On the Features, Functions and Translation Methods of Public Signs
语言学重要概念梳理(中英文对照版)完整版

第一节语言的本质一、语言的普遍特征(Design Features)1.任意性 Arbitratriness:shu 和Tree都能表示“树”这一概念;同样的声音,各国不同的表达方式2.双层结构Duality:语言由声音结构和意义结构组成(the structure ofsounds and meaning)3.多产性productive: 语言可以理解并创造无限数量的新句子,是由双层结构造成的结果(Understand and create unlimited number withsentences)4.移位性 Displacemennt:可以表达许多不在场的东西,如过去的经历、将来可能发生的事情,或者表达根本不存在的东西等5.文化传播性 Cultural Transmission:语言需要后天在特定文化环境中掌握二、语言的功能(Functions of Language)1.传达信息功能 Informative:最主要功能The main function2.人际功能 Interpersonal:人类在社会中建立并维持各自地位的功能establish and maintain their identity3.行事功能 performative:现实应用——判刑、咒语、为船命名等Judge,naming,and curses4.表情功能 Emotive Function:表达强烈情感的语言,如感叹词/句exclamatory expressions5.寒暄功能 Phatic Communion:应酬话phatic language,比如“吃了没?”“天儿真好啊!”等等6.元语言功能 Metalingual Function:用语言来谈论、改变语言本身,如book可以指现实中的书也可以用“book这个词来表达作为语言单位的“书”三、语言学的分支1. 核心语言学 Core linguistic1)语音学 Phonetics:关注语音的产生、传播和接受过程,着重考察人类语言中的单音。
功能学派翻译理论

Holz-Mäntarri (行为理论)
• Holz-Mäntarri (1984) • Translational action(行为理论) A communicative process involving: • The initiator(翻译的发起者) • The commissioner(委托人) • The ST producer(原文生产者) • The TT producer(译文生产者) • The TT user(译文使用者) • The TT receiver(译文接受者)
Functional school
___by team three
General description about functionalist school
The school holds that translation is a particular variety of translatonal action based on source text. Any translational action, including translation itself, is an action. Any action has an aim or purpose. The Skopostheorie brought forth by Vermeer is the most important functionalist translation theory.
莱斯根据布勒的的语言功能,把文本类 型分为三种:信息型、表情型、操作型
Text type
Language function
Language dimension
informative
Representing objects and facts logical Contentfocused Transmit referential content Plain prose, explicitation as required
电子围栏说明书英文版

Directory preface (2)safe note (3)一.function and feature (4)1.1 function (4)1.2 feature (4)二、instruction (5)2.1 design basis (5)2.2.characteristic (5)2.3 technical index (5)2.4 LEDindicated (5)2.5The front ,bottom of controller (6)2.6wire connection instruction (6)三、design requirement (6)3.1safety level (6)3.2 security (7)四、installation (7)4.1 install mode (7)4.2 install angle (9)4.3terminal rod/insulator install (9)4.4 middle rod/insulator install (10)4.5middle pull rod/insulator install (10)4.6 tensioner install (10)4.7lightning-arrester (10)4.8alloy wire connection (11)4.9grounding (11)4.10ground wirin g (11)五、controller install &connection (11)5.1controller install (11)5.2controller and fence connection (12)六、debug and alarm test (12)6.1 electrify check (12)6.2alarm test................................... .. (12)七、install rule & security note (15)8.1 install rule (15)8.2 safety note and some removing methods (16)Prefaceshanghai Yi Section Company is specialized in manufacturing electric security fence with its own R&D department factory located in Shenzhen.Yi Section company persistently follows the principle"do everything with whole hearted", through superior quality and zealous service, Yi Section tries every efforts to respond the users that gives uslong-term support and trust.Pulse electric fence energizer and perimeter accessory that Yi Section products and develops conforming to the requirements of GB/T7946-2008 《pulse electric fence installation and safe operation》and meet CEI-61011 international electric fence standard. The patent water proof casing design, professional output method, advanced difference voltage output technology, unique LCD display function, make the products as the leader in the industry. The products sell to Unite States, Europe, Australia, Africa, and Asian country , Yi Section will be the base of international production and service. The products are widely used in ware house, industry areas, power station, private house, villa, school, car parks, plant areas, airpot, prison, military base.etc.Security note:◆Unless high/low voltage output disarmed of electric fence, don’t touch the conductor part of the system.◆Don’t install electric fence during the thunder, don’t install controller in moist place.◆Before install and use the product, do the safe educational technology training work.◆During the installation, in order to use normally in future, follow the instructions to make sure the grounding of products. Refer to “4.9 item”◆When the product is electrified , the inside of controller is 220v 50Hz AC current, 5-10KV high voltage pulse on the output port and electric fence, don’t touch it to make sure security.◆The user must know the related electric device knowledge, and read the user instruction and operate it carefully before install it with correct use methods, otherwise , our side don’t bear any responsibility if any accident arise.◆Once abnormal case happen, should make sure the power off, then professional staffs check and analysis repair or contact our company, don’t repack it by your own.