南京工程学院(英文文献翻译201110715)
中英文翻译
南京工程学院Nanjing Institute of Technology毕业设计英文资料翻译Translation of the English Material of Graduation Design学生姓名：陈建伟学号： 240112608Name: Chen Jianwei Number: 240112608 班级： K 暖通 111Class: K-Nuantong111所在学院：康尼学院College:K a n g n i C o l l e g e专业：建筑环境与设备工程Profession: Building Environment and Equipment Engineering指导教师：刘明会Tutor: Liu Minghui2015年 3月 7日CHAPTER 1THERMODYNAMICS AND REFRIGERATION CYCLESTHERMODYNAMICS ................................................................... 1.1First Law of Thermodynamics ......................................................... 1.2Second Law of Thermodynamics ..................................................... 1.2Thermodynamic Analysis of Refrigeration Cycles ........................... 1.3Equations of State .............................................................................. 1.3Calculating Thermodynamic Properties ............................................. 1.4COMPRESSION REFRIGERATION CYCLES ................................ 1.6Carnot Cycle ........................................................................................ 1.6Theoretical Single-Stage Cycle Using a Pure Refrigerantor Azeotropic Mixture .................................................................. 1.8Lorenz Refrigeration Cycle ................................................................... 1.9Theoretical Single-Stage Cycle Using ZeotropicRefrigerant Mixture ..................................................................... 1.10 Multistage Vapor Compression Refrigeration Cycles ........................ 1.10Actual Refrigeration Systems ................................................................ 1.12ABSORPTION REFRIGERA TION CYCLES ..................................... 1.14Ideal Thermal Cycle .............................................................................. 1.14Working Fluid Phase ChangeConstraints ............................................... 1.14Working Fluids ....................................................................................... 1.15Absorption Cycle Representations .......................................................... 1.16Conceptualizing the Cycle ....................................................................... 1.16Absorption Cycle Modeling ..................................................................... 1.17Ammonia-Water Absorption Cycles ........................................................ 1.19Nomenclature for Examples .................................................................... 1.20 THERMODYNAMICS is the study of energy, its transformations, and its relation to states of matter. This chapter covers the application of thermodynamics to refrigeration cycles. The first part reviews the first and second laws of thermodynamics and presents methods for calculating thermodynamic properties. The second and third parts address compression and absorption refrigeration cycles, two common methods of thermal energy transfer.THERMODYNAMICSA thermodynamic system is a region in space or a quantity of matter bounded by a closed surface. The surroundings include everything external to the system, and the system is separated fromthe surroundings by the system boundaries. These boundaries can be movable or fixed, real or imaginary. Entropy and energy are important in any thermodynamic system. Entropy measures the molecular disorder of a system. The more mixed a system, the greater its entropy; an orderly or unmixed configuration is one of low entropy. Energy has the capacity for producing an effect and can be categorized into either stored or transient forms.Stored EnergyThermal (internal) energy is caused by the motion of molecules and/or intermolecular forces.Potential energy (PE) is caused by attractive forces existing between molecules, or the elevation of the system.mgzPE=(1)wherem =massg = local acceleration of gravityz = elevation above horizontal reference planeKinetic energy (KE) is the energy caused by the velocity of molecules and is expressed as22m VKE=(2)whereV is the velocity of a fluid stream crossing the system boundary.Chemical energy is caused by the arrangement of atoms composing the molecules.Nuclear (atomic) energy derives from the cohesive forces holding protons and neutrons together as the atom’s nucleus.Energy in TransitionHeat Q is the mechanism that transfers energy across the boundaries of systems with differing temperatures, always toward the lower temperature. Heat is positive when energy is added to the system (see Figure 1).Work is the mechanism that transfers energy across the boundaries of systems with differing pressures (or force of any kind),always toward the lower pressure. If the total effect produced in the system can be reduced to the raising of a weight, then nothing but work has crossed the boundary. Work is positive when energy is removed from the system (see Figure 1).Mechanical or shaft work W is the energy delivered or absorbed by a mechanism, such as a turbine, air compressor, or internal combustion engine.Flow work is energy carried into or transmitted across the system boundary because a pumping process occurs somewhere outside the system, causing fluid to enter the system. It can be more easily understood as the work done by the fluid just outside the system on the adjacent fluid entering the system to force or push it into the system. Flow work also occurs as fluid leaves the system.Flow work =pv (3)where p is the pressure and v is the specific volume, or the volume displaced per unit mass evaluated at the inlet or exit.A property of a system is any observable characteristic of the system. The state of a system is defined by specifying the minimum set of independent properties. The most common thermodynamic properties are temperature T, pressure p, and specific volume v or density ρ. Additional thermodynamic properties include entropy, stored forms of energy, and enthalpy.Frequently, thermodynamic properties combine to form other properties. Enthalpy h is an important property that includes internal energy and flow work and is defined aspvuh+≡(4)where u is the internal energy per unit mass.Each property in a given state has only one definite value, and any property always has the same value for a given state, regardless of how the substance arrived at that state.A process is a change in state that can be defined as any change in the properties of a system.A process is described by specifying the initial and final equilibrium states, the path (if identifiable), and the interactions that take place across system boundaries during theprocess.A cycle is a process or a series of processes wherein the initial and final states of the system are identical. Therefore, at the conclusion of a cycle, all the properties have the same value they had at the beginning. Refrigerant circulating in a closed system undergoes acycle.A pure substance has a homogeneous and invariable chemical composition. It can exist in more than one phase, but the chemical composition is the same in all phases.If a substance is liquid at the saturation temperature and pressure,it is called a saturated liquid. If the temperature of the liquid is lower than the saturation temperature for the existing pressure, it is called either a subcooled liquid (the temperature is lower than the saturation temperature for the given pressure) or a compressed liquid (the pressure is greater than the saturation pressure for the given temperature).When a substance exists as part liquid and part vapor at the saturation temperature, its quality is defined as the ratio of the mass of vapor to the total mass. Quality has meaning only when the substance is saturated (i.e., at saturation pressure and temperature).Pressure and temperature of saturated substances are not independent properties.If a substance exists as a vapor at saturation temperature and pressure, it is called a saturated vapor. (Sometimes the term dry saturated vapor is used to emphasize that the quality is 100%.) When the vapor is at a temperature greater than the saturation temperature, it is a superheated vapor. Pressure and temperature of a superheated vapor are independent properties, because the temperature can increase while pressure remains constant. Gases such as air at room temperature and pressure are highly superheated vapors.FIRST LAW OF THERMODYNAMICSThe first law of thermodynamics is often called the law of conservation of energy. The following form of the first-law equation is valid only in the absence of a nuclear or chemical reaction.Based on the first law or the law of conservation of energy for any system, open or closed, there is an energy balance asNet amount of energy Net increase of stored=added to system energy in systemor[Energy in] – [Energy out] = [Increase of stored energy in system]Figure 1 illustrates energy flows into and out of a thermodynamic system. For the general case of multiple mass flows with uniform properties in and out of the system, the energy balance can be written=-++++-+++∑∑W Q gz V pv u m gz V pv u m out out in in )2()2(22 []system i i f f gz V pv u m gz V pv u m )2()2(22++-++ (5) where subscripts i and f refer to the initial and final states,respectively.Nearly all important engineering processes are commonly modeled as steady-flow processes. Steady flow signifies that all quantities associated with the system do not vary with time. Consequently,0)2()2(22=-+++-++∑∑W Q gz V h m gz V h m leavingstream all entering