过程中的初期阶段，它的发展比预期的要慢得多，图像分析(也称为图像理解)领域则处在图像处理和计算机视觉两个学科之间

第47页：译文 9.1 图像处理概念

• 从图像处理到计算机视觉这个连续的统一体内并没有明确的界线。然而，在这个连续的统一体中可以考虑3种典型的计算处理(即低级、中级和高级处理)来区分其中的各个学科。低级处理涉及初级操作，如降低噪声的图像预处理、对比度增强和图像锐化。低级处理是以输入、输出都是图像为特点的处理。中级处理涉及分割(把图像分为不同区域或目标物)以及缩减对目标物的描述，以使其更适合计算机处理及对不同目标的分类(识别)。中级图像处理是以输入为图像，但输出是从这些图像中提取的特征(如边缘、轮廓及不同物体的标识等)为特点的。高级处理涉及在图像分析中对被识别物体的总体理解，以及执行与视觉相关的识别函数(处在连续统一体边缘)等
• 根据上述讨论可知，图像处理和图像分析两个领域合乎逻辑的重叠部分是图像中特定区域或物体的识别这一块。因此，在本章中界定数字图像处理为输入和输出均是图像的处理，同时也包括从图像中提取特征及识别特定物体的处理。下面举一个简单的文本自动分析方面的例子来具体说明这一概念。在自动分析文本时首先获取一幅包含文本的图像，对该图像进行预处理，提取(分割)字符，然后以适合计算机处理的形式描述这些字符，最后识别出这些字符，而所有这些操作都在本书界定的数字图像处理的范围内。理解一页的内容可能要根据理解的复杂度从图像分析或计算机视觉领域考虑问题。这样，本章定义的数字图像处理的概念将在有特殊社会和经济价值的领域内通用。在以下各节中给出的概念是在那些应用领域所使用的方法的基础

第48页：9.2 数字图像处理的起源

• 9.2数字图像处理的起源
• 数字图像处理最早应用于报纸业，当时，图像第一次通过海底电缆从伦敦传往纽约。早在20世纪20年代曾引入Bart lane电缆图片传输系统，把横跨大西洋传送一幅图片所需的时间从一个多星期减少到3个小时。为了用电缆传输图片，首先要进行编码，然后在接收端用特殊的打印设备重构该图片
• 这些早期数字图像视觉质量的改进工作，涉及打印过程的选择和亮度等级的分布等问题。这种打印方法到1921年底就被彻底淘汰了，转而支持一种基于光学还原的技术，该技术在电报接收端用穿孔纸带打出图片。与打印方法相比，这种方法在色调质量和分辨率方面的改进都很明显
• 早期的Bart lane系统可以用5个灰度等级对图像编码，到1929年已增加到15个等级。在这一时期，由于引入了一种用编码图像纸带去调制光束而使底片感光的系统，使图像复原过程得到明显改善

第49页：译文 9.2 图像处理起源：数字图像要求

• 刚才引用的数字图像的例子并没有考虑数字图像处理的结果，这主要是因为没有涉及计算机。因此，数字图像处理的发展历史与数字计算机的发展密切相关。事实上，数字图像要求非常高的存储和计算能力，因此数字图像处理领域的发展要依赖于数字计算机及数据存储、显示和传输等相关技术的发展
• 计算机的概念可追溯到5000多年前中国算盘的发明。近两个世纪以来算盘的一些发展也为计算机的出现奠定了基础。然而，现代计算机的基础还要回溯到20世纪40年代由约翰·冯·诺依曼提出的两个重要概念：(1)保存程序和数据的存储器；(2)条件分支。这两个概念是中央处理单元(CPU)的基础。现在，它是计算机的心脏。从冯·诺依曼开始，产生了一系列重要技术进步，使得计算机的功能变得足够强大，可以用于数字图像处理领域

第50页：译文 9.2 图像处理起源

• 简单说，这些进步可归纳为如下几点
• (1)1948年贝尔实验室发明了晶体三极管
• (2)20世纪50年代到20世纪60年代高级编程语言(如COBOL和FORTRAN)的开发
• (3)1958年得州仪器公司发明了集成电路(IC)
• (4)20世纪60年代早期操作系统的开发
• (5)20世纪70年代Intel公司开发了微处理器(由中央处理单元、存储器和输入输出控制器组成的单个芯片)
• (6)1981年IBM公司推出了个人计算机
• (7)20世纪70年代出现的大规模集成电路(LI)引发了元件微小化革命，20世纪80年代出现了VLSI (超大规模集成电路)，现在已出现了ULSI