一.Function and featureThere are two kinds of existing security products. One kind is a fence, like traditional brick wall, cement wall, glass wall, iron fence and so on, has the obvious contours, encircles the defense area, has the impediment external invasion function, but does not have the alarm and the surveillance function. Another kind is the alarm system, like the infrared correlation alarm apparatus, divulge electric cable, CCTV and so on, they have the alarm or the surveillance function, but does not have the impediment function.The smart pulse electric fence system integrate the prevents and alarm functions together, and increase the high voltage deterrent effect, not only have visible fence, but also have smart multi-zone controller, the alarm system becomes brand new perimeter security system.1.1 system function1.1.1 have complete and explicit dividing line fence,and formidable impediment and deter function.1.1.2 have low rate of false alarm function,when the outer fence alarm,the display windows rapidly display which zone suffers break,short-circuit and tamper, reflect the current condition.1.1.3 The DC 12V alarm connection,dry contact signal output, can linkage with other security system, improve the system’s safe rank.1.1.4 distinction function of electric fence and can detect all kinds of invasion rank, can differentiate the accidental invade from forceful invade ability.The accidental intruder see the alarm or suffer from the electric shock, the leave, the alarm apparatus do not send out alarm. The intruder invade forcefully , so that the electric fence will be destroyed, in this case, the system will send out alarm. Only the intruder really invade and destroy the system, then will send alarm, will not alarm mistakenly and fails to alarm.1.2 system feature1.2.1Absolute safety and alarm perceiveThe traditional high voltage pulse electric net system doesn’t have alarm perceive function, only through high voltage, strong current to impediment intruder. It is very easy to get intruder disable, even dies and so serious results. The smart pulse electric fence system adopts low energy pulse high voltage((5~10KV),because the energy is extremely low,the response time is extremely short, will not cause the damage to the humanbody. Once touched, will just leave dut to electric shock.1.2.2The low rate of false alarm and strong AdaptabilityThe smart pulse electric fence system is not affected by environment (for example. trees, animals, shake.etc ) and climate ( wind, snow, rain and fog,etc) and not limited by terrain height and winding degree. The rate of false alarm is very low.1.2.3 impediment and alarm dual functionThe new concept of smart pulse electric fence system is to prevent the intruder out of from the fence area,without the crime is the goal, give the intruder one kind of deterrent feeling and impediment function, dare not act rashly, achieve the security, reduces the crime.1.2.4 continuous work,set armed/disarmed .1.2.5 use the 12V/4Ah battery as backup power supply,when the power supply is off,can work continually.1.2.6 According to user’s request and site geographical environment, carry out the design and installation, and workwith multiple modern security products, such as CCTV monitored system, television supervisory system andGSM system to improve the system security ranks.1.2.7 Absolute safety, conforms to the international CE certification and is also in conformity with IEC 60335-2-76 andthe GB/T7946-2008 request, and get approved from public security form test.二、product instructions2.1 product design basis:GB/T7946-2008 "Pulse electric fence installation And Safe operation"2.2 product characteristic:● difference voltage output technology: every wire has voltage, have the differential voltage between the two adjacent wires.●LCD display the monitor state indicating the voltage on every wire.● high/low voltage switch manually, long-distance equipment auto switch function.●short circuit, break and tamper alarm, equipment fault auto-detect..●waterproof casing and downward appearance design guiding the industry development.●RS 485 main line control, keypad, computer, network and multiple long distance management solution.●DC12V, open/close dry contact alarm output, can work with other modern security products..2.3 product technical index2.3.1 power supply: AC180V-240v 50Hz2.3.2 use environment2.3.2.1 Temperature :-40-+50℃2.3.2.2 moisture: :≤95%2.3.3 output index:Output peak voltage:5KV~10KVOutput low voltage: 700~1000VPeak current of pulse: <10APulse duration: ≤0.1sPulse interval time: 1sMax quantity of electricity of pulse: 2.5mCMax energy of pulse: ≤5.0JSystem power consumption: 10w2.4 LED indicated lamp condition instructions:2.4.1 AC 220V or DC12V,when power supply input ,the green indicated lamp is bright.2.4.2 when armed,the yellow indicated lamp is bright. When disarmed,it shows off.2.4.3 when alarm break,the red indicated lamp is bright.2.4.4 when alarm short circuit,the red indicated lamp is bright2.4.5 when open cover, the red indicated lamp of tamper is bright.2.5 the front and bottom of controller instructions.2.6 the wire connection of controller instruction( above drawing)2.6.1 the controller comprise two parts: input and output,the positive pulse from output part”1st line high voltage output port”transfer to fence wire, then return back to the receiver part “ 1st line high voltage input port”, negative pulse from output”2nd line high voltage output port” transfer to fence wire, then return back to the receiver part” 2nd line high voltage input port” ,so form positive and negative pulse closed circuit on the front fence.2.6.2 high voltage earth output: can share the same earth output with lightning arrester, strong/weak earth separate.2.6.3 voltage switch: use for switching high/low voltage mode or automatic working mode.2.6.4 keypad output: can connect keypad directly or RS485 transfer RS 232,then connect to computer by long distance control, The RS 485 connection of each controller needs to go hands by hands connection, multiple connection is not allowed.2.6.5 switch output: output open/close dry contact signal, needs to connect with other linkage device.2.6.6 alarm output: when alarm, output DC 12V voltage, connect with alarm device DC 12v below 10W.2.6.7 power switch: control the energizer on/off.2.6.8 power supply input: connect AC 220V power source.