stream all (6)where h = u + pv as described in Equation (4).A second common application is the closed stationary system for which the first law equation reduces to[]system i f u u m W Q )(-=- (7)SECOND LAW OF THERMODYNAMICSThe second law of thermodynamics differentiates and quantifies processes that only proceed in a certain direction (irreversible) from those that are reversible. The second law may be described in several ways. One method uses the concept of entropy flow in an open system and the irreversibility associated with the process. The concept of irreversibility provides added insight into the operation of cycles. For example, the larger the irreversibility in a refrigeration cycle operating with a given refrigeration load between two fixed temperature levels, the larger the amount of work required to operate the cycle. Irreversibilities include pressure drops in lines and heat exchangers, heat transfer between fluids of different temperature, and mechanical friction. Reducing total irreversibility in a cycle improves cycle performance. In the limit of no irreversibilities, a cycle attains its maximum ideal efficiency. In an open system, the second law of thermodynamics can be described in terms of entropy asdI s m s m dS e e i i T Q system +-+=δδδ(8) wheredS = total change within system in time dt during process systemδm s = entropy increase caused by mass entering (incoming)δm s = entropy decrease caused by mass leaving (exiting)δQ/T = entropy change caused by reversible heat transfer between system and surroundings at temperature TdI = entropy caused by irreversibilities (always positive)Equation (8) accounts for all entropy changes in the system. Rearranged, this equation becomes[]I d dS s m s m T Q sys i i e e -+-=)(δδδ (9)In integrated form, if inlet and outlet properties, mass flow, and interactions with the surroundings do not vary with time, the general equation for the second law isI ms ms T Q S S out in revsystem i f +-+=-∑∑⎰)()(/)(δ (10)In many applications, the process can be considered to operate steadily with no change in time. The change in entropy of the system is therefore zero. The irreversibility rate, which is the rate of entropy production caused by irreversibilities in the process, can be determined by rearranging Equation (10):∑∑∑--=surrin out T Q ms ms I )()( (11) Equation (6) can be used to replace the heat transfer quantity.Note that the absolute temperature of the surroundings with which the system is exchanging heat is used in the last term. If the temper-ature of the surroundings is equal to the system temperature, heat istransferred reversibly and the last term in Equation (11) equals zero.Equation (11) is commonly applied to a system with one mass flow in, the same mass flow out, no work, and negligible kinetic or potential energy flows. Combining Equations (6) and (11) yields []surr inout in out T h h s s m I ---=)( (12)In a cycle, the reduction of work produced by a power cycle (or the increase in work required by a refrigeration cycle) equals the absolute ambient temperature multiplied by the sum of irreversibilities in all processes in the cycle. Thus, the difference in reversible and actual work for any refrigeration cycle, theoretical or real, operating under the same conditions, becomes∑+=I T W W reversible actual 0 (13)THERMODYNAMIC ANAL YSIS OFREFRIGERA TION CYCLESRefrigeration cycles transfer thermal energy from a region of low temperature T to one of higher temperature. Usually the higher-T R temperature heat sink is the ambient air or cooling water, at temperature T 0, the temperature of the surroundings.The first and second laws of thermodynamics can be applied to individual components to determine mass and energy balances and the irreversibility of the components. This procedure is illustrated in later sections in this chapter.Performance of a refrigeration cycle is usually described by a coefficient of performance (COP), defined as the benefit of the cycle (amount of heat removed) divided by the required energy input to operate the cycle:Useful refrigerating effectCOP ≡Useful refrigeration effect/Net energy supplied from external sources (14) Net energy supplied from external sources For a mechanical vapor compression system, the net energy supplied is usually in the form of work, mechanical or electrical, and may include work to the compressor and fans or pumps. Thus,net evapW Q COP = (15)In an absorption refrigeration cycle, the net energy supplied is usually in the form of heat into the generator and work into the pumps and fans, ornet gen evapW Q Q COP += (16)In many cases, work supplied to an absorption system is very small compared to the amount of heat supplied to the generator, so the work term is often neglected.Applying the second law to an entire refrigeration cycle shows that a completely reversible cycle operating under the same conditions has the maximum possible COP. Departure of the actual cycle from an ideal reversible cycle is given by the refrigerating efficiency:tev R COP COP)(=η (17)The Carnot cycle usually serves as the ideal reversible refrigeration cycle. For multistage cycles, each stage is described by a reversible cycle.EQUATIONS OF STATEThe equation of state of a pure substance is a mathematical relation between pressure, specific volume, and temperature. When the system is in thermodynamic equilibrium,(18)The principles of statistical mechanics are used to (1) explore the fundamental properties of matter, (2) predict an equation of state based on the statistical nature of a particular system, or (3) propose a functional form for an equation of state with unknown parameters that are determined by measuring thermodynamic properties of a substance. A fundamental equation with this basis is the virial equation. The virial equation is expressed as an expansion in pressure p or in reciprocal values of volume per unit mass v as(19)(20)where coefficients B ’, C ’, D ’, etc., and B, C, D, etc., are the virial coefficients. B ’ and B are second virial coefficients; C ’ and C are third virial coefficients, etc. The virial coefficients are functions of temperature only, and values of the respective coefficients in Equations (19) and (20) are related. For example, B ’ = B/RT and C ’ = (C – B2)/(RT)2. The ideal gas constant R is defined as(21)where (pv)T is the product of the pressure and the volume along an isotherm, and tp is the defined temperature of the triple point of water, which is 491.69°R. The current best value of R is 1545.32 ft·lbf/(lb mole·°R).The quantity pv/RT is also called the compressibility factor; i.e., Z = pv/RT orAn advantage of the virial form is that statistical mechanics can be used to predict the lower order coefficients and provide physical significance to the virial coefficients. For example, in Equation(22), the term B/v is a function of interactions between two molecules, C/v2 between three molecules, etc. Since the lower order interactions are common, the contributions of the higher order terms are successively less. Thermodynamicists use the partition or distribution function to determine virial coefficients; however, experimental values of the second and third coefficients are preferred. For dense fluids, many higher order terms are necessary that can neither be satisfactorily predicted from theory nor determined from experimental measurements. In general, a truncated virial expansion of four terms is valid for densities of less than one-half the value at the critical point. For higher densities, additional terms can be used and determined empirically.Digital computers allow the use of very complex equations of state in calculating p-v-T values, even to high densities. The Benedict-Webb-Rubin (B-W-R) equation of state (Benedict et al. 1940) and the Martin-Hou equation (1955) have had considerable use, but should generally be limited to densities less than the critical value. Strobridge (1962) suggested a modified Benedict-Webb-Rubin relation that gives excellent results at higher densities and can be used for a p-v-T surface that extends into the liquid phase.The B-W-R equation has been used extensively for hydrocarbons (Cooper and Goldfrank 1967):where the constant coefficients are Ao, Bo, Co, a, b, c, α, γ. The Martin-Hou equation, developed for fluorinated hydrocarbon properties, has been used to calculate the thermodynamic property tables in Chapter 20 and in ASHRAE ThermodynamicProperties of Refrigerants (Stewart et al. 1986). The Martin-Houequation is as