第51页：译文 9.2 图像处理起源

• 伴随着这些技术的进步，大规模的存储和显示系统也发展起来，这两者均是数字图像处理的基础
• 第一台可以执行有意义的图像处理任务的大型计算机出现在20世纪60年代早期。数字图像处理技术的诞生可追溯至这一时期对这些机器的使用和空间项目的开发，这两大发展把人们的注意力集中到数字图像处理的潜能上。利用计算机技术改善空间探测器发回的图像的工作，始于1964年美国加利福尼亚的喷气推进实验室。当时由“旅行者7号”卫星传送的月球图像由一台计算机进行了处理，以校正航天器上电视摄像机中各种类型的图像畸变。图9.1显示了由“旅行者7号”于1964年7月31日上午(东部白天时间)9点09分在光线影响月球表面前约17分钟时摄取的第一张月球图像[痕迹(称为网状痕迹)用于几何校正]，这也是美国航天器取得的第一幅月球图像。“旅行者7号”传送的图像可作为改善的增强和复原图像(例如，来自“探索者”登月飞行、“水手号”系列空间探测器及阿波罗载人登月飞行的图像)方法的基础

第52页：图9.1 美国航天器传送的第1张月球图

图9.1 美国航天器传送的第1张月球图

本页为图示页，图中给出译文 9.2 图像处理起源的结构或流程，具体图形见课件原图。

第53页：译文 9.2 图像处理起源：进行空间应用

• 进行空间应用的同时，数字图像处理技术自20世纪60年代末和20世纪70年代初开始还被用于医学图像、地球遥感监测和天文学等领域
• 早在20世纪70年代发明的计算机轴向断层术(CAT，简称计算机断层CT)是图像处理在医学诊断领域最重要的应用之一
• 计算机轴向断层术是一种处理方法，在这种处理中，一个检测器环围绕着一个物体(或病人)，并且一个X射线源(与检测器环同心)绕着物体旋转
• X射线穿过物体并由位于对面环中的相应检测器收集起来
• 当X射线源旋转时，这一过程不断重复
• 断层技术由一些算法组成，这些算法用感知的数据去重建通过物体的“切片”图像
• 当物体沿垂直于检测器的方向运动时就产生一系列这样的“切片”，这些切片组成了物体内部的再现图像
• 断层技术是由Godfiey N. Houns field先生和Allan M. Coimack教授发明的，他们共同获得了1979年诺贝尔医学奖

第54页：译文 9.2 图像处理起源

• 从20世纪60年代至今，图像处理领域已得到了生机勃勃的发展
• 除了医学和空间项目外，数字图像处理技术现在已应用于更多的领域
• 可以使用计算机程序增强对比度或将亮度编码为彩色，以便分析X射线和用于工业、医学及生物科学等领域的其他图像
• 地理学用相同或相似的技术从航空和卫星图像中研究污染模式
• 图像增强和复原过程用于处理不可修复物体的已损图像或者造价昂贵不可复制的实验结果
• 在考古学领域，使用图像处理方法已成功地复原了模糊的图片，这些图片是丢失或损坏的稀有物品唯一现存的记录
• 在物理学和相关领域，利用计算机技术通常能增强如高能等离子和电子显微镜等领域的实验图像
• 图像处理技术还被成功地应用在天文学、生物学、核医学、法律实施、国防及工业领域中

第55页：译文 9.2 图像处理起源

• 这些例子说明图像处理的结果可用于人类进行分析
• 本章开始曾提到，数字图像处理技术的第二个主要应用是解决机器感知问题
• 在这种情况下要将注意力集中在以更适合计算机处理的形式从图像中提取信息的过程
• 通常，这一信息有点类似于人类用于解释一幅图像内容的视觉特征
• 用于机器感知的信息有统计矩、傅里叶变换系数和多维距离测度
• 在机器感知中，通常使用图像处理技术解决的典型问题有字符自动识别、用于产品生产线及检测的工业机器视觉、军事识别、指纹的自动处理、X射线和血样分类处理、用于天气预报和环境鉴定的航空与卫星图像的机器处理
• 计算机价格性能比的不断下降和通过万维网及因特网实现的网络和通信带宽的扩展，已为数字图像处理技术的持续发展创造了前所未有的机会

第56页：9.3 数字图像处理的基本步骤

• 9.3数字图像处理的基本步骤
• 把后续各节涉及的内容划分为在9.1节定义的两个主要类别是有帮助的。一类的输入和输出都是图像，另一类的输入可能是图像，但输出是从图像中提取的特征属性。这种组织在图9.2中做了总结。该图并不意味着每一步处理都用于图像。更确切的目的是给出所有方法的一个概念，这些方法可以针对不同的目的，也许具有不同的目标
• 图像获取是图9.2所示过程的第一步。在9.2节的讨论中给出了关于数字图像来源的说明，下面将对此进行更具体的研究。本节还将介绍一些基本的数字图像概念。获取图像与给出一幅数字形式的图像一样简单。通常，图像获取包括如设置比例尺等预处理操作