三、design request of electric fence3.1 according to different security rank, configurate suitable electric fence,usually divide into Ι、Ⅱ、III th e three rank.Ι—general security rank, uses 4 lines system, the monitor zone partition does not surpass 500 meters.Ⅱ—medium security rank, uses 8 lines system, the monitor zone partition does not surpass 250 meters.III—higher security rank, uses 12 lines system, the monitor zone partition does not surpass 100 meters.Every zone must configurate independent controller, and have different triggered alarm apparatus, can indicate the alarm zone number. The alarm output usually linkage with CCTV, red-infrared correlation, spotlight, alarm apparatus etc. When applied, the length of monitor zone should set according to the perimeter total length, the terrain and the objective site environment.3.2 Security3.2.1 don’t allow get the AC power supply on electric fence. when the product gets faulty, please make sure the electricfence without any AC power supply.3.2.2 product used the rectification voltage dropping DC 12V, then voltage-rise to condenser charge, finally electriccapacity pulse discharge boosting transformer, the output energy is limited by rectification, primary voltage-rise, electric capacitance discharge, so many links, without harm to human body and absolute safety.3.2.3 association factor and the influence of electric shock:The electric pulse will not injure the intruder, to avoid the associated responsibility, firstly, electric fence warning signs should hang up at every other 10 meters striking position noted” electric fence, forbid climbing”, secondly, the height of electric fence should be above 1.8meters, if the height of electric fence is not high, should install isolated netting or wall to avoid accident.3.2.4 electric fence should not be with other power circuit and telecommunication wires on the same pole.3.2.5 electric fence should keep a certain safe distance from the power circuit. The minimum distance as shown in thetable 3-13.2.6 horizontal distance between electric fence and public rod border should be above 5m( except the electric fence on top of the wall)3.2.7 electric fence has pulse high voltage, when bad connection or close to charged conductor, will have fire spark.Thus, electric fence should install without any flammable gas, and without any flammable fluid place. Accordingto the related national standard, keep enough safe distance, or take the safe measures to isolate.3.2.8 When inspect the place to install electric fence, pls make sure the electric fence without any conflict with theelectric wire, pipeline from the ground. And without any sundry goods at nearby place, whether the electric fence is influenced strongly by transmitting station and other high frequency device. If have, must mark the signal wire with STP shielded twisted pair at the construction drawing.四、Electric fence install4.1 install mode:4.1.1 install on top of the wall.Install the electric fence on top of wall, the height of fence wall should be above 1.8meters, the wire pole/rod may have welding ,clip or buried the three modes, should choose suitable mode according to the fence structure condition. For example, install at the steel fence, can take the welding, install at the concrete wall, can take buried mode to make sure stable, artistic, may use other modes also.4.1.2 The attached installThe installation attaches on the fence or top or inside, the fence wall directly withstanding the pressure and wire tension, before installation, pls make sure the structural strength of the wall, if unreliable, should reinforce in advance. The metallic conductor wire from the front end of electric fence spacing to the wall should be smaller than 700mm.4.1.3 stand alone electric fenceInstall the electric fence around the building, because the height is above 1800mm, the wire number reaches 8-20, the strength of the wire to the terminal rod/pull rod tensity must be big, hence, terminal rod/ pull rod must have enough strength, and the embedment must be stable. If the soil is solid, may directly bury the wire of terminal rod and pull rod ends to underground 60cm to do fixed. If the rigidity of terminal wiring rod is insufficient, should increase the support rod. Although the middle pull rod does not withstand the tensity of wire, but must support the pressure of multiple wires. Therefore, also needs to install it stably.May take buried mode installation. Follow the chart 12 wires stand alone installation.4.2 electric fence installation angle: ( the angle with wall) see the right drawing.4.2.1 according to the site condition and the first party request((0°、22.5°、45°、67.5°、90°、112.5°、135°、157.5°、180°)and the inclined direction( inclination angle, camber angle, vertical or