follows:第1章工程热力学和制冷循环热力学........................................................................................................................1.1热力学第一定律.......................................................................................................1.2热力学第二定律.......................................................................................................1.2制冷循环的热力学分析.......................................................................................1.3状态方程.....................................................................................................................1.3计算热力学性质........................................................................................................1.4压缩制冷循环............................................................................................................1.6卡诺循环.....................................................................................................................1.6理论单级循环使用纯制冷剂或共沸混合物.....................................................1.8洛伦兹制冷循环........................................................................................................1.9理论单级循环使用非共沸的制冷剂混合物......................................................1.10多级蒸汽压缩制冷循环...........................................................................................1.10实际制冷系统..............................................................................................................1.12吸收制冷周期............................................................................................................1.14理想的热循环..............................................................................................................1.14工作流体相变约束条件............................................................................................1.14工作液..............................................................................................................................1.15吸收循环表示...............................................................................................................1.16概念化循环....................................................................................................................1.16吸收循环建模................................................................................................................1.17氨水吸收周期...............................................................................................................1.19实例命名........................................................................................................................1.20工程热力学是研究能量及其转换和能量与物质状态之间的关系。
南京工程学院专业介绍.doc
南京工程学院专业介绍南京工程学院是江苏省属普通本科高校，坐落于历史文化名城南京。
学校是全国应用型本科院校专门委员会主任委员单位，全国效劳特需硕士专业学位研究生培养单位联盟副理事长单位，也是国家“卓越工程师教育培养方案”、“CDIO工程教育模式改革研究与实践”首批试点高校之一和江苏省协同创新中心培育建立单位。
学校还依托特色学科和行业优势，积极探索和构建多元化的科技创新与孵化机制，充分发挥产学合作的优势和产业园区的科技孵化功能，实现了学校科技产业的良性互动开展。
学校与企业申报共建了8个省级工程技术研究中心和1个博士后科技工作站，“南京工程学院技术转移中心”成为同类高校中首个省级技术转移中心;先后有3项新产品获国家重点新产品称号，3项产品获江苏省高新技术产品称号，康尼公司已成为中国最大的轨道交通门系统高新技术企业。
校办产业销售总额年均产值近12亿元，在全国高校科技产业中名列前茅。
xx年8月，康尼机电股份成功上市，成为江苏省在上证交易所上市的首家校资企业，校资产业效劳学校教学科研的能力不断增强。
今后一段时期，学校将切实遵照第二次党代会提出的“全力开创特色鲜明的高水平应用型工程大学建立新局面”的奋斗目标，狠抓“突出一个重点，构建两大特色，实现四项提升”的工作重点，坚持以工程教育为主体，以本科教育为根本，以应用型品牌专业和工程技术特色学科建立为抓手，突出师资队伍、平台载体和体制机制三大重点，努力深化内涵建立，不断彰显办学特色，着力打造应用型本科质量名校，确保我校在新一轮改革与开展中继续走在全国同类高校前列。
南京工程学院是一所工科类普通本科院校，拥有共3个最好专业(特色专业)。
南京工程学院自动化、热能与动力工程、电气工程及其自动化等专业可以说是南京工程学院最好最有特色的专业了，这些专业为同类型高校相关专业和本校的专业建立与改革起到示范带动作用。
专业名称：国际经济与贸易，本科，学制4年。
培养目标：本专业培养适应社会主义现代化建立和未来社会与经济开展需要的、德、智、体、美等全面和谐开展与安康个性相统一，富有工程意识、实践能力和创新精神，具有经济、管理和贸易方面的理论素养，能在涉外经济贸易部门、各类企业及外贸公司从事经贸工作的应用型人才。
南京工程学院毕业论文任务书范文模板
文档编号：0000 193
优秀毕业设计
NANJING INSTITUTE OF TECHNOLOGY
南京工程学院毕业论文任务书
目录
目录. (2)
南京工程学院（设计）论文任务书（模板一） (2)
南京工程学院（设计）论文任务书（模板二） (6)
南京工程学院（设计）论文任务书（模板三） (9)
南京工程学院（设计）论文任务书（模板四） (10)
南京工程学院（设计）论文任务书（模板五） (12)
南京工程学院（设计）论文任务书（模板六） (14)
南京工程学院（设计）论文任务书（模板七） (15)
南京工程学院（设计）论文任务书（模板八） (17)
南京工程学院（设计）论文任务书（模板九） (19)
南京工程学院（设计）论文任务书（模板十） (21)
本文精选了南京工程学院各个学院毕业（设计）论文任务书，总共10 个不同类别的任务书模板范文，都选自南京工程学院同学的优秀毕业论文，包含各个专业类别的任务书，适合各个学院同学们撰写毕业设计论文时参考和研究。
撰写毕业论文是检验学生在校学习成果的重要措施，也是提高教学质量的重要环节。
大学生在毕业前都必须完成毕业论文的撰写任务。
申请学位必须提交相应的学位论文，经答辩通过后，方可取得学位。
愿此文能助毕业班同学们一臂之力，欢迎下载。
南京工程学院（设计）论文任务书（模板一）
（包括原始数据、技术要求、工作要求）
5月20日：论文答辩
九、主要参考文献
主要参考文献
南京工程学院(设计)论文任务书(模板二)
3•对本毕业设计(论文)课题成果的要求〔包括毕业设计论文、图表、实物样品等〕：。
外文翻译2
南京工程学院Nanjing Institute of Technology毕业设计英文资料翻译Translation of the English Material of Graduation Design学生姓名：殷典学号：207100533 Name:YinDian Number:207100533班级：集控102Class:Jikong102所在学院：能源与动力工程学院College:College of Energy and Power Engineering专业：热能与动力工程Profession:Thermal Energy and Power Engineering指导教师：王金平Tutor:WangjinPing2014年3月5日汽轮机超速跳闸保护作者是总部位于德克萨斯州艾力市Lyondell/Equistar化学公司高级工程顾问查尔斯·鲁坦查尔斯（查理）鲁坦是一个高级工程顾问（工程院士），在这个位置上他可以享用总部位于德克萨斯州艾力市的Lyondell/Equistar化学公司的资源。
他的专长是在旋转设备领域，也擅长热开孔/堵漏和一些特殊的问题。
鲁坦先生曾经在孟山都化学公司，康菲公司，杜邦公司，该隐和西方等地方工作过。
他拥有两项专利，并研究思考世界各地关于涡轮机，热开孔和封堵的问题。
鲁坦先生在1973年从美国德州理工大学获得机械工程理学士学位，他已经出版并在文章中介绍过烃加工，ASME，AICHE，泵系统，振动研究所，休斯顿商务局表，得克萨斯A＆M的透平机械和国际泵用户专题讨论会，南方动力机械及气体压缩发布会和预测性维护技术大会。
鲁坦先生是AICHE过程燃气用户委员会委员，得克萨斯技术学院工程师，水利研究所/ANSI泵标准审查委员会委员，也是涡轮机研讨会咨询委员会委员。
摘要本文讨论了汽轮机超速跳闸保护以前的设计，历史的发展，当前的标准和设计，以及涡轮机超速和转子时间常数之间的关系。
毕业设计说明书论文-南京工程学院实践教学管理平台
南京工程学院毕业设计说明书(论文)作者：蒋为良学号：*********系部：能源与动力工程学院专业：能源与动力工程题目：660MW火电机组配套锅炉的烟、风系统设计（贾汪烟煤）指导者：辛洪祥副教授评阅者：潘效军教授2014 年 5 月18 南京毕业设计说明书（论文）中文摘要毕业设计说明书（论文）外文摘要目录第一章绪论 (1)第二章设计原理 (3)2.1 锅炉风、烟系统 (3)2.2 风、烟系统的设备及工作原理 (3)2.2.1送风机和一次风机 (4)2.2.2轴流式风机工作原理 (5)2.2.3空气预热器 (6)2.2.4除尘设备 (7)2.2.5烟囱 (7)2.3 系统介绍 (8)2.3.1二次风系统 (8)2.3.2一次风系统 (9)2.3.3烟气系统 (10)2.4 研究方向 (10)第三章热力计算 (11)3.1 设计思路 (11)3.2 计算步骤 (11)3.3 风、烟系统设备选型汇总 (29)结论 (30)参考文献 (31)致谢 (32)附录 (33)第一章绪论电厂锅炉以其容量大、参数高区别于一般工业锅炉。
电厂锅炉在火电厂中是提供动力的关键设备，因而电厂锅炉技术的进步对电力生产的发展有着直接影响。
20世纪50年代以前，电厂锅炉的发展一直落后于汽轮发电机，这限制了机组容量的提高。
最初，电厂采用火管锅炉。
这种锅炉容量小,压力低，效率低,适应不了电厂对动力日益增长的需求，因而被水管锅炉代替。
水管锅炉经历了由直水管向弯水管形式的发展。
后者与中、高参数机组配套，是电厂锅炉发展史上的一大进步。
随着材料、制造工艺、水处理技术、热工控制技术的进步，锅炉能够燃烧充分、着火稳定、运行可靠。
风烟系统中安装了省煤器、空气预热器等大大提高了燃料的利用率。
20世纪30年代，德国和苏联开始应用直流锅炉；40年代美国开发了多次强制循环锅炉。
到80年代，世界上最大的单台多次强制循环锅炉已可与1000 MW机组匹配。
西欧则发展了低倍率强制循环锅炉，最大的单台容量可配600MW机组。
南京工程学院图书馆
南京工程学院图书馆文献传递服务部分资源简介1. Springer-LINK（全文）网址1：/app/home数据库介绍：德国施普林格(Springer-Verlag)是世界上著名的科技出版集团, 通过Springer LINK系统提供学术期刊及电子图书的在线服务。
Springer公司和 EBSCO/Metapress 公司现已开通Springer LINK电子期刊服务。
目前Springer LINK按学科分为以下11个“在线图书馆”：生命科学、医学、数学、化学、计算机科学、经济、法律、工程学、环境科学、地球科学、物理学与天文学，是科研人员的重要信息源。
Springer-LINK还提供Kluwer电子期刊。
2. ACM（全文）网址1：/网址2：/portal.cfm数据库介绍：ACM Digital Library数据库收录了美国计算机协会（Association for Computing Machinery）的各种电子期刊、会议录、快报等文献。
请注意：目前该数据库中大多数内容可看到全文(pdf格式)，但有些文献只能看到文摘；各种文献的收录年代范围也不统一，有的收录自创刊起直到当前的最新内容，有的只收录了某几年的内容。
（目前镜像站点网址1正在升级，如无法正常访问，可选择主站点网址2访问，但需要自付国际流量费。
）3． Aerospace & High Technology Database网址1：/htbin数据库介绍：宇航及高技术数据库，是CSA众多数据库中的一个。
美国科学信息出版公司（Cambridge Scientific Abstracts，CSA）出版的剑桥科学文摘包括60多个数据库，覆盖的学科范围包括：生命科学、水科学与海洋学、环境科学、计算机科学、材料科学以及社会科学。
检索结果为文献的题录文摘信息。
4． BIOSIS Previews网址1：/数据库介绍：BIOSIS Previews由美国生物科学信息服务社 (BIOSIS)出版，是世界上最大的关于生命科学的文摘索引数据库。
论诗歌的翻译标准“传神达意”——以汪榕培译《枫桥夜泊》为例
Standards of Translating Poetry：Conveying Spirits and Sense of the Original——A Case Study of ＂Maple Bridge Night Mooring＂ Translated by Wang
Rongpei
作者： 刘性峰[1]
作者机构： [1]南京工程学院外语系,南京211167
出版物刊名： 哈尔滨工业大学学报：社会科学版
页码： 109-113页
年卷期： 2010年 第2期
主题词： 汪榕培;枫桥夜泊;传神达意;诗歌翻译标准
摘要：人们普遍认为诗歌不可译,或者说无法再现原作的原姿原貌。
翻译家汪榕培先后将
《诗经》、《老子》、《易经》、《邯郸记》、《陶渊明集》、《墨子》等翻译成英语,并赢得
了普遍认可。
汪榕培在翻译实践中一直遵守＂传神达意＂的翻译标准。
结合汪榕培翻译的《枫
桥夜泊》,从＂传神＂和＂达意＂两个方面,与《枫桥夜泊》的其他几个英文译本进行比较分析。
结果表明,汪榕培翻译的《枫桥夜泊》更能做到＂形神兼备＂,即＂传神地达意＂。
进一步分析＂传神＂与＂达意＂的关系,并指出＂传神达意＂是较好的诗歌翻译标准。
英文论文及中文翻译格式(附件1)
附件1南京工程学院英文论文及中文翻译英文论文题目：Multilevel Inverters for ElectricVehicle Applications中文翻译题目：多电平逆变器在电动汽车中的应用专业：车辆工程（车辆电子电气）班级：车电气101学号：215100439学生姓名：许兵指导教师：朱华教授张开林副教授说明：本页开始可直接附上打印好的pdf格式英文论文（注：不要把pdf格式的英文论文转化为word文档或其它格式），选择英文论文的要求如下：1.选择与你毕业设计（论文）内容相关或与你所学专业相关且公开发表的英文论文。
2.所选英文论文的页数4—5页为宜。
3.不要选择英文产品说明书或英文教材上的内容。
4.不要选择由中国学者翻译发表的英语论文。
多电平逆变器在电动汽车中的应用莱昂·托尔伯特，彭方Z. 托马斯·G. Habetler美国橡树岭国家实验室* 佐治亚理工学院邮政信箱2009 电气和计算机工程学院橡树岭，TN37831-8038 亚特兰大,佐治亚州30332-0250电话：（423）576-6206 电话：（404）894-9829传真：（423）241-6124 传真：（404）894-9171邮箱：tolbertlm@邮箱：tom.habetler 邮箱：pengfz@摘要：本文介绍了多电平逆变器在纯电动汽车（EV）和混合动力汽车（HEV）电机驱动上的应用。
二极管钳位逆变器和串联H桥型逆变器（1）能够产生只含基频近似的正弦电压，（2）几乎没有任何电磁干扰（EMI）和共模电压，（3）使得电动汽车更方便/更安全和对大多数电动汽车动力系统可能的开放性线路。
本文探讨了电动汽车的好处并讨论了使用电动汽车马达驱动或并联式混合动力汽车驱动和二极管钳位逆变器的一系列混合动力汽车电机驱动的串联逆变器的控制方案。
分析、仿真和实验结果表明了这些多电平逆变器在这个新领域中的优越性。
1．背景电动和混合动力汽车的发展，尤其是以牵引电机驱动为主的发展[1]，为电力电子行业提供了众多新机遇和挑战。
南京工程学院
南京工程学院
毕业设计开题报告
课题名称：电力负荷管理系统数据传输规约
研究与应用（中心站）
学生姓名：吉继梅
指导教师：张惠刚教授
所在院部：电力工程学院
专业名称：电气工程及其自动化(电网监控技术)
南京工程学院
2008年3月8日
说明
1．根据南京工程学院《毕业设计(论文)工作管理规定》，学生必须撰写《毕业设计（论文）开题报告》，由指导教师签署意见、教研室审查，系教学主任批准后实施。
2．开题报告是毕业设计（论文）答辩委员会对学生答辩资格审查的依据材料之一。
学生应当在毕业设计（论文）工作前期内完成，开题报告不合格者不得参加答辩。
3．毕业设计开题报告各项内容要实事求是，逐条认真填写。
其中的文字表达要明确、严谨，语言通顺，外来语要同时用原文和中文表达。
第一次出现缩写词，须注出全称。
4．本报告中，由学生本人撰写的对课题和研究工作的分析及描述，应不少于2000字，没有经过整理归纳，缺乏个人见解仅仅从网上下载材料拼凑而成的开题报告按不合格论。
5．开题报告检查原则上在第2～4周完成，各系完成毕业设计开题检查后，应写一份开题情况总结报告。
毕业设计(论文)开题报告。
南京工程学院毕业(设计)论文系统快速使用指南
南京工程学院毕业（设计）论文系统快速使用指南系统流程简单概述必选流程：1.指导老师进行“课题申报”（在填写课题申报时老师可指定学生做该课题或选择“盲选”给本专业的所有学生进行选择）。
2.教研室主任审核教师申报的课题。
3.教研室主任审核通过后，系教学主任进行发布课题。
4.学生进入选题。
注：如果是被老师指定选题的学生则不需要进行课题选择。
5.指导老师进行选题审核，并确认选题的学生。
6.老师审核完成后交由系教学主任审核选题并且进行“发布双选结果”（审核完成后必须进行发布否则无法进行以下流程）。
然后教研室主任就可以分配评阅教师。
7.指导老师填写任务书。
8.教研室主任进行审核任务书。
9.系教学主任进行审核任务书。
10.学生查看任务书，并且可以提交开题报告11.指导老师审核开题报告12.教研室主任审核开题报告13.系教学主任审核开题报告15.指导教师进行审核论文并进行评分。
16.指导教师审核并评分完成后由教研室主任分配论文评阅教师。
17.评阅教师进行论文评阅。
18.指导教师进行学生答辩资格审核19.答辩资格审核通过后教研室主任设置论文答辩组，并指定答辩组中答辩的论文学生及答辩秘书。
20.答辩秘书录入答辩审核意见并评分。
21.系教学秘书发布论文最终成绩并录入工作总结。
可选流程：1、学生提交校外毕业设计申请书。
2、指导教师审阅外出毕业设计申请。
3、专业负责审核外出申请。
4、系教学主任审核外出申请。
5、指导教师申报团队课题。
6、系教学主任审核团队课题。
7、团队课题组长提交项目总结。
8、指导教师下达外文翻译原文9、学生提交外文翻译10、学生提交中期检查教学秘书1 流程管理流程管理包括：各专业教师申报课题数、院内公告、优秀论文展示、汇总查询、工作总结、删除课题指定学生、取消学生答辩资格审核、成绩总评、论文终稿退回、教师审核退回、评阅审核退回、学院专家管理设置、推荐优秀指导教师、学院推荐优秀论文、优秀论文学院意见、推荐团队优秀、团队优秀论文学院意见、抽检学生信息查看、优秀论文学院专家意见；2特殊情况处理特殊情况处理包括总评成绩调整申请3截止日期设置截止日期设置分为课题申报截止日期，学生选题截止日期、提交任务书截止日期、提交开题报告截止日期、提交论文截止日期。
外文翻译