第57页：图9.2 数字图像处理的基本步骤

图9.2 数字图像处理的基本步骤

本页为图示页，图中给出译文 9.3 图像处理步骤的结构或流程，具体图形见课件原图。

第58页：译文 9.3 图像处理步骤：增强技术

• 图像增强是数字图像处理中最简单和最有吸引力的领域。从本质上讲，增强技术的思想是显现那些被模糊了的细节，或简单地突出一幅图像中用户感兴趣的特征。一个图像增强的例子是增强图像的对比度，使其看起来好一些。应记住，增强是图像处理中非常主观的领域，这一点很重要
• 图像复原也是改进图像外貌的一个处理领域。然而，不像增强，图像增强是主观的，而图像复原是客观的。从某种意义上说，复原技术倾向于以图像退化的数学或概率模型为基础。另一方面，好的增强效果的构成是基于人的主观偏好确定的
• 彩色图像处理已经成为一个重要领域，因为基于互联网的图像处理应用在不断增长

第59页：译文 9.3 图像处理步骤：虽然存储技术

• 压缩，顾名思义，就是减少图像的存储量，或者在传输图像时降低频带。虽然存储技术在过去的10年内有了很大改进，但在传输量方面还不能这样说，尤其在互联网上更是如此，互联网是以大量的图片内容为特征的。图像压缩技术因图像文件扩展名而被大多数计算机用户所熟悉，如. jpg文件扩展名用于JPEG(联合图片专家组)图像压缩标准
• 形态学处理涉及提取图像元素的工具，它在表现和描述形状方面非常有用。正如9.1节指出的，本章将从输出图像的处理向输出图像特征的处理的转换开始
• 分割过程用于将一幅图像划分为组成部分或对象。通常，自主分割是数字图像处理中最为困难的任务之一。复杂的分割过程导致要求将各个对象分别识别出来的图像问题还要进行大量的研究。另一方面，不健壮且不稳定的分割算法几乎总会导致最终失败。通常，分割越准确，识别越成功

第60页：译文 9.3 图像处理步骤

• 表示与描述几乎总是在分割步骤的输出后边，这一输出通常是未加工的数据，其构成不是区域的边缘(区分一个图像区域和另一个区域的像素集)就是区域本身的所有点。无论哪种情况，把数据转换成适合计算机处理的形式都是必要的。首先，必须确定数据是应该被表示为边界还是整个区域。当注意的焦点是外部形状特性(如拐角和曲线)时，表示成边界是合适的。当注意的焦点是内部特性(如纹理或骨骼形状)时，则表示成区域是合适的。在某些应用中，这些表示方法是互补的。选择一种表示方式仅是把原始数据转换为适合计算机后续处理的形式的一部分。为了描述数据以使用户感兴趣的特征更明显，还必须确定一种方法。描述也称为特征选择，涉及提取特征的技术，该特征是某些用户感兴趣的定量信息或是区分一组目标与其他目标的基础
• 识别是基于目标的描述给目标赋以符号的过程。如9.1节所讨论的那样，我们用识别个别目标方法的开发推出数字图像处理的覆盖范围

第61页：译文 9.3 图像处理步骤

• 到目前为止还没有谈到图9.2中关于先验知识及知识库与处理模块之间的交互这部分内容
• 关于问题域的知识以知识库的形式被编码载入一个图像处理系统
• 这一知识可能如处理图像细节区域那样简单，在这里，用户感兴趣的信息被定位，这样，限制性的搜索就被引导到寻找的信息处
• 知识库也可能相当复杂，如材料检测问题中所有主要缺陷的相关列表或者图像数据库(该库中包含和变化检测应用相关区域的高分辨率卫星图像)
• 除了引导每一个处理模块的操作，知识库还要控制模块间的交互
• 这一特性在图9.2中的处理模块和知识库间用双向箭头表示
• 而单向箭头连接的是处理模块

第62页：译文 9.3 图像处理步骤：中任何阶段

• 虽然这里没有专门地讨论图像显示，但图像显示还是很重要的
• 可在图9.2中任何阶段的输出上图像处理结果
• 此外要注意不是所有的图像处理应用都需要图9.2中所显示的复杂交互
• 事实上，在某些情况下并不是所有那些模块都需要
• 例如，用于人可视化解释的图像增强很少需要使用图9.2中的其他步骤
• 然而，随着图像处理任务复杂度的增加通常要做更多处理才能解决问题