horizontal ) installation.4.2.2according to the perimeter condition:if resident area, school nearby, suggested byinclination angle or vertical installation,spacious area suggested to install bycamber angle, if the height of wall is above2.5meters, take horizontal installation.4.2.3 According to the protection object: if prevent from outside invasion, suggest camber angle installation, ifprevent from interior climbing, suggest inclination angle installation.4.3 terminal rod/terminal rod insulator installationFix the terminal rod insulator on the terminal rod throughterminal insulator clip, considering the distance is longer, the pullingstrength is bigger, every 100 meters or big corner and divided zonemust be set a terminal rod, see the right drawing:4.4 middle pull rod/middle pull rod insulator installation.The middle pull insulator assembles on the rod through screws or use pinlock to fix it. Should note thedirection of installing, usually every 20 meters to set a middle pull rod. See the rightdrawing:4.5 tensioner installationThe middle tensioner usually hang on the front of electric fence,when install alloy wire, penetrates the wire from the tensioner side’s round hole, aim and pass the middle hole, finally gets out from another side round hole. See the chart. Note, don’t make the tight line excessive, heat expanded and cold contracted, it maybe break in winter weather.4.6 lightning arrester installationFirstly fix the lightning arrester on the top of terminal rod,then through the own nuts fixing on the frame, the lightning arrester usually installs on top of pulse energizer, every zone must install two lightning arresters.4.7 a lloy wire connectionConnect it by wire linker, need connect the lead end to penetrate thewire linker, then use nut to contract wire in the wire linker. See theright drawing:4.8 warning signs installationAccording to the condition, every 10 meters set as one warning signs.4.9Grounding/earthgrounding request: according to “GB/T 7946-2008”, earth principle and weak electricity separated. The electric fence system includes: controller high voltage output earth, lighting arrester earth, protective earth output, the three kinds of earth. High voltage earth and lightning arrester earth can share the same earth, but must separate from protective earth, and the earth pole distance is above 10meters, the earth pole depth of burying is above 1.5m, the resistance value of weak electricity should be below 4Ω, the resistance of high voltage earth should be below 10Ω,the lightning arrester earth connects reliably by square millimeter copper wires.。
1 THE GEOMETRY OF MARKOV DIFFUSION GENERATORS

3
ห้องสมุดไป่ตู้
TABLE OF CONTENTS
INTRODUCTION 1. GEOMETRIC ASPECTS OF DIFFUSION GENERATORS 1.1 Semigroups and generators 1.2 Curvature and dimension 1.3 Functional inequalities 2. INFINITE DIMENSIONAL GENERATORS 2.1 Logarithmic Sobolev inequalities 2.2 L´ evy-Gromov isoperimetric inequality
1
THE GEOMETRY OF MARKOV DIFFUSION GENERATORS
Z¨ urich, November 1998
MICHEL LEDOUX Institut de Math´ ematiques Universit´ e Paul-Sabatier 31062, Toulouse (France) ledoux@math.ups-tlse.fr
´ R´ esum´ e. Ces notes sont un r´ esum´ e d’un mini-cours pr´ esent´ e` a l’Ecole Polytechnique F´ ed´ erale de Z¨ urich en novembre 1998. Elles ont pour but d’exposer quelqu’unes des id´ ees g´ eom´ etriques de l’´ etude des g´ en´ erateurs de diffusion d´ evelopp´ ees par D. Bakry et ses collaborateurs au cours des derni` eres ann´ ees. Les notions abstraites de courbure et dimension, qui ´ etendent les d´ efinitions g´ eom´ etriques, sont mises ` a profit dans des d´ emonstrations fonctionnelles de th´ eor` emes de comparaison riemanniens. Constantes optimales et mod` eles de r´ ef´ erence forment un aspect essentiel de l’analyse. Ces notes, qui pour l’essentiel ne comportent pas de d´ emonstrations, se proposent de d´ ecrire les grandes lignes et les principes g´ en´ eraux de cette ´ etude.
哈尔滨工程大学学报_总第207_218期_2014年总目次

哈尔滨工程大学学报第35 卷( 总第207 ~ 218 期) 2014 年2014 年总目次浮筏隔振及声辐射同步定相控制研究………………周刘彬,杨铁军,M ichael J.BRENNAN,刘志刚( 1,1)舰载机甲板调运过程避碰路径规划研究………………………………张智,林圣琳,夏桂华,朱齐丹( 1,9)基于倒谱和偏相干分析的噪声源分离方法………………………杨德森,韩闯,时胜国,于树华,时洁(1,16)平板喷气流场特性影响因素试验…………………………………………………叶青,董文才,欧勇鹏( 1,25)载人潜器载人舱布局优化………………………………………………………刘峰,韩端锋,韩海辉( 1,30)水下爆炸冲击波作用下船体刚体动态响应特性………………………………王军,郭君,孙丰,杨棣(1,38)高速滑行艇在规则波中的纵向运动数值研究…………………………王硕,苏玉民,庞永杰,刘焕兴( 1,45)磁性目标干扰下的舰船磁场建模……………………………………姚振宁,刘大明,周国华,喻洲( 1,53)新型船用螺旋桨加工装置刀具位姿控制算法……………………………………………………王瑞( 1,58)超音速横向流作用下射流的二次破碎模型研究…………杨东超,朱卫兵,陈宏,郭金鑫,刘建文( 1,62)船舶燃气轮机发电机组实时仿真新方案……………………………………李淑英,李铁磊,王志涛( 1,69)幅度与频率联合测量的LOFAR浮标定位算法………………………………陶林伟,王英民,苟艳妮( 1,74)小Kasami 序列的正交码元移位键控扩频水声通信…………………………………于洋,周锋,乔钢( 1,81)基于二阶差分滤波器的水下目标纯方位角跟踪适用于宽波束的多接收阵SAS 波数域成像算法…………………………………………王宏健,徐金龙,么洪飞,张爱华( 1,87)张学波,唐劲松,钟何平,张森( 1,93)FrFT 解线调浅剖信号增强算法及FPGA 实现………………………………朱建军,李海森,杜伟东( 1,102)利用相对时钟实现水声通信网络时分复用………………………………………………章佳荣,乔钢( 1,109)混流式喷水推进泵台架试验……………………………………………………………靳栓宝,王永生( 1,115) M 元仿海豚叫声隐蔽水声通信…………………………………………刘凇佐,刘冰洁,尹艳玲,乔钢( 1,119)非静压方程与波作用谱模型的波浪传播嵌套模拟……………………………………邹国良,张庆河( 1,126)渤海湾风浪预报数值模式的研究………………………………………李大鸣,潘番,罗浩,解以扬( 1,132)前视声呐多特征自适应融合跟踪方法…………………………………马珊,庞永杰,张铁栋,张英浩( 2,141)基于小波矩的自主式水下机器人目标识别………………………………万磊,黄蜀玲,张铁栋,王博( 2,148)钢悬链线立管与海底相互作用和疲劳分析………………REZAZADEH Kosar,陆钰天,白勇,唐继蔚( 2,155)水平管道内冰浆流动阻力特性实验研究…………………………王继红,王树刚,张腾飞,贾廷贵( 2,161)沿海进港航道通航水位仿真优化…………唐国磊,王文渊,郭子坚,于旭会,宋向群,王秉昌,张冉( 2,166) SH 