南京工程学院毕业设计外文资料翻译学生姓名：郗萍学号：240076001班级名称：K建工ZB074所在院系：康尼学院2012年3月10日Construction EngineeringAbstract: Construction engineering is a specialized branch of civil engineering concerned with the planning, execution, and control of construction operations for such projects as highways, buildings, dams, airports, and utility lines. Planning consists of scheduling the work to be done and selecting the most suitable construction methods and equipment for the project. Execution requires the timely mobilization of all drawings, layouts, and materials on the job to prevent delays to the work. Control consists of analyzing progress and cost to ensure that the project will be done on schedule and within the estimated cost.Keywords: planning, Execution, Control, Preparation of site,Earthmoving, Foundation treatment, Steel erection, Concrete construction, Asphalt paving and so on.Construction engineering is a specialized branch of civil engineering concerned with the planning, execution, and control of construction operations for such projects as highways, buildings, dams, airports, and utility lines.Planning consists of scheduling the work to be done and selecting the most suitable construction methods and equipment for the project. Execution requires the timely mobilization of all drawings, layouts, and materials on the job to prevent delays to the work. Control consists of analyzing progress and cost to ensure that the project will be done on schedule and within the estimated cost.Planning. The planning phase starts with a detailed study of construction plans and specifications. From this study a list of all items of work is prepared, and related items are then grouped together for listing on a master schedule. A sequence of construction and the time to be allotted for each item is then indicated. The method of operation and the equipment to be used for the individual work items are selected to satisfy the schedule and the character of the project at the lowest possible cost.The amount of time allotted for a certain operation and the selection of methods of operation and equipment that is readily available to the contractor. After the master or general construction schedule has been drawn up, subsidiary detailed schedules or forecasts are prepared form the master schedule. These include individual schedules for procurement of material, equipment, and labor, as well as forecasts of cost and income.Execution. The speedy execution of the project requires the project requires the ready supply of all materials, equipment, and labor when needed. The construction engineer is generally responsible for initiating the purchase of most construction materials and expediting their delivery to the project. Some materials, such as structural steel and mechanical equipment, require partial or complete fabrication by a supplier. For these fabricated materials the engineer must prepare or check all fabrication drawings for accuracy and case of assembly and often inspect the supplier’s fabrication.Other construction engineering duties are the layout of the work by surveying methods, the preparation of detail drawings to clarify the design engineer’s drawings for the construction crews, and the inspection of the work to ensure that it complies with plans and specifications.On most large projects it is necessary to design and prepare construction drawings for temporary construction facilities, such as drainage structures, access roads, office and storage buildings, formwork, and cofferdams. Other problems are the selection of electrical and mechanical equipment and the design of structural features for concrete material processing and mixing plants and compressed air, water, and electrical distribution systems.Control. Progress control is obtained by comparing actual performance on the work against the desired performance set up on the master or detailed schedules. Since delay on one feature of the project could easily affect the entire job, it is often necessary to add equipment or crews to speed up the work.Cost control is obtained by comparing actual unit costs for individual work items against estimated or budgeted unit costs, which are set up at the beginning of the work.A unit cost is obtained by dividing the total cost of an operation by the number of units in that operation.Typical units are cubic yards for excavation or concrete work and tons for structural steel. The actual unit cost for any item at any time is obtained by dividing the accumulated costs charged to that item by the accumulated units of work performed.Individual work item costs are obtained by periodically distributing job costs, such as payroll and invoices to various work item accounts. Payroll and equipment rental charges are distributed with the aid of time cards prepared by crew foremen. The cards indicate the time spent by the job crews and equipment on the different element of the work. The allocation of material costs is based on the quantity of each type of material used for each specific item.When the comparison of actual and estimated unit costs indicates an overrun; an analysis is made to pinpoint the cause. If the overrun is in equipment costs, it may be that the equipment has insufficient capacity or that it is not working properly. If the overrun is in labor costs, it may be that the crews have too many men, lack of proper supervision, or are being delayed for lack of materials or layout. In such cases time studies are invaluable in analyzing productivity.Construction operations are generally classified according to specialized fields. These include preparation of the project site, earthmoving, foundation treatment, steel erection, concrete placement, asphalt paving, and electrical and mechanical installations. Procedures for each of these fields are generally, the same, even when applied to different projects, such as buildings, dams, or airports. However, the relative importance of each field is not the same in all cases.Preparation of site. This consists of the removal and clearing of all surface structures and growth from the site of the proposed structure. A bulldozer is used for small structures and trees. Larger structures must be dismantled.Earthmoving. This includes excavation and the placement of earth fill. Excavation follows preparation of the site, and is performed when the existing grade must be brought down to a new elevation. Excavation generally starts with the separate stripping of the organic topsoil, which is later reused for landscaping around the new structure. This also prevents contamination of the nonorganic material whichis below the topsoil and which may be required for fill. Excavation may be done by any of several excavators, such as shovels, draglines, clamshells, cranes, and scrapers.Efficient excavation on land requires a dry excavation area, because many soils are unstable when wet and cannot support excavating and hauling equipment. Dewatering becomes a major operation when the excavation lies below the natural water table and intercepts the groundwater flow. When this occurs, dewatering and stabilizing of the soil may be accomplished by trenches, which conduct seepage to a sump from which the water is pumped out. Dewatering and stabilizing of the soil may in other cases be accomplished by wellpoints and electroosmosis.Some materials, such as rock, cemented gravels, and hard clays, require blasting to loosen or fragment the material. Blast holes are drilled in the material; explosives are then placed in the blast holes and detonated. The quantity of explosives and the blast-hole spacing are dependent upon the type and structure of the rock and the diameter and depth of the blast holes.After placement of the earth fill, it is almost always compacted to prevent subsequent settlement. Compaction is generally done with sheep’s-foot, grid, pneumatic-tired, and vibratory-type rollers, which are towed by tractors over the fills it is being placed. Hand-held, gasoline-driven rammers are used for compaction close to structures where is no room for rollers to operate.Foundation treatment. When subsurface investigation reveals structural defects in the foundation area to be used for a structure, the foundation must be strengthened. Water passages, cavities, fissures, faults, and other defects are filled and strengthened by grouting. Grouting consists of injection of fluid mixtures under pressure. The fluids subsequently solidify in the voids of the strata. Most grouting is done with cement and mixtures, but other mixture ingredients are asphalt, cement and clay, and precipitating chemicals.Steel erection. The construction of a steel structure consists of the assembly at the site of mill-rolled or shop-fabricated steel section. The steel sections many consist of beams, columns, or small trusses which are joined together by riveting, bolting, or welding. It is more economical to assemble sections of the structure at a fabricating shop rather than in the field, but the size of preassembled units is limited by the capacity of transportation and erection equipment. The crane is the most common type of erection equipment, but when a structure is too high or extensive in area to be erected by a crane, it is necessary to place one or more derricks on the structure to handle the steel. In high structures the derrick must be constantly dismantled and reerected to successively higher levels to raise the structure. For river bridges the steel may be handled by cranes on barges, or, if the bridge is too high, by traveling derricks which ride on the bridge being erected, Cables for long suspension bridges are assembled in place by special equipment that pulls the wire from a reel, set up at one anchorage, across to the opposite anchorage, repeating the operation until the bundle of wires is of the required size.Concrete construction. Concrete construction consists of several operations: forming, concrete production, placement, and curing. Forming is required to contain and support the fluid concrete within its desired final outline until it solidifies and cansupport itself. The form is made of timber or steel sections or a combination of both and is held together during the concrete placing by external bracing