第63页：9.4 图像处理系统的部件

• 9.4图像处理系统的部件
• 20世纪80年代中叶，世界各地售出的各种型号的图像处理系统，确切地说，基本上都是由主机及与主机相配合的外设装置构成的。20世纪80年代末90年代初，市场已转为将图像处理硬件设计为与工业标准总线兼容并能配合工程工作站机箱和个人计算机的单板形式。除了降低价格外，这一市场还如催化剂一样出现了大量的新公司，这些公司的任务是开发用于图像处理的软件
• 虽然针对大规模图像应用(如处理卫星图像)的大规模图像处理系统一直在出售，但发展趋势是朝着小型化和通用化的小型机并带有专用图像处理软件的混合型系统方向发展。图9.3显示了用于数字图像处理的典型通用系统的组成部件。每个组件的功能将在下面的段落中从图像感知开始进行介绍

第64页：图9.3 通用图像处理系统组成

图9.3 通用图像处理系统组成

本页为图示页，图中给出译文 9.4 系统部件的结构或流程，具体图形见课件原图。

第65页：译文 9.4 系统部件：需要两个部件

• 关于感知，需要两个部件以获取数字图像。第一个是物理设备，该设备对人们希望成像的物体发射的能量很敏感。第二个称之为数字化器，数字化器是一种把物理感知装置的输出转换为数字形式的设备。例如，在数字视频摄像机中，由传感器产生一个与光强成比例的输出，由数字化器把该输出转化为数字数据
• 特殊的图像处理硬件通常由刚刚谈到的数字化器与执行其他原始操作[如算术逻辑单元(ALU)]的硬件组成，算术逻辑单元对图像执行并行操作。一个使用ALU的例子是与数字化一样快的图像取平均操作，这一操作的目的是降低噪声。有时这种类型的硬件称为前端子系统，它的显著特点是速度快。换句话说，这个单元负责执行需要数据快速吞吐的功能(例如，在30帧/s的速度下数字化和求视频图像的平均值)，典型的主机不能胜任该工作

第66页：译文 9.4 系统部件：图像处理系统

• 图像处理系统中的计算机是通用的计算机，从PC到超级计算机都是。有时在专门的应用中也采用特殊设计的计算机以达到所要求的性能水平。但是，这里令人感兴趣的还是通用的图像处理系统。在这些系统中，几乎所有配置较好的PC对于离线图像处理任务都是适合的
• 图像处理软件由执行特殊任务的特殊模块组成。一个设计优良的软件包还包括为用户写代码的能力，至少能够使用专用的模块。较复杂的软件包允许将那些模块和至少用一种计算机语言编写的通用软件命令集成到一起
• 大规模存储能力在图像处理中是必需的。一幅图像的尺寸是1024×1024像素，每像素的亮度是8比特，如果图像不压缩需1 MB的存储空间。当处理1000甚至100万幅图像时，在图像处理系统中很难提供足够的存储空间。图像处理应用的数字存储分成3个主要类别：(1)用于处理时的短期存储；(2)关系到快速调用的在线存储；(3)档案存储，其特点是频繁访问。存储是以字节(B)、千字节(KB)、兆字节(MB)、吉字节(GB)、太字节(TB)为单位的

第67页：译文 9.4 系统部件：另一种方法

• 提供短期存储的一种方法是使用计算机内存。另一种方法采用专用板，该存储板称为帧缓存(缓冲存储器)，它们可以存储一帧或多帧图像并可快速访问，通常以视频速率(30帧/s)访问。后一种方法允许实际意义上的快速图像变焦以及卷动(垂直移动)和摇(水平移动)。帧缓存通常放在专业的图像处理硬件单元中，如图9.3所示。在线存储一般采取磁盘或光介质存储。在线存储的关键特性参数是对存储数据的访问频率。档案存储是以海量存储要求为特点的，但无须频繁访问。放在类似于投币电唱机的盒子内的磁带和光盘是通常使用的档案存储介质
• 今天人们使用的图像显示器主要是彩色电视监视器(平面屏幕更好)。监视器由图像和图形显示卡驱动，它们是计算机系统的组成部分。对于图像显示应用不可能没有这样的要求，即作为计算机系统的一部分，其显示卡应满足商用性要求，在有些场合还要求立体显示。立体显示是采用戴在用户头上的头革(目镜上嵌有两个小的显示屏)来实现的

第68页：译文 9.4 系统部件

• 用于记录图像的硬拷贝装置包括激光打印机、胶片照相机、热敏装置、喷墨装置和数字单元(如CD-ROM)等。胶片的分辨率最高，但相纸作为记录材料是首先要选择的介质。为了描述图像，可将图像显示在透明胶片上或者使用图像投影设备将其显示在数字介质上。后面的方法作为图像描述标准正在被接受
• 联网在今天所用的计算机系统中几乎都是默认的功能，因为在图像处理应用中涉及大量数据，所以在图像传输中考虑的主要问题是带宽。在专用网络中这不是一个典型问题，但通过互联网远程通信就不是总有效了。幸运的是，这一状况由于光纤和其他宽带技术的发展正在迅速得到改进