波入射覆盖层半空间圆孔及圆夹杂的动力分析……………………………………陈冬妮,齐辉( 2,171)基于特征参数准则的舰船燃机燃烧室贫熄预测……………………………张智博,郑洪涛,李雅军( 2,177)一种水平蒸汽发射系统的基本性能优化设计…………………………………严志腾,金家善,朱泳( 2,184)台架试验的生物柴油发动机使用特性……………………刘玉梅,张敬师,崔丽,熊明烨,宋捷,李梦( 2,189)复合控制导弹的模糊控制算法及实现…………………………………史震,马文桥,张玉芳,林强( 2,195)双均流母线的均流电路及其控制方法……………………………………张强,张敬南,姚绪梁,霍虹( 2,202)浅海中单矢量水听器高分辨方位估计方法…………………………双卷流燃烧室与油束夹角匹配对柴油机排放的影响李楠松,朴胜春宋海岩,张海刚( 2,208)………………………………………………………高浩卜,李向荣,耿文耀,赵陆明,刘文鹏,刘福水( 2,216)基于产形线切齿法的弧齿锥齿轮精确建模……………………………豆书强,呼咏,荣国灿,杨兆军( 2,221)高速切削锯齿形切屑形成过程的有限元模拟………………………………段春争,王肇喜,李红华( 2,226)辅助起立机器人人体运动学建模与试验研究……………………王志强,姜洪源,KAMNIK Roman ( 2,233)一种用于高光谱图像特征提取的子空间核方法………………………………刘振林,谷延锋,张晔( 2,238)基于P SO-RV M 算法的发动机故障诊断…………………………………………毕晓君,柳长源,卢迪( 2,245) TinySBSec—新型轻量级WSN 链路层加密算法………………………………………白恩健,朱俊杰( 2,250)隐性知识对企业竞争优势作用机理的实证研究…………………………………………肖振红,李妍( 2,256)基于参数化小波的主冷却剂泵故障特征识别………………………………………………李彬,夏虹( 2,261)基于预测模型的模糊参数自寻优S 面控制器…………………………何斌,万磊,姜大鹏,张国成( 3,267)正交各向异性功能梯度夹层板的自由振动分析……………………李华东,朱锡,梅志远,张颖军( 3,274)泥浆快速化学脱稳与固液分离试验研究…………………………………冷凡,庄迎春,刘世明,陈兵( 3,280) MEMS 型水听器的自噪声分析………………………………………………方尔正,洪连进,杨德森( 3,285)单矢量水听器频域极化加权MUSIC 算法……………………………………刘伟,朴胜春,祝捍皓( 3,289)潜礁上孤立波传播的数值模拟……………………………房克照,刘忠波,唐军,邹志利,尹继伟( 3,295)台风浪对掠海飞行安全性的影响………………………………………无人水下航行器近海底空间路径规划方法……………………………郑崇伟,邵龙潭,林刚,潘静( 3,301)严浙平,邓超,赵玉飞,李本银( 3,307)大件运输中简支梁桥的动力响应分析及监测……………………刘波,王有志,王涛,丛侃,徐永芝(3,313)单侧有限宽度开缝竖通道内的旋转热流体特性…………………………霍岩,邹高万,李树声,郜冶( 3,320)面向起飞质量的小型飞航导弹参数优化方法………………………………陈阳阳,陈卫东,吴限德( 3,325)对冲式止回阀低流量关闭过程动态特性分析………………杨志达,陈炳德,黄伟,韩伟实,张志明( 3,331)蒸汽发生器多目标优化设计……………………………………………………陈磊,阎昌琪,王建军( 3,336)单井循环系统在不同初始地温下的特性……………………………………………宋伟,倪龙,姚杨( 3,342)泵用不可压缩流体密封刚度系数分析…………………………………张盟,王晓放,徐胜利,万学丽( 3,347)大数据分区管理模型及其应用研究…………………………………………张文燚,项连志,王小芳( 3,353)旋成体环形射流表面优化设计与减阻机理分析…………………………………赵刚,李芳,臧东阳( 3,361)快速高效无损图像压缩系统的低功耗硬件实现………………………………薛金勇,黑勇,陈黎明( 3,368)模拟退火遗传禁忌搜索的多用户检测算法…………………………………………………刁鸣,邹丽( 3,373)冻融循环对沥青混合料力学性能的影响……………………………李兆生,谭忆秋,吴思刚,杨福祺( 3,378)有无稀土CuCrSnZn 合金形变时效组织和性能……………苏娟华,杨哲,任凤章,魏世忠,陈志强( 3,383)水平轴潮流能水轮机设计与数值模拟…………………………………………盛其虎,赵东亚,张亮( 4,389)有限元/间接边界元法求解浸水板振动特性……………………………………刘城,洪明,刘晓冰( 4,395)浅海温跃层对水声传播损失场的影响………………潘长明,高飞,孙磊,王璐华,王本洪,李璨( 4,401)基于基面力的有限变形广义拟Hamilton 原理…………………………………刘宗民,梁立孚,樊涛( 4,408)简谐成分的盲源分离适用性研究………………………………………董建超,杨铁军,李新辉,代路( 4,413)不同边界条件正交各向异性圆柱壳结构固有振动分析…… 张爱国,李文达,杜敬涛,代路,杨铁军( 4,420)长输油气管道裂纹初始扩展特性………………………………………………………马有理,赵丽丽( 4,426)混合距离算法在蒸汽发生器优化设计中的应用………………………………阎昌琪,陈磊,王建军( 4,432)压力容器外部冷却两相自然循环特性理论分析……………………………赵国志,曹欣荣,石兴伟( 4,437)舰载机着舰的动力学建模…………………………………………………高旋弹丸背景涡流磁场建模与补偿………………………………………车用柴油机瞬态工况试验及性能评价方法………………………………夏桂华,董然,孟雪,朱齐丹( 4,445) 向超,卜雄洙,祁克玉,于靖( 4,457) 张龙平,刘忠长,田径,许允( 4,463)发动机多工况有机朗肯循环中混合工质研究……………尉庆国,王震,杨凯,杨富斌,张健,张红光( 4,469) Mindlin 矩形板在任意弹性边界条件下的振动特性分析………薛开,王久法,李秋红,王威远,王平(4,477)消除冗余循环前缀的水声信道OFDM 频域均衡算法………………………冯成旭,许江湖,罗亚松( 4,482)提高系统稳定性的可控串联补偿控制策略……………………………罗远翔,杨仁刚,蔡国伟,刘铖( 4,488)钢结构建筑物磁场特征定量评估………………………………………………………郭成豹,刘大明( 4,493)采用椭球波脉冲的多波段超宽带性能研究………………………………………陈丽丽,于欣,窦峥( 4,499)基于平均场均衡的Ad hoc 网络路由协议…………………………………张旭,钱志鸿,刘影,王雪( 4,504)铝合金微弧氧化放电火花对膜层粗糙度的影响……………………刘兴华,朱立群,刘慧丛,李卫平( 4,510) MWCNT 纳米纸/形状记忆聚合物复合材料导电性能研究……………………………张阿樱,吕海宝( 4,516)多桨型船舶轴频电磁场在浅海中传播特性分析………………………刘德红,王向军,朱武兵,嵇斗( 5,521)声矢量阵自适应抵消稳健波束形成算法…………………………………………梁国龙,陶凯,范展( 5,527)实船阻力及流场数值预报方法…………………………………………易文彬,王永生,杨琼方,李剑( 5,532)高速电磁阀动态响应特性响应面预测模型的研究………刘鹏,范立云,马修真,王昊,白云,宋恩哲( 5,537)全息数据外推与插值技术的极限学习机方法………………………弹性地基复合曲梁抗弯力学性能研究………………………………孙超,何元安,商德江,刘月婵( 5,544)郝扣安,王振清,周利民,王欣( 5,552)近断层地震激励下的隔震桥梁参数优化………………………………………………马涌泉,邱洪兴( 5,558)微小型自主式水下机器人系统设计及试验…………魏延辉,田海宝,杜振振,刘鑫,郭志军,赵大威( 5,566)舰载机阻拦过程动力学仿真…………………………………张智,闻子侠,朱齐丹,张雯,喻勇涛( 5,571)封锁雷智能防排系统………………………………………………孙宇嘉,王晓鸣,贾方秀,于纪言( 5,580)具有输入与状态时变时滞线性系统的鲁棒H∞控制……………………………侯晓丽,邵诚,李永凤( 5,585) LTS 仿真模型组合验证方法……………………………………………………冯晓宁,王卓,王金娜( 5,589)支持高效查询检索的大数据资源描述模型…………………………………张文燚,项连志,王小芳( 5,594)类人型五指手构型的优化设计…………………………樊绍巍,陈川,姜力,曾博,刘宏,邱景辉( 5,602)齿轮激光跟踪在位测量的姿态调整模型………………………………………张白,石照耀,林家春( 5,607)圆筒型永磁直线电机齿槽力解析………………………………………………黄克峰,李槐树,周羽( 5,613)微型二维光流传感器设计……………………………………………张洪涛,张广玉,李隆球,王林( 5,619)头型对回转体非定常空化流动特性影响的实验研究…………………………………胡常莉,王国玉( 5,624)基于多因素模糊判决的多模型机动目标跟踪……………………………………刘扬,闫新庆,国强( 5,630)利用独立性约束非负矩阵分解的高光谱解混算法……………………………………杨秀坤,王东辉( 5,637)遮光剂对SiO2气凝胶热辐射特性影响的理论研究…………………………苏高辉,杨自春,孙丰瑞( 5,642)造船供应链利益分配问题的Shapley 值法分析…………………………………………范德成,胡钰( 5,649)企业产学研合作原始创新系统的演化机制……………………………………李柏洲,罗小芳,苏屹( 5,654) GPU 在SPH 方法模拟溃坝问题的应用研究……………………………杨志国,黄兴,郑兴,段文洋( 6,661)水面舰船风浪环境适应性模糊综合评价方法………………………………焦甲龙,孙树政,任慧龙( 6,667)基于总体坐标法的大变形锚泊线的静力分析…………………………………………… 桅杆RCS 可视化计算方法改进……………………………………………………………马刚,孙丽萍( 6,674)崔俊伟,杨飏( 6,679)深海中潜艇腐蚀相关磁场全空间分布特征分析………………………陈聪,李定国,蒋治国,龚沈光( 6,684)基于弹性基础梁理论的球壳承压能力分析……………………………………………熊志鑫,罗培林( 6,690) EMD 与样本熵在往复压缩机气阀故障诊断中的应用………………张思阳,徐敏强,王日新,高晶波( 6,696)基于模型预测控制的动力定位过驱动控制设计……………………………梁海志,李芦钰,欧进萍( 6,701)不确定条件下海上危化品运输安全演变机理…………………………………李建民,齐迹,郑中义( 6,707)功能梯度热障涂层热应力的有限体积法研究……………龚京风,张文平,宣领宽,明平剑,韩春旭( 6,713)点源模型在旋转声源声场计算中的推广应用……………………………………………付建,王永生( 6,719)新型轮机仿真平台实操考试自动评估算法…………………………张巧芬,孙建波,史成军,孙才勤( 6,725)基于椭球约束的载体三维磁场补偿方法……………………………………于振涛,吕俊伟,稽绍康( 6,731)基于双主交替领航的多AUV 协同导航方法……………………………高伟,刘亚龙,徐博,唐李军( 6,735)迭代寻优的稳健波束形成…………………………………………………王燕,吴文峰,范展,梁国龙( 6,741)基于Schmidt 正交单位化的稀疏化定位算法……………………………赵春晖,许云龙,黄辉,崔冰( 6,747)联合多元信道编码调制-网络编码方案………………………………………佟宁宁,赵旦峰吴宇平( 6,753)高光谱图像的同步彩色动态显示………………………………………………刘丹凤,王立国,赵亮( 6,760) Deltamax 高压缓冲器的研制…………………………………………………………………谢飞,李格( 6,766)加速加载条件下沥青路面结构动力响应……………………………陈静云,刘佳音,刘云全,周长红( 6,771)中国航空航天制造业自主创新效率研究……………………………………………………陈伟,周文( 6,777)摇摆对单相强迫循环流量波动特性的影响机理……………………………幸奠川,阎昌琪,孙立成( 6,784)岸式振荡水柱波能转换装置的数值模拟………………………………基于水动力-结构模型的波浪载荷计算方法…………………………宁德志,石进,滕斌,赵海涛( 7,789)任慧龙,孙葳,李辉,童晓旺( 