or internal ties. The forms and ties are designed to withstand the temporary fluid pressure of the concrete.The usual practice for vertical walls is to leave the forms in position for at least a day after the concrete is placed. They are removed when the concrete has solidified or set. Slip-forming is a method where the form is constantly in motion, just ahead of the level of fresh concrete. The form is lifted upward by means of jacks which are mounted on vertical rods embedded in the concrete and are spaced along the perimeter of the structure. Slip forms are used for high structures such as silos, tanks, or chimneys.Concrete may be obtained from commercial batch plants which deliver it in mix trucks if the job is close to such a plant, or it may be produced at the job site. Concrete production at the job site requires the erection of a mixing plant, and of cement and aggregate receiving and handling plants. Aggregates are sometimes produced at or near the job site. This requires opening a quarry and erecting processing equipment such as crushers and screens.Concrete is placed by chuting directly from the mix truck, where possible, or from buckets handled by means of cranes or cableways, or it can be pumped into place by special concrete pumps.Curing of exposed surfaced is required to prevent evaporation of mix water or to replace moisture that dose evaporate. The poper blance of water and cement is required to develop full design strength.Concrete paving for airports and highways is a fully mechanized operation. Batches of concrete are placed between the road forms from a mix truck or a movable paver, which is a combination mixer and placer. A series of specialized pieces of equipment, which ride on the forms, follow to spread and vibrate the concrete, smooth its surface, cut contraction joints, and apply a curing compound.Asphalt paving. This is an amalgam of crushed aggregate and a bituminous binder. It may be placed on the roadbed in separate operations or mixed in a mix plant and spread at one time on the roadbed. Then the pavement is compacted by rollers.建筑工程摘要：建筑工程是土木工程的一个专业分支，涉及对诸如高速公路、建筑物、水坝、机场和公用事业管线项目的规划、实施和控制。
外文翻译
南京工程学院毕业设计外文翻译（原文及译文）原文名称：Features and Kernels for Audio EventRecognition课题名称：绝缘缘破坏放电声事件检测算法学生姓名孙冉学号208130506指导教师王青云所在系部通信工程学院专业名称电子信息工程2017年 3 月7日Features and Kernels for Audio Event RecognitionAbstractOne of the most important problems in audio event detection research is absence of benchmark results for comparison with any proposed method. Different works consider different sets of events and datasets which makes it difficult to comprehensively analyze any novel method with an existing one. In this paper we propose to establish results for audio event recognition on two recent publicly-available datasets. In particular we use Gaussian Mixture model based feature representation and combine them with linear as well as non-linear kernel Support Vector Machines.Index Terms: Audio Event Detection, Audio Content Analysis.1. IntroductionIn recent years automatic content analysis of audio recordings has been gaining attention among the audio research community. The goal is to develop methods which can automatically detect the presence of different kinds of audio events in a recording. Audio event detection (AED) research is driven by its application in several areas. These include areas such as multimedia information retrieval or multimedia event detection[1] where the audio component contains important information about the content of the multimedia data. This is particularly important for supporting content-based search and retrieval of multimedia data on the web. Other applications such as surveillance [2], wildlife monitoring [3] [4], context aware systems [5] [6], health monitoring etc. are also motivating audio event detection research. A variety of methods have been proposed for AED in the last few years. A GMM-HMM structure, similar to that used in automatic speech recognition was presented in [7]. A simple yet effective approach is the bag of words representation [8][9] [10]. This approach also features in audio event detection in noisy environments [11]. Given the complexity of audio event detection, deep neural networks which are known for modeling highly non-linear functions have also been explored for AED [12] [13]. Some other interesting approaches involve use of acoustic unit descriptors for AED [14], spectral exemplars for AED [15], and matrix factorization of spectral representations [16]. Lack of supervised data for audio events has also led to some weakly supervised audio event detection works [17] [18].All of the methods have been shown to be effective and reasonable upto an extent on datasets and events on which they were evaluated. However, a major problem is that we fail to understand how they compare against each other. Most of them use different datasets or sets of sound events and given the variations we encounter across datasets and more importantly across sound events it is difficult to assess different methods. Another factor lacking in several of the current works is analysis of the proposed method on a reasonably large vocabulary of sound events. Usually, only small sets of audio events are considered. The absence of publicaly available An important reason for these concerns in current AED literature is theabsence of publicly available standard sound event databases which can serve as a common testbed for any proposed method. More importantly, the dataset should be of reasonable size in terms of total duration of audio and in terms of the vocabulary of events. Open-source dataset helps in standardization of research efforts on a problem and give a better understanding of results shown by any proposed method. For example, in image recognition and classification results are prominently reported and compared on well known standard datasets such as PASCAL [19], CIFAR [20] and Image Net[21].Recently, the audio community have attempted to address this problem by creating standard sound event databases meant for audio event detection tasks [22] [23] [24]. The UrbanSounds8k [22] dataset contains 10 different sound events with a total of about 8.75 hours of data. The ESC-50 [23] dataset contains labeled data for 50 sound events and a total of 2.8 hours of audio. An important goal of our work is to perform a comprehensive analysis of these two datasets for audio event detection. These two public datasets addresses event detection concerns in terms of both total duration and vocabulary of sound events. The total duration of the audio is reasonably large and overall they allowed us to present results on 56 unique sound events. Specifically, we analyze Gaussian Mixture Models (GMMs) for feature representation. We inspect two forms of feature representations using GMMs. The first one is a soft-count histogram representation obtained using a GMM and the second uses maximum a posteriori (MAP) adaptation of GMM parameters to obtain a fixed-dimensional feature representation for audio segments. The first feature is similar to the bag of audio words representation. The second feature obtained through MAP adaptation tries to capture the actual distribution of low-level features of a recording. On the classifier side we use Support Vector Machines (SVMs) where we explore different kernels for audio event detection.The ESC-50 dataset also allowed us to study audio event detection at different hierarchies. The 50 events of ESC-50 broadly belong to 5 different higher semantic categories. We report and analyze results for both lower-level events and higher semantic categories.It is important to note that our analysis on these two datasets is very different from the simple multi-class classification analysis done by the authors of these datasets. In particular, we are interested in audio event detection where our goal is to detect presence or absence (binary) of an event in a given recording. Our analysis is in terms of ranked measure (AP) for each event as well as area under detection error curves for each event.The rest of the paper is organized as follows. In Section 2 we provide a brief description of ESC-50 and Urban Sounds8k datasets. We describe the features and kernels used for audio event detection in Section 3 and report results in Section 4. Section 5 concludes our work.2. DatasetsUrban Sounds8k: This dataset contains a total of 8732 audio clips over 10 differentsound event classes. The event classes are air conditioner (AC), car horn (CH), children playing (CP), dog barking (DB), drilling (DR), engine idling (EI), gun shot (GS), jackhammer (JK), siren (SR) and street music (SM). The total duration of the recordings is about 8.75 hours. The length of each audio clip is less than or equal to 4 seconds. The dataset comes with a presorted 10 folds and we use the same fold structure in our experiments.ESC-50: ESC-50 contains a total of 2000 audio clips over 50 sound events. Examples sound classes are dog barking, rain, crying baby, fireworks, clock ticking etc . The full list of sound events can be found in [23]. The entire set of 50 events in the dataset belong 5 groups, namely , animals (e.g dog, cat), natural soundscapes (e.g rain, seawaves), non-speech human sounds (e.g sneezing, clapping), interior domestic sounds (e.g door knock, clock), and exterior sounds (e.g helicopter, siren). The total duration of the recordings in this dataset is around 2.8 hours. We divide the dataset into 