7,795)基于改进人工鱼群算法的无人艇控制参数优化…………………………廖煜雷,刘鹏,王建,张铭钧( 7,800)气体燃料船用主机工作过程三维数值模拟…………………………冯立岩,田江平,翟君,隆武强( 7,807)多孔介质内预混合燃烧的二维数值模拟……………………………………刘宏升,张金艳,解茂昭( 7,814)欠驱动水面船舶非线性信息融合航迹跟踪控制………………………………胡洲,王志胜,甄子洋( 7,820)柔性飞艇及其悬挂屏体系结构力学性能分析……………………陈宇峰,陈务军,何艳丽,张大旭( 7,826)渤海海域软表层土的动力特性…………………………………………………………荣棉水,李小军( 7,833)加筋土拉拔界面作用的离散元细观模拟………………………………王家全,周健,吴辉琴,徐华( 7,839)木结构组合板蒙皮效应研究…………………………王卫锋,崔楠楠,颜全胜,黄仕平,朱竞翔,夏珩( 7,846)地震总输入能量谱峰值估计……………………………………………………………李亚楠,王国新( 7,851)改进的分水岭图像分割算法…………………………………………………孙惠杰,邓廷权,李艳超( 7,857)多属性决策的IPv6 移动网络路由选择算法………………………………赵蕾,李小平,谭帅帅,石磊( 7,865)实现连续柔性成形的临界摩擦系数的研究……………………………王蜜,蔡中义,李琳琳,李明哲( 7,871)具有复杂缓冲的分段多流水车间调度问题……………………………………张志英,代乙君,隋毅( 7,875) TBM 刀盘支撑筋结构设计及静动态特性分析………………………………霍军周,杨静,孙伟,张旭( 7,883)柔性滤波驱动机构的自适应动态面模糊控制…………………………罗绍华,王家序,李俊阳,石珍( 7,889)基于ANSYS 二次开发的风力机叶片结构优化…………………陈进,马金成,汪泉,郭小锋,李松林(7,895)基于四叉树的海浪磁场快速仿真算法………………………………熊雄,杨日杰,韩建辉,郭新奇( 7,901)自适应匹配追踪的OFDM 系统窄带干扰检测……………………………………………毕晓君,邵然( 7,908)共形阵列天线超宽频带波达方向实时估计………………司伟建,万良田,刘鲁涛,田作喜,蓝晓宇( 7,913)制造业企业产品与工艺创新知识流耦合过程与控制……………………………………毕克新,黄平( 7,919)改进的单轴旋转SINS 初始对准方法……………………………………………………王东,李国林( 8,925)浅海中船舶尾流产生的感应电磁场…………………………………………张伽伟,姜润翔,龚沈光( 8,931)钢结构厚板及其对接焊缝Z 向性能试验……………………………………王元清,张元元,石永久( 8,936)基于逐步等效平面法的可靠性敏度分析方法………………………………王杰方,安伟光,宋向华( 8,942)柴油机缸内辐射换热三维数值模拟………………………………付丽荣,张文平,明平剑,罗跃生( 8,948)多谐次相位法在柴油机故障诊断上的应用……………………肖小勇,向阳,钱思冲,李瑞,周强( 8,954)基于Petri 网的情境感知服务建模及干扰发现…………………………………………胡志芳,卢涛( 8,961)动态部署与分块存储策略的数据恢复模型…………………………………黄春梅,姜春茂,曲明成( 8,968)工程陶瓷脆性域旋转超声磨削加工切削力研究…………………………魏士亮,赵鸿,薛开,荆君涛( 8,976)直立柔性窄结构上的海冰挤压同时破坏………………………………………………王译鹤,岳前进( 8,982)四缸双作用斯特林发动机平衡改进设计…………………………霍军周,李涛,张旭,杨静,路林( 8,987)弹性支撑阶梯柱侧向位移与稳定性的精确分析…………………………………都亮,陆念力,兰朋( 8,993)基于粗集理论桥式起重机能效模糊评价……………………………叶伟,童一飞,李东波,李向东(8,997)谐振式微型管道机器人设计与实验……………………………………刘磊,李娟,李伟达,秦佳( 8,1002)基于MUSIC 算法多脉冲采样的ARM 抗诱饵测向误差分析…………………司伟建,朱曈,张梦莹( 8,1008)微带天线对高功率电磁脉冲响应的时域分析……………………………………………李磊,张昕( 8,1015)分块GPCA 和多视点图像融合………………………………………………吴远昌,孙季丰,李万益( 8,1022)中国地方政府税收竞争行为特性与激励机制……………………………………………梁河,西宝( 8,1028)厚规格铝压型板辊弯成型工艺与裂纹缺陷分析……………………………陈泽军,李军超,黄光杰( 8,1035)周期力场下窄通道内汽泡滑移实验研究………………李少丹,谭思超,高璞珍,许超,胡健,郑强( 8,1040)液舱晃荡与船体非线性时域耦合运动计算………………黄硕,段文洋,游亚戈,姜金辉,王文胜( 9,1045)欠驱动船舶简捷鲁棒自适应路径跟踪控制…………………………………张国庆,张显库,关巍( 9,1053)不同头型弹体低速入水空泡试验研究……………………………杨衡,张阿漫,龚小超,姚熊亮( 9,1060)带支腿浮式结构水动力建模及波浪力分析…………………………………黄亚新,武海浪,陈徐均( 9,1067)船用电缆剩余寿命快速检测方法研究………………………………王鹤荀,纪玉龙,李根,孙玉清( 9,1076)舰船抗沉性的抗鱼雷攻击极限能力分析………………………………………郭君,孙丰,曹冬梅( 9,1082)云重心方法在舰炮维修性评价中的应用………………………………………刘勇,徐廷学,孙臣良( 9,1087) Myring 型回转体直航阻力计算及艇型优化…………………………庞永杰,王亚兴,杨卓懿,高婷( 9,1093)自主式水下机器人故障特征增强方法……………………………张铭钧,刘维新,殷宝吉,王玉甲( 9,1099)竖直窄矩形通道内弹状流特性的实验研究…………………闫超星,阎昌琪,孙立成,王洋,周艳民( 9,1106)多管火箭弹射击精度的复合形法优化……………王巍,陈卫东,张宝财,吴限德,路胜卓,于佳( 9,1112)提高轴系扭振信号时频转换精度的方法………………郭宜斌,李玩幽,蔡鹏飞,卢熙群,吕秉琳( 9,1117)双线圈点火系统特性仿真与试验研究……………………………马修真,靖海国,杨立平,宋恩哲( 9,1124)低快拍下MIMO 雷达收发角度联合估计方法………………………王咸鹏,王伟,马跃华,王君祥( 9,1129)电动汽车复合制动系统过渡工况协调控制策略………………………………朱智婷,余卓平,熊璐( 9,1135)高速公路汽车追尾碰撞预警关键参数估计………………………………………宋翔,李旭,张为公( 9,1142)面向聚类分析的邻域拓扑势熵数据扰动方法…………………………张冰,杨静,张健沛,谢静( 9,1149)基于梯度增强和逆透视验证的车道线检测…………………………王超,王欢,赵春霞,任明武( 9,1156)面向虚拟手术的碰撞检测优化算法…………………………于凌涛,王涛,宋华建,王正雨,张宝玉( 9,1164)基于可拓理论的低碳服务业集聚竞争力评价…………………………………………………贾立江( 9,1171)污泥改良盐碱土中污染物的淋滤行为……………………………………孟繁宇,姜珺秋,赵庆良,王琨,刘纯甫,马立,杨俊晨,郑振( 9,1176)复合材料天线罩螺栓连接结构损伤失效分析………………………杨娜娜,王伟,董一帆,姚熊亮( 10,1183)海上超大型浮体的水弹性响应预报…………………………………………万志男,翟钢军,程勇( 10,1189)舰载风速仪测量误差与安装位置的关系研究……………郜冶,谭大力,李海旭,王金玲,刘长猛( 10,1195)已服役20 年预应力结构现存预应力的试验研究……………………………………黄颖,房贞政( 10,1201)碳纳米管在水性体系中的分散性能及机理……………………………王宝民,韩瑜,葛树奎,张源( 10,1206)曲线箱梁桥的空间传递矩阵…………………………………………………………闫仙丽,李青宁( 10,1212)利用船舶运动数据估计海浪方向谱的研究……………赵大威,丁福光,谢业海,杭栋栋,边信黔( 10,1219)基于相位校正的零点约束直达波抑制方法………………………黄聪,张殿伦,孙大军,兰华林( 10,1224)抗扰动移动对等覆盖网的构建及性能评价…………………………………李军,张国印,王向辉( 10,1231)相似度质心多层过滤策略的动态文摘方法………………………于洋,范文义,刘美玲,王慧强( 10,1236)基于小波变换和陀螺的高旋弹角运动测量技术…………杨登红,李东光,申强,曾广裕,杨瑞伟( 10,1242)基于Bloom 滤波器的快速路由查找方法……………………………………于明,王振安,王东菊( 10,1247)基于干扰对齐的自适应频谱共享算法…………………………………………李记,赵楠,殷洪玺( 10,1253)宽零点约束的圆阵宽带波束形成研究…………………………………张曙,栾晓明,李亮,蒋毅( 10,1260)基于L 邻域分割的结构性纹理合成方法…………………………张雨禾,耿国华,刘伦椿,周子骏( 10,1265)战场宽带数据链分布式虚拟骨干网的构建…………………………陶凯,杨春兰,史海滨,范立耘( 10,1272)全驱动式自主水下航行器有限时间编队控制………………………………袁健,张文霞,周忠海( 10,1276)基于贝叶斯压缩感知多目标定位算法………………………………吴哲夫,许丽敏,陈滨,覃亚丽( 10,1282)基于贝叶斯网络的海洋工程装备故障诊断模型…………………………赵金楼,成俊会,岳晓东( 10,1288)介质阻挡放电辅助甲烷蒸汽重整的动力学分析…………………………刘倩,郑洪涛,杨仁,陈曦( 10,1294)矩形通道弹状流液膜特性…………………………………王洋,阎昌琪,孙立成,闫超星,刘国强( 10,1301)环流理论与泵理论相结合的导管桨设计优化………………………刘业宝,苏玉民,赵金鑫,张赫( 11,1307)海上风机半潜式基础概念设计与水动力性能分析……………………唐友刚,桂龙,曹菡,秦尧( 11,1314)饱和海床土渗流-应力耦合损伤及液化破坏规律( Ⅰ)…………刘红军,李洪江,王虎,吕小辉( 11,1320)水声阵列信号处理对角减载技术…………………………………赵安邦,周彬,宋雪晶,毕雪洁( 11,1327) Park-Ang 三维损伤模型的参数确定和性能等级划分…………………郭进王君杰韩鹏王文彪( 11,1332)公路减隔震桥梁的地震反应简化分析………………………………………李闯,叶爱君,余茂峰( 11,1339)改善型波形钢腹板PC 组合梁抗弯强度试验研究……………………………充量温度对某中速柴油机燃烧和NO x 排放的影响逯彦秋,安关锋,程进( 11,1345)……………………冷先银,魏胜利,田江平,何爽,隆武强,郭海娥,张旭东,钟兵,李焕英( 11,1351)散射体旋转角对二维声子晶体带隙结构影响分析………………………杜敬涛,贺彦博,冯浩成( 11,1358)遥现中基于显著特征的深度图像滤波算法……………………………………………冯策,戴树岭( 11,1364)带极点约束离散广义分段仿射系统的H!