10 folds for our experiments. For future use of our setup by others, the details of this fold structure is available on this webpage [25].For both datasets we convert all audios into single channel and resample all audio to 44.1KHz sampling frequency.3. Features and Kernels3.1. FeaturesBefore we can go into the details of different feature representation we need low-level characterization of audio signals. We use Mel-Frequency Cepstral Coefficients (MFCC) as low-level representations of audio signals. Let these D-dimensional MFCC vectors for a recording be represented as t x , where t = 1 to T , T being thetotal number of MFCC vectors for the recording.Broadly, we employ two higher level feature representations for characterizing audio events. Both of these features attempt to capture the distribution of MFCC vectors of a recording. We will refer to these as α and β features and the subtypes will be represented using appropriate sub-scripts and superscripts.The first step to obtain higher-level fixed dimensional representation is to train a GMM on MFCC vectors of the trainingdata. Let us represent this GMM by G = {k ω,N(k μ,k ∑),k =1 to M}, where k ω, k μ and k ∑are the mixture weight, meanand covariance parameters of the th k Gaussian in G. We will assume diagonal covariance matrices for all Gaussians and k σwill represent the diagonal vector of k ∑.3.2. KernelsAs stated before we use SVM [29] classifiers. We analyze different kernel functions for training SVMs on the above features.The first one is linear kernel (LK). The other two kernels canbe described in a general form for feature vectors f and 'fby K (f ,'f ) = ),('exp f f D γ-. For 22''||, || ),(f f f f D = we obtain the radial basisfunction kernel (RK). For histogram features (αfeatures) exponential Chi-square distance kernels (CK) are known to work well for vision tasks [30]. The Chi-squaredistance [31] is given by ∑=+-=p i i i i i f f f f f f D 1'2'')( ),(, where p is dimensionality of feature vectors f . We use this kernel only for histogram features (α). Linear and RBF kernel are used for all features. It is worth noting that M s βhave been designedspecifically for linear kernels. In our experiments we try M s βwith RBF kernels as well.4. Experiments and ResultsWe report results on both Urban Sounds8k and ESC-50 datasets.First, we divide the datasets into 10 folds. Urban Sounds8k already comes with presorted folds and we follow the same foldstructure. 9 folds are used as training data and the 10th fold isused for testing. This done in all 10 ways such that each foldbecomes test data once. This allows us to obtain results on theentire dataset and aggregate results over entire dataset are reported here. 20-dimensional MFCC vectors are extracted foreach audio recording using 30 ms windows and a 15 ms hopsize. We experiment with 4 different values of number of mixture components (M ) in GMM G . These are 32, 64, 128, 256.The relevance factor r is set to 20 for all experiments. For training SVMs we use LIBSVM [32]. The slack parameter in SVMtraining and the parameter γfor RBF and Chi-Square kernel isobtained by performing cross validation over the given trainingdata.ESC-50: Table 1 shows MAP values over all 50 eventsof ESC-50 . It shows MAP for different M values, features and kernels. Table 2 shows the mean AUC values fordifferent cases. We can observe that for αfeatures, exponential Chi-square kernel (CK) is much superior compared to linear (LK) and RBF (RK) kernels. There is an absolute improvement of 20-30% in MAP using the exponential Chi-square kernel. The MAP also improves with increasing M. Among the βfeatures M s βseems to be the best variant. Adding adapted variance k ^σleads to inferior performance. Both M σβand M s σβperform poorly compared to other features. Between M s βand M β, the latter is clearlysuperior for any given kernel and M . An important point worth noting is thatvariance for M βfeatures the performance goes down as M increases. This is notobserved for M s βwhere MAP is more or less constant across all M . For M s βlinear kernels are in general better than RBF kernels, although, the difference is small.Figure 1: ESC-50:Event AP values for 128α+ CK (BLUE) and 128s β+ LK(RED), Bottom Right Corner - MAP over HigherSemantic CategoriesFigure 1 (bottom right) also shows MAP over each higher semantic category. It can be observed that animal sounds are the easiest to detect and interior sounds are the hardest. The maximum difference between αand βfeatures is for non-speech humanUrban Sounds8k: Table 3 and Table 4 shows MAP and MAUC values respectively on the Urban Sounds8k dataset for different features and kernels. M σβand M s σβhave been removed from analysis due to their inferior performance. Once again we notice that αfeatures with the exponential Chi-square kernel is in general the best feature and kernel pair. Interestingly, in this case M αwith CK is better than all βfeatures for lower M values as well. At low M MAP gain is between 3−8%whereas for large M the difference goes upto 16−17% in absolute terms. The primary reason for this is the fact that for βfeatures performance goes down as we increase M whereas for αfeatures the performance goes up. Among the βfeatures,M s βis ingeneral better than M β, though, for M = 32 and M = 64 under suitable kernel both features give more or less similar performance. However, we had previously observed for the ESC-50 dataset that M s βis always better than M β. Another new observation forM s βon the Urban Sounds8k dataset is that MAP goes down (MAUC goes up) withincreasing M which was not the case in ESC-50.Figure 2: Urabn Sounds8k: Left - Event AP values, Right-Event AUC values (128α+ CK (BLUE) and 128s β+ LK (RED)).Figure 3: ESC-50 and Urban Sounds8k: Left - Event AP values,Right-Event AUC values (128α+ CK (BLUE) and 128s β+ LK (RED)).Event wise results (AP and AUC) for 10 events of Urban-Sounds8k is shown in Figure 2. The event names have been encoded in the figure as per the description in Section 2. Events like dog barking, Siren and Gun Shot are again easy do detect. The source of the original audios for both datasets is Freesound [34]. This allows us to make general comments for events which are common to both datasets, namely, Car Horn, Dog bark, Engine and Siren. The results are shown in Figure3. We observe that we do well on Dog barking on both datasets whereas complex events such as Engine are inherently difficult to detect and poor performance is obtained on both datasets.5. ConclusionsIn this work we have performed a comprehensive analysis of audio event detection using two publicly available datasets. Overall, we presented results for 56 unique events across the two datasets. Our analysis included two broad sets of features and overall 5 different feature types. The features are GMM-based representations. The classifier side included SVMs with 3 different kernel types.ESC-50 dataset also allowed us to study audio event detection at lower and higher semantic levels. We observed that events in the Animal group are in general easier to detect. Interior sound events which includes events such as Washing Machine, Vacuum Cleaner, Clock Ticking are relatively harder to detect. Together both datasets allowed us to study audio event detection in an exhaustive manner for the features and kernels used in this work. This is potentially very helpful in standardizing AED methods and results.6. References[1] P. Over, G. Awad, M. Michel, J. Fiscus, W. Kraaij, A. Smeaton, G. Quenot, and R. Ordelman, “Trecvid 2015–an overview of the goals, tasks, data, evaluation mechanisms and metrics,” in Proceedings of TRECVID 2015. NIST, USA, 2015.[2] G. Valenzise, L. Gerosa, M. Tagliasacchi, F. Antonacci, and A. Sarti, “Scream and gunshot detection and localization for audio-surveillance systems,” in Advanced Video and Signal Based Surveillance, IEEE Conference on, 2007, pp. 21–26.[3] D. Stowell and M. D. Plumbley, “Automatic large-scale classification of birdsounds is strongly improved by unsupervised feature learning,”Peer J, 2014.[4] J. Ruiz-Muñoz, M. Orozco-Alzate, and G. Castellanos-Dominguez, “Multiple instance learning-based birdsong classification using unsupervised recording segmentation,”in Proceedings of the 24th Intl. Conf. on Artificial Intelligence, 2015.[5] D. Battaglino, A. Mesaros, L. Lepauloux, L. Pilati, and N. Evans, “Acoustic context recognition for mobile devices using a reduced complexity svm,”in Signal Processing Conference (EUSIPCO), 2015 23rd European. IEEE, 2015, pp. 534–538.[6] A. J. Eronen, V. T. Peltonen, J. T. Tuomi, A. P. Klapuri, S. Fagerlund, T. Sorsa, G. Lorho, and J. Huopaniemi, “Audio-based context recognition,”Audio, Speech, and Language Processing, IEEE Transactions on, vol. 14, no. 1, pp. 321–329, 2006. [7] X. Zhuang, X. Zhou, M. A. Hasegawa-Johnson, and T. S. Huang, “Real-world acoustic event detection,”Pattern Recognition Letters, vol. 31, no. 12, pp. 1543–1551, 2010.[8] S. Pancoast and M. Akbacak, “Bag-of-audio-words approach for multimedia event classification,”in Interspeech, 2012.[9] A. Plinge, R. Grzeszick, G. Fink et al., “A bag-of-features approach to acoustic event detection,”in IEEE ICASSP, 2014.[10] A. Kumar, R. Hegde, R. Singh, and B. Raj, “Event detection in short duration audio using gaussian mixture model and random forest classifier,”in Signal Processing Conference (EUSIPCO), 2013 Proceedings of the 21st European. IEEE, 2013, pp. 1–5.[11] P. Foggia, N. Petkov, A. Saggese, N. Strisciuglio, and M. Vento, “Reliable detection of audio events in highly noisy environments,”Pattern Recognition Letters, vol. 65, pp. 22–28, 2015.[12] K. Ashraf, B. Elizalde, F. Iandola, M. Moskewicz, J. Bernd, G. Friedland, and K. Keutzer, “Audio-based multimedia event detection with DNNs and sparse sampling,”in Proc. of the 5th ACM International Conference on Multimedia Retrieval, 2015. [13] O. Gencoglu, T. Virtanen, and H. Huttunen, “Recognition of acoustic events using deep neural networks,”in Signal Processing Conference (EUSIPCO), 2014 Proceedings of the 22nd European. IEEE, 2014, pp. 506–510.[14] A. Kumar, P. Dighe, R. Singh, S. Chaudhuri, and B. Raj, “Audio event detection from acoustic unit occurrence patterns,”in IEEE ICASSP, 2012, pp. 489–492. [15] X. Lu, Y. Tsao, S. Matsuda, and C. Hori, “Sparse representation based on a bag of spectral exemplars for acoustic event detection,”in IEEE ICASSP, 2014, pp. 6255–6259.[16] A. Mesaros, T. Heittola, O. Dikmen, and T. Virtanen, “Sound event detection in real life recordings using coupled matrix factorization of spectral representations and class activity annotations,”in ICASSP. IEEE, 2015, pp. 606–618.[17] A. Kumar and B. Raj, “Audio event detection using weakly labeled data,”in 24th ACM International Conference on Multimedia. ACM Multimedia, 2016.[18] ——, “Weakly supervised scalable audio content analysis,”in 2016 IEEE International Conference on Multimedia and Expo (ICME). IEEE, 2016.[19] M. Everingham, L. Van Gool, C. K. I. Williams, J. Winn, and A. Zisserman, “The PASCAL Visual Object Classes Challenge 2011 (VOC2011) Results,”/challenges/VOC/voc2011/workshop/index.html. [20] A. Krizhevsky and G. Hinton, “Learning multiple layers of features from tiny images,”2009.[21] J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,”in Computer Vision and Pattern Recognition, 2009. CVPR 2009. IEEE Conference on. IEEE, 2009, pp. 248–255. [22] J. Salamon, C. Jacoby, and J. P. Bello, “A dataset and taxonomy for urban sound research,”in Proceedings of the ACM International Conference on Multimedia. ACM, 2014, pp. 1041–1044.[23] K. J. Piczak, “Esc: Dataset for environmental sound classification,”in Proceedings of the 23rd Annual ACM Conference on Multimedia Conference. ACM, 2015, pp. 1015–1018.[24] A. Mesaros, T. Heittola, and T. Virtanen, “Tut database for acoustic scene classification and sound event detection,”in 24th European Signal Processing Conference, 2016.[25] A. Kumar. Aed. /%7Ealnu/AED.htm Copy and Paste in browser if clicking does not work.[26] J. Gauvain and C. Lee, “Maximum a posteriori estimation for multivariate gaussian mixture observations of markov chains,”Speech and audio processing, IEEE Trans. on, 1994.[27] F. Bimbot and et al., “A tutorial on text-independent speaker verification,”EURASIP journal on applied signal processing, vol. 2004, pp. 430–451, 2004. [28] W. Campbell, D. Sturim, and A. Reynolds, “Support vector machines using gmm supervectors for speaker verification,”Signal Processing Letters, IEEE, pp. 308–311, 2006.[29] T. Hastie, R. Tibshirani, J. Friedman, and J. Franklin, “The elements of statistical learning: data mining, inference and prediction,”The Mathematical Intelligencer, vol. 27, no. 2, pp. 83–85, 2005.[30] J. Zhang, M. Marszałek, S. Lazebnik, and C. Schmid, “Local features and kernels for classification of texture and object categories: A comprehensive study,”International journal of computer vision, vol. 73, no. 2, pp. 213–238, 2007.[31] P. Li, G. Samorodnitsk, and J. Hopcroft, “Sign cauchy projections and chi-square kernel,”in Advances in Neural Information Processing Systems, 2013, pp. 2571–2579.[32] C.-C. Chang and C.-J. Lin, “Libsvm: a library for support vector machines,”ACM Transactions on Intelligent Systems and Technology (TIST), vol. 2, no. 3, p. 27, 2011.[33] A. Martin, G. Doddington, T. Kamm, M. Ordowski, and M. Przybocki, “The det curve in assessment of detection task performance,”DTIC Document, Tech. Rep., 1997.[34] Free Sound, “https:///,”2015.原文出处：音频事件识别特征核摘要音频事件检测研究中最重要的问题之一就是没有基准结果与任何提出的方法进行比较。
英文文献
附录：英文技术资料翻译英文原文：PowerBuilder Brief IntroductionPowerBuilder is a popular database application development tool manufacturer which is called PowerSoft introduced product （PowerSoft has been purchased by the database manufacturer called Sybase）, it entirely designed in accordance with the client / server architecture, in the client / server architecture, it uses the client, as a database application development tool to exist. Because of using object-oriented and visualization technologies, providing visual application development environment in PowerBuilder, allows us to be developing data and database management system database applications in the background server easily and quickly.At present, network technology is developing rapidly, with the development of the OLE, OCX, cross-platform technology, and in the latest version of PowerBuilder provides the full support of these technologies. In a word, in the database development tools, PowerBuilder is a very good product of them, we can use it to develop powerful database applications.PowerBuilder provides support in most of the current popular relation database management systems, because in PowerBuilder applications for database access part of the general international standard database query language SQL, makes using PowerBuilder development of the application can be used in different background of database management systems without revision or do a small amount of revision. That is to say, with PowerBuilder development of the application server is independent of the database management system.As the most of the windows application procedures, PowerBuilder is also event driven of the method to work. In this work method, programfor the operation has no fixed processes, the code in the program is also possible programmed for the preparation of the events, when the program began running, it can be accepted the events that are from the system, users or other applications trigger, and then execute the relevant code. Event driven methods of work is closely related to the object-oriented technology. In the PowerBuilder application, acceptance of the events is often the various visual interface object in the program.PowerBuilder is an object-oriented development tool, various common application windows, menus, and other controls in Windows are one and one targets in PowerBuilder. We can also create own user object in PowerBuilder. We should point out that the PowerBuilder provides object-oriented method of comprehensive technical support, we can make use of object-oriented methods in the object of encapsulation, inheritance, polymorphism, and other features make the our development applications have great reusability and scalability, which is on the application in the process of the important goals in software engineering.At present, due to the development of network technology, and many different types of operating system platforms are using on the internet at the same time, this is a higher request for the development of the cross-platform application, and PowerBuilder provides a good cross-platform, for example, in PowerBuilder, using Windows platform for the development of all kinds of objects can be easily applied to the UNIX platform, because PowerBuilder supports cross-platform object. This allows the applications from one platform to another platform to become not complicated.In order to provide the user all aspects of supports, PowerBuilder has its own programming language which is called power script, in addition to providing the basic flow control statements, but also toprovide hundreds of functions to manipulate objects and all kinds of information supports such as DDE, OLE, and other aspects in this language . Furthermore, we can also define our own functions and deal with specific incidents. When we learn PowerBuilder, a part of the time is to understand and familiar with the functions in PowerBuilder.In PowerBuilder, a great feature is to propose the concept of data window. Data window object is one of the objects in PowerBuilder, it is different from other objects, because data window is specifically targeted to visit background database services, in the data window, we define data sources and data display style, so that we will be able to concentrate fully on the operation of process control procedures in application procedures, and couldn’t concern about data sources, because we have been well defined data sources in the data window. If you need to use different data in the database as long as the data is revised target window on it. In particular, we should point out that the PowerBuilder provides a wealth of data display mode in data window object, and this can meet a variety of needs.In a new version of PowerBuilder, it provides a foundation class library PFC, it provides many reusable predefined classes and objects for application procedures of development, using the basic class library PFC can enable the rapid development of high-quality reuse of good application procedures. It is really rolling power of object-oriented programming.Finally we should point out that the there are three different versions: desktop type, professional type, and enterprise type. Desktop type is for personal using of the desktop applications, where you can use the built-in PowerBuilder database management system Sybase SQL anywhere to create and use of local