保性能控制……………………王茂,周振华,王学翰( 11,1369)多自由度微陀螺结构参数对其动态性能影响分析……………………………郝燕玲,刘博,胡钰( 11,1378)低重力模拟系统控制策略…………………………………………………朱齐丹,陈力恒,卢鸿谦( 11,1384)一种面向不确定图的SimRank 算法………………………………董宇欣,王莹洁,宁鹏飞,张耀元( 11,1390)基于雅可比旋量统计法的发动机三维公差分析……………陈华,唐广辉,陈志强,李志敏,金隼( 11,1397)不同模式下TBM 刀群三维回转切削仿真与优化设计………………霍军周,杨静,孙伟,李庆宇( 11,1403)考虑周期预防性维护的异速并行机集成调度研究………………………江才林,陆志强,崔维伟( 11,1409)机械弹性车轮结构参数对牵引性能的影响……………臧利国,赵又群,李波,陈月乔,李小龙( 11,1415)并联六自由度机构运动学与动力学标定对比………………………………皮阳军,王骥,胡玉梅( 11,1422)补0 的TD-AltBO C 多信号分量联合捕获方法………………………网络下遥感影像实时压缩及渐进传输系统设计……………………杨建雷,金天,黄智刚,秦红磊( 11,1427)石翠萍,张钧萍,曲海成,张晔( 11,1434)北方地区供暖情况下室内热环境数值分析………………………Mg /A l 液固双金属复合材料的界面及相组成王烨,王靖文,王良璧,张文霞( 11,1441)………………………………………………赵成志,李增贝,张贺新,杜德顺,符策鹄,余娇娇( 11,1446)水翼剖面多目标粒子群算法优化………………………………………………黄胜,任万龙,王超( 12,1451)阻抗梯度变化介质的声学特性………………………………杨德森,孙玉,胡博,韩闯,靳仕源( 12,1458)可刚性固定的同振圆柱型矢量水听器的设计…………………………刘爽,李琪,贾志富,刘勇( 12,1467) AUV 水声跳频通信调制解调器的设计与实现……………………范巍巍,张殿伦,董继刚,张友文( 12,1473)饱和海床土渗流-应力耦合损伤及液化破坏规律( Ⅱ)……………李洪江,刘红军,王虎,张冬冬( 12,1480)腐蚀预应力钢绞线的疲劳试验分析……………………………………余芳,贾金青,姚大立,吴锋( 12,1487)管间距对水平管降膜蒸发流动形态和传热的影响………沈胜强,陈学,牟兴森,王耀萱,高宏达( 12,1492) Lock-up 装置的作用机理与分析模型………………………………………夏修身,崔靓波,李建中( 12,1497)磁弹性理论中的守恒定律和路径无关积分………………………………刘宗民,周健生,宋海燕( 12,1503)壳装高能固体推进剂的殉爆实验与数值模拟…………………………………路胜卓,罗卫华,陈卫东,王巍,张丰超,于艳春,李广武( 12,1507)配筋活性粉末混凝土梁抗剪承载力………………………………………邓宗才,周冬至,程舒锴( 12,1512)基于全量补偿算法的结构损伤识别………………………王凤刚,凌贤长,徐训,张锋,赵莹莹( 12,1519)叶片积垢对压气机性能衰退的影响…………………………王松,王国辉,韩青,王忠义,任翱宇( 12,1524)复杂形状波力直线发电装置的优化……………………林礼群,吴必军,王幸,吴春旭,王文胜( 12,1529)光纤捷联惯导系统的双轴旋转调制方案………………………………………………于飞,阮双双( 12,1536)活塞航空发动机复合增压技术仿真分析…………………………………潘钟键,何清华,张祥剑( 12,1543)磁力弹簧式压电振动送料器的设计与试验…………………田晓超,杨志刚,刘勇,沈燕虎,吴越( 12,1548)船舶主尺度设计的高维多目标多方向进化算法………………………毕晓君,张永建,苍岩,肖婧( 12,1553)基于特征空间算法的非圆相干信源DOA 估计…………………刁鸣,丁兆明,高洪元,李晨琬( 12,1559) GNSS 信号捕获的包络损耗及其补偿方法………………………………………………李健,陈杰( 12,1564)核安全级仪控软件可靠性评估模型构建………………………………………( 第35 卷卷终)迟淼,史丽萍,刘盈( 12,1570)Journal of Harbin Engineering UniversityVol.35( Sun N o.207~218) 2014ContentsOn synchrophasing control of vibration and sound radiation for a floating raft vibration isolation system ……………………………………ZHO U Liubi n,YANG T i ejun,M i c hae l J.BRE NNAN,LIU Zhigang( 1,1) Collision avoidance path planning for an aircraft in scheduling process on deck…………………………………………………ZHA NG Zhi,LIN Shenglin,XIA Guihua,ZHU Qi dan( 1,9) Separation of noise sources based on the cepstrum and partial coherence theor y…………………………………YANG Desen,HAN Chuang,S HI Shengguo,YU Shuhua,SHI Jie( 1,16)An experimental study of the flow field around the flat plate with air injection………………………………………………………………Layout optimization of a human occupied vehicle manned cabin ………………………………………………………………YE Q i ng,DONG W encai,O U Y ongpeng( 1,25) LIU F eng,HA N Duanfeng,HA N Haihui( 1,30)Dynam ic response of the rigid ship's hull subjected t o under w ater shock w aves…………………………………………………………WA NG J un,GUO J un,S UN Feng,YA NG Di( 1,38) Numerical study on longitudinal m otions of a high-speed planing craft in regular waves……………………………………………WANG Shuo,SU Yumin,PANG Yongjie,LIU Huanxi ng( 1,45)Research on the m odeling method f or a ship's magnetic f ield w ith magnetic target's inter f erence ……………………………………………YA O Z henning,LIU D aming,ZHOU G uohua,Y U Z hou( 1,53) The cutter position control algorithms for a new marine propeller processing deviceModeling the secondary breakup of a liquid jet in supersonic cross flows…………WA NGRui( 1,58)…………………………YANG D ongchao,ZHU W eibing,CHE N H ong,GUO J i nxin,LIU J i anwen( 1,62)A new scheme for real-time simulation of a marine gas tur bine generator set……………………………………………………………………LI Shuying,LI Tielei,WANG Zhitao( 1,69) LOFAR sonobuoy localization algorithm based on the measurements of am plitude and frequency ……………………………………………………………TA O Linwei,W A NG Y i ngmin,G O U Y anni( 1,74) Orthogonal code shift keying spread spectrum underwater acoustic communications employing the small Kasami se-quence …………………………………………………………………Y U Yang,ZHOU Feng,QIAO Gang( 1,81) Research on second order divided difference filter algorithm for underwater target bearing-only tracking …………………………………………WANG Hongjian,XU Jinlong,YAO Hongfei,ZHANG Aihua( 1,87) Wavenumber-domain imaging algorithm for wide-beam multi-receiver synthetic aperture sonar …………………………………………ZHA NG Xuebo,TANG Jinsong,ZHONG Heping,ZHANG Sen( 1,93) FrFT dechirping sub-bottom profiling signal enhancement algorithm and FPGA implementation ………………………………………………………………ZHU Jianjun,LI Haisen,DU W eidong( 1,102)A relative clock based TDMA protocol f or underwater acoustic communication networks…………………………………………………………………………ZHA NG Jiarong,QIAO Gang( 1,109) Bench test of the mixed-flow waterjet pump…………………………JIN S huanbao,WA NG Y ongsheng( 1,115) M-ray covert underwater acoustic communication by mimicking dolphin sounds…………………………………………………LIU S ongzuo,LIU B ingjie,YIN Y anling,QIAO G ang( 1,119) Simulation of wave transformation by nesting the non-hydrostatic equation and wave action spectrum model ………………………………………………………………………A study of the numerical forecast model of wind waves in Bohai BayZOU Guoliang,ZHA NG Qi nghe( 1,126) ………………………………………………………LI Daming,PAN Fan,LUO Hao,XIE Yiy ang( 1,132) An adaptive fusion method used in forward looking sonar multi-feature tracking……………………………………MA Shan,PANG Yongjie,ZHANG Tiedong,ZHANG Yi nghao( 2,141) Object recognition system for an autonomous underwater vehicle based on the wavelet invariant moment …………………………………………WAN Lei,HUANG Shuling,ZHANG Tiedong,WANG Bo( 2,148) The interaction between the steel catenary riser and the seabed and the analysis of fatigue…………………………………………RE ZA ZA DE H K osar,LU Y utian,BAI Y ong,TA NG J i wei( 2,155)。
翻译理论

一外国翻译史和翻译名家西方翻译史可大致归纳为四种主要的翻译研究方法:①语义学翻译法(philological approach)②语言学翻译法(linguistic approach)③交际学翻译法(communication approach)④社会符号学翻译法(socio-semiotic approach)社会符号学翻译法:源于美国翻译理论家,《圣经》翻译家尤金.奈达(Eugene A.Nida)创导的社会符号学翻译法语言学翻译法:着眼的基本问题是原文的字面意义(the literal character of the source text)篇章的主题结构和风格(the thematic structur and style of the discouse).这种方法强调直译。
(公元前一世纪,古罗马翻译家兼演说家西塞罗:翻译不应拘泥与原文的词语而注重原文的思想,坚持不可逐字死译而要符合译文的语言规则与特性)。