databases, the application for personal services; professional type of PowerBuilder most important point is to provide the support of Microsoft ODBC （databaseconnectivity standard interface）,in professional version of PowerBuilder, we can use the inheritance, polymorphism and more features; in the enterprise versions of the most important point is to provide for the development of large-scale database applications and comprehensive support, the development of large-scale applications of the many auxiliary tools, such as C++, class builder and so on.Database front-end development tool and the background database management system in a way is a very important issue. PowerBuilder provides two ways to visit the background database, the one way is standard interface through ODBC, and the second is through a dedicated interface with the database linked to the background.ODBC is called Open Database Connectivity in Chinese, it is the database connectivity standards which is proposed by Microsoft, the use of ODBC database connectivity of the first step is creating data sources, for example, we can use windows control panel under the "32 BIT ODBC" election of the drive to create a database of data sources, after finishing creating data sources, we will be able to use computers in the local definition of good background data source to access to the data in the database.We can also link to the background through a dedicated interface with the database, because dedicated interface is designed against the background of a specific database management system, by this way to access to data faster than using ODBC way to access to data. If our application is only for specific types of background database, of course, using a special interface to background information will be more faster.中文译文：PowerBuilder简要介绍PowerBuilder是著名的数据库应用开发工具生产厂商PowerSoft公司推出的产品（PowerSoft现已被数据库厂商Sybase所收购），它完全按照客户机/服务器体系结构研制设计，在客户机/服务器结构中，它使用在客户机中，作为数据库应用程序的开发工具而存在。
南京工程学院专业英语复习整理.docx
圳闰滑lubricate(v.) lubrication(n・) 冷去卩cool (v.) cooling(n..) 单件小批量small batches or one-of-a-kind items 机床ma ch ine cool 算初L数控computer! zed numeri cal co门trol 电火花力口工机elec trical discharge machine 车床turning machine 铳床mil ling ma ch ine 计算机辅助设计computer aided design 计算机辅助制造computer ai ded manufacturi ng 成品件finished part 正转/反转forward/reverse direction 轴axis axes (复数)台式钻床drill press 刀库tool magazine 自动换刀装置automatic tool changer (ATC) 进给速度feed rate 力口工中心ma ch ining center 分布式数字控制distributive numerical control (DNC) 绝对的absolute 增量的increniental 普通机床conventional machine tool 滚珠丝杠ball screw 丝杠lead screw 直角坐标系rectangular coordinate system 极坐标系polar coordinate system 编程零点program zero (work zero) 驱动电机drive motor 导轨way 床头箱headstock 尾架tailstock 转塔刀架turret 粗车rough turning 精镇finish boring 夹具work hoi di ng devi ce 装酉己图assembly drawing 零件图part draAving 冋参考点，冋零点zero return 机床厂，机床制造商machine tool builder (MTB) 主轴档位，主轴齿轮级spindle range 切槽加工grooving 轮廓铳削contour milling 精纟堂加工finish boring 凹槽铳削neck milling螺纹加工threading平面铳削face milling 浅孔钻short-hole drill 直线插补linear interpolation 圆弧插补circular i门terpolation 螺纹铳削thread milling 加工循环machining cycle 快速定位rapid motion1•数控机床包括车床、多轴机床、铳床和电火花加工机床等，只不过以前有人工完成的功能现在由计算机控制模块完成。
“一带一路”倡议下中医典籍翻译术语库的构建标准、原则和技术
第20卷第3期2020年9月南京工程学院学报（社会科学版）Journal of Nanjing Institute of Technology! Social Science Edition)V 〇1.20,N 〇.3 S e p.，2020d o i : 10. 13960/j. i n . 2096 -238X . 2020. 03. 008 投稿网址：http ://x b . n jit. edu. cn“ 一带一路”倡议下中医典籍翻译术语库的 构建：标准、原则和技术朱姝颖(南京工程学院外国语学院，江苏南京，211167)摘要：中医典籍的翻译促进中医药逐渐走向“一带一路”沿线国家，其翻译术语库构建能规范外译并提升中医话语权。
构建中医典籍翻译术语库应首先确定标准和原则，进而运用相关技术进行术语提取，运用工具构建术语库。
关键词：$—带一路”倡议；中医典籍翻译;翻译术语库；术语提取中图分类号:H 315.9“一带一路”合作倡议下，中医药的发展、传播 和共享，体现了“ 一带一路”倡议的互利共赢精神，推动了人类命运共同体的建设。
中医典籍作为中 国医学发展的基础，其外译能使中医药从走向“一 带一路”沿线国家到走入这些国家。
而中医典籍翻译术语库的构建能规范典籍外译并促进“一带一 路”语言服务的发展。
本文首先分析中医典籍翻译 术语库构建标准、原则，为术语的确定和翻译术语 库的建设提供依据，进而探讨术语提取的方法和翻 译术语库构建的工具，提供构建中医典籍翻译术语 库的思路。
一、国内中医典籍术语的研究概况近十年，国内学者对典籍术语的研究仍侧重翻 译策略，对中医典籍术语的研究也不例外。
例如，蒋 学军从文化图式角度探讨了中医典籍中部分术语的翻译策略[1];吴纯瑜、张斌、潘大为从各自角度分析 《黄帝内经》中相关术语的翻译原则或翻译方法[2_3];甘胜男强调翻译中医药术语时需考虑语言、交际、相 关哲学、文化等多方面因素[4];江楠探讨中医英语翻 译在术语、语句、语篇方面的策略[5];李成华、孙慧明 等从文化自信的角度，论述民族性原则、音译法、直 译法、异化翻译等能够有效传递中医文化[6]。
典籍翻译中的“根”自源说
第20卷第3期2020年9月南京工程学院学报（社会科学版）Jo u r na l o f N a n ji ng I n stit u t e o f T ec hn olo gy! S oci a l S cie n c e Ed itio n)V〇1.20,N〇.3S e p.，2020d o i: 10. 13960/j. i n. 2096 -238X. 2020. 03. 001 投稿网址：http ://x b.n jit. edu. cn典籍英译研究特约主持人:朱源教授主持人语：自从2002年首届全国典籍英译研讨会暨中国英汉语比较研究会典籍英译专业委员会成立以来，在国家中国 文化走出去的号召下，我国的典籍英译事业方兴未艾。
国内诸多院校成立了典籍英译研究所，开设了典籍英译课程，招收了 典籍英译研究方向硕博研究生。
相关的学术研讨会以及论文专著不断增加。
本栏目选取英语语言文学专业典籍英译研究方向的论文四篇。
魏建国的论文《典籍翻译中的“根”自源说》颇具创意，通 过构建典籍英译批评术语“根词”的概念，强化了核心语词以及相互的逻辑关系在哲学典籍英译阐释中的特别作用和方法，显示出作者较强的抽象逻辑思维建构能力。
同时作者也提供了代表性译例，努力做到理论与实践的统一。
显然，此建构也并 非空穴来风，而是受到了结构主义基本理念的影响。
希望作者进一步深人研究相关经典译本，拓展理论视野，争取在典籍英 译批评领域构建更全面系统的概念及与之相适应的术语。
呼媛媛的《中国先唐民歌英译述评》选题有新意，综述全面细致，并且对主要译本及相关概念进行了一定的评析和界定。
按照文类框架研究典籍英译可谓司空见惯，但选择先唐民歌英译作 为研究对象具有特别的视角和学术价值。
作者来自延安，对于西北民歌及英译研究情有独钟，扩展至先唐民歌英译研究具有 比较坚实的基础。
相信作者沿着这条学术之路深人下去，定会发掘出新的译品，归纳出新的规律，丰富典籍英译的研究领域。
- 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)。
在小尺寸精密测量上改进的最佳最佳极限值分割法Fu Jihua, Wang Jianjun, Li Zhitao, Wu Ronghui地壳运动学院, 中国地震局,中国北京fujh@Wang Zhongyu仪器科学与光电工程学院, 北京航天航空大学,中国北京mewan@摘要——为了提高小尺寸测量的精度和效率，一种改进的最佳极限分割方法被提出来。
传统的最佳极限方法可以通过计算丢失的分类可能性用于图像分割。
在很强的噪声背景情况下，丢失的分类可能性分布是未知的。
最佳极限值获得的倾向于局部最佳方法。
所有的这些都会导致测量错误。
为了解决问题，一个过度分割系数引进最佳极限方法中来控制局部最优解决方案。
由引进的过度分割系数导致的过度分割现象将会由小尺寸测量系统来补偿。
通过仿真和实验分析，这种方法可以实施小尺寸精密的测量，而且很轻易就可以处理，有小运算量和给力的实时特点。
关键词：图像分割；最佳极限值；过度分割；小尺寸精密测量；机器视觉I.引进最小尺寸的测量是一种几何平均值范围的精确测量，机器视觉方法被应用到提高测量系统的灵活性，自动化性能和效率。
在机器视觉方法中，图像拆分是目标寻找和实物测量的关键技术之一。
很多学者在近几年中已经完成了他们图像拆分的努力。
具有分水岭的方法是在数学表面形态上研究出一个有力的拆分工具，但是在强大的背景噪音环境下，这种方法也许会经常引起一种称为过度拆分的问题[1]。
表面形态的方法被用来实现图像拆分过程中降低噪音[2]。
尽管这种表面形态法也能够提高原图的对比，但是这种在结构上不正确的选择也许会引起目标的变形和测量的错误。
一些新方法已经成功地应用到图像拆分，例如神经网络，微波和区域增长[3-5]。
然而，这些方法通常是复杂而且不适合现场测量。
对于少量的计算和很强的实时特性，最理想的极限方法是广泛使用图像拆分[6]。
不幸的是，在错过的可能分类上的分布是未知的，最理想的极限方法是很难为很大的噪声原因做局部图像分析的。
为了弥补理想极限方法的缺陷，一个改良的理想极限方法被提出用于现场小尺寸的精密测量。
II.理论和原理A.最佳的极限分割评估错失的分类可能性，最佳的极限分割方法计算最佳分割来分离目标和背景[7]。
目标的尺寸对分割结果有小的影响。
因此，最佳的极限分割方法在小尺寸精密测量中广泛使用。
通过最佳极限分割方法，目标和背景的原始图像像素通常假定其遵循平常的分布，并且他们的可能密度函数是P1(x）和P2（x），他们同时在图1中显示。
1μ 2μ 像素灰度目标图像灰度的平均值和方差是μ1和σ1,背景的图像灰度的平均值和方差是μ2和σ2。
目标像素占的百分比是Q ，背景的像素占的百分比是（1-Q ），并且可能的混合密度可以由 P(X)=QP 1(x)+(1−Q)P 2(x)=()()⎥⎦⎤⎢⎣⎡+⎥⎦⎤⎢⎣⎡21221212112-x -exp 2Q -12-x -exp 2Q σμσσμσππ （1） 如果选择的极限是T ，那么目标像素被错误地分类为背景像素的可能性是E 1（T ）=⎰∞T 1dX x P ）（ （2） 背景像素被错误地分类为目标像素的可能性是E 2（T ）=⎰∞T-2dX x P ）（ （3） 全部错误的可能性是E （T ）=）（）（）（T E Q -1T QE 21+。
（4） 当E （T ）/T 等于零的时候，全部错误的可能性达到了最小值。
并且最佳极限可以通过公式 22222121122-T -2-T -Q -1Q ln σμσμσσ）（）（）（= （5）获得。
B.最佳极限分割的迭代法对于大多数的图像，目标和背景的灰度值可以被看做跟随正态分布。
但是它们的优先性不能够预先决定。
因此，最佳极限很难通过理论Eθ来获得（5）。
对于快速算法，最佳极限可以通过在[8]中提到的迭代法来获得。
迭代法的步骤在下面演示：1）如果最小和最大灰度值是min Z 和max Z ，然后设置初始极限值T 0=2Z Z max min ）（+ （6）2）通过极限值T κ把原始图片分离为目标和背景。
计算他们的平均灰度值μA 和μB 。
这里μA = ∑∑≤≤k k T j i x Tj i x j i N j i N j ),(),(),(),(*,i x ）（ （7）∑∑=K K T j i x T j i x B j i N j i N j i x ),(),(),(),(*),(μ （8）这里，),(j i x 是像素灰度值，j)N(i,是比重系数3）计算新极限T κ+1=（μA + μB ）/2 （9） 4) 如果|T κ-T κ+1|x ≤，那么就停止迭代过程，那么T K 就是最佳极限，x 给出了无穷小的数字来控制循环的精确性。
否则，将T κ+1赋值给T κ并跳转到步骤2），继续迭代过程。
X. 改进最佳极限分割在很强的背景噪点中，错失的分类中的可能性分布也将会被影响到。
最佳极限将趋于一个有局限性的最佳方法。
为了控制局限性的最佳方法，背景像素被错误的分类为目标像素的可能性应该要减少。
结果是，目标像素被错误地分类为背景像素将会增加，并且这将会引起过度分割的程度。
幸运的是，过度分割可以被控制在很小的程度上，而且可以由测量系统的校核来弥补。
为了实现改进最佳极限分割和简化计算，唯一的改进是被下列取代迭代法Eq.(9)的方程式： T k+1=B A -1μλλμ）（+ （10）这里 λ 是过度分割的系数，并且 [)1,5.0∈λ 。
如果 λ 比0.5大，那么背景像素被错误的分类为目标像素的可能性会稍微减少点。
当 λ 等于0.5，Eθ.(10)是和Eθ.（9）一样的。
III. 仿真和实验分析A ．不同分割方法的操作比较为了显示改进的最佳极限方法，这种方法和原来的被同时用来做图像分割实验。
并且其中之一的处理结果在表1中列出来。
原始图像是混凝土裂缝的图像被640*480分辨率的CCD 摄像头捕捉到。
在实验中，混凝土裂缝是所要的目标。
在分割处理后，只有目标能留下来，获得了一个二进制图像。
（a）CCD摄像头获得的原始图像（b）最佳极限方法获得的二进制图像（c）改进的最佳极限方法获得的二进制图像在表1中的图（a）显示了被CCD摄像头捕获的原始图片。
由于测量环境的限制和测试对象，捕获的图像包含了很多干涉，例如测量噪点，纹理特征和其他非目标因素。
图像分割的任务是得到不失真的混凝土裂缝，最大化控制干涉。
图（b）是通过最佳极限分割法获得的混凝土的二进制图像。
正如图（b）显示的，由于最佳极限方法的缺陷，也有很多的非目标物存在里面。
图（c）是通过改进的最佳极限分割法获得的混凝土裂缝二进制图像，它的过度分割系数 =0.75。
通过比较由改进的最佳极限分割法获得的二进制图像，它几乎不包含非目标物。
它显示出改进的最佳极限分割法是比原来那个图像分割法好。
但是混凝土裂缝比真实的裂缝稍微细一点，它是由过度分割系数造成的。
通过大量的测试，过度分割现象显示了统计学现象，并且可以由测量系统的校核来弥补。
B、过度分割补偿的校核实验为了理解过度分割现象的统计学特征，我们做了一些校核实验。
在实验中，使用了6个标准宽度的被测物体，它们在表3中列了出来。
图3. 被测物体不同标准宽的数码图像CCD摄像头捕获了每个被测物体的10贞数码图片。
使用改进的最佳极限分割法，原始图片被转换成仅包含被测物体的二进制图像，它的过度分割系数也等于0.75。
通过计数像素的数量，被测物体的像素宽度就可以获得。
通过被测物体的像素宽度和实际宽度，可以得到过度分割值。
像素宽度和实际宽度的关系可以由公式y=Mx+s （11）来描述。
这里y是被测物体的实际宽度，x是像素宽度，M测量系统的电子放大率，s是静态干扰补偿。
S主要来自于量化误差和过度分割。
因此，得到s值意味着过度分割得到了补偿。
表I显示了过度分割补偿实验数据。
像素宽度的平均值也包含在表中。
通过最小平方拟合法得到M和s的值。
M的估计值在0.0233，s的估计值在0.0034.表I . 过度分割补偿的校准数据C .被测物体的测量实验为了更多地了解改进最佳极限分割法的性能和小尺寸测量系统的精确性，另外两个标准的被测物体被用来做测量实验。
每一个标准的被测物体被测10次，它们的平均值和标准偏差的数据都在表II.中显示。
表 . 标准被测物体的测量数据正如表II.中显示的，平均值很接近实际宽度，并且标准偏差是很小的。
根据在错误理论中3σ原则，小尺寸测量系统有着高测量精度，它的测量不确定度少于0.036mm。
IV. 结论通过仿真和实验分析，改进的最佳极限分割法比其他途径拥有以下的特点：●在很强的背景噪点情况下引进过度分割系数避免了局部最佳极限，它提高了小尺寸精密部件的测量精度。
●改进的最佳极限分割方法继承了原先好的性能，没有增加计算的复杂而且它保持了一个好的实时特点。
●轻微的过度分割可以被测量系统的计算来补偿，并且不会太大影响测量结果。
知识这篇文章得到了地壳动力学, 中国地震局(No.ZDJ2009-09), 和中国国家自然科学基金(No. 50908215).授予的研究的支持。
参考[1]R. Rodríguez, T. E. Alarcón, O. Pacheco, “新战略获得强劲的标记血管使用分水岭分割方法” Computers in Biology and Medicine, 卷35, 665-686页, 2005.10.[2]Z. Zhou, S., Zhao, Z. An, “研究自动涡轮叶片实时射线照相检验技术,”.中国机械工程日报, 卷41, 180-184页, 2005. 4.(中国)[3]J. Nuneza, J. Llacer, “天文图像分割的自组织神经网络和波,” 神经网络,卷16,411-417页, 2003.3.[4]S. K. Sinha, P. W. Fieguth, “神经模糊网络的分类埋管的缺陷” 建造自动化,卷15, 73-83页,2006.1.[5]X. Luo, “一种改进的区域生长算法对CT图像分割的植物根,”中国农业机械汇报, 卷37, 122-125页, 2006. 12.(in Chinese)[6]N. Sang, H. Li, W. Peng, et al, “以知识为基础的数字减影血管造影图像的自适应阈值分割,”图像与视觉计算，卷25, 1263-1270页，2007.8[7]K. C. Lin, “在计算速度的提高首先进行图像阈值,” 电子成像杂志, 卷14, 023011-1-6页, 2005.5[8]B. He, T. Ma, Y. Wang, et al, Visual C++数字图像处理, 人民邮电出版社. 北京, 2001. (中国) 702。