18世纪中叶,爱丁堡大学的历史学教授泰特勒在《翻译的原则》一书中提出著名的三原则:①译文应完整的再现原文的思想内容——That the translation should give a complete transcript of the ideas ofthe original work。
②译文的风格,笔调应与原文的性质相同——That the style and manner of writing should be of the samecharacter with that of the original)③译文应像原文一样流畅自然——That the translation should have all the ease of the original composition.)语言学翻译法:现代语言文学发展的产物,它主张在对比语言学(contrastive linguistics)的基础上制定的一系列规则以实现等值(equivalence)这一学派的代表人物:英国语言学家卡特福特catford,法国的穆南和前苏联的巴尔胡达罗夫等人。
Bayesian Modeling of Dynamic Scenes

Bayesian Modeling of Dynamic Scenesfor Object DetectionYaser Sheikh,Student Member,IEEE,and Mubarak Shah,Fellow,IEEE Abstract—Accurate detection of moving objects is an important precursor to stable tracking or recognition.In this paper,we present an object detection scheme that has three innovations over existing approaches.First,the model of the intensities of image pixels as independent random variables is challenged and it is asserted that useful correlation exists in intensities of spatially proximal pixels.This correlation is exploited to sustain high levels of detection accuracy in the presence of dynamic backgrounds.By using a nonparametric density estimation method over a joint domain-range representation of image pixels,multimodal spatial uncertainties and complex dependencies between the domain(location)and range(color)are directly modeled.We propose a model of the background as a single probability density.Second,temporal persistence is proposed as a detection criterion.Unlike previous approaches to object detection which detect objects by building adaptive models of the background,the foreground is modeled to augment the detection of objects(without explicit tracking)since objects detected in the preceding frame contain substantial evidence for detection in the current frame.Finally,the background and foreground models are used competitively in a MAP-MRF decision framework,stressing spatial context as a condition of detecting interesting objects and the posterior function is maximized efficiently by finding the minimum cut of a capacitated graph.Experimental validation of the proposed method is performed and presented on a diverse set of dynamic scenes.Index Terms—Object detection,kernel density estimation,joint domain range,MAP-MRF estimation.æ1I NTRODUCTIONA UTOMATED surveillance systems typically use stationarysensors to monitor an environment of interest.The assumption that the sensor remains stationary between the incidence of each video frame allows the use of statistical background modeling techniques for the detection of moving objects such as[39],[33],and[7].Since“interesting”objects in a scene are usually defined to be moving ones, such object detection provides a reliable foundation for other surveillance tasks like tracking([14],[16],[5])and is often also an important prerequisite for action or object recognition.However,the assumption of a stationary sensor does not necessarily imply a stationary background.Exam-ples of“nonstationary”background motion abound in the real world,including periodic motions,such as a ceiling fans,pendulums,or escalators,and dynamic textures,such as fountains,swaying trees,or ocean ripples(shown in Fig.1).Furthermore,the assumption that the sensor remains stationary is often nominally violated by common phenomena such as wind or ground vibrations and to a larger degree by(stationary)hand-held cameras.If natural scenes are to be modeled,it is essential that object detection algorithms operate reliably in such circumstances.Back-ground modeling techniques have also been used for foreground detection in pan-tilt-zoom cameras[37].Since the focal point does not change when a camera pans or tilts, planar-projective motion compensation can be performed to create a background mosaic model.Often,however,due to independently moving objects motion compensation may not be exact,and background modeling approaches that do not take such nominal misalignment into account usually perform poorly.Thus,a principal proposition in this work is that modeling spatial uncertainties is important for real world deployment,and we provide an intuitive and novel representation of the scene background that consistently yields high detection accuracy.In addition,we propose a new constraint for object detection and demonstrate significant improvements in detection.The central criterion that is traditionally exploited for detecting moving objects is background difference,some examples being[17],[39],[26],and[33].When an object enters the field of view it partially occludes the background and can be detected through background differencing approaches if its appearance differs from the portion of the background it occludes.Sometimes,however,during the course of an object’s journey across the field of view, some colors may be similar to those of the background and, in such cases,detection using background differencing approaches fail.To address this limitation and to improve detection in general,a new criterion called temporal persistence is proposed here and exploited in conjunction with background difference for accurate detection.True foreground objects,as opposed to spurious noise,tend to maintain consistent colors and remain in the same spatial area(i.e.,frame to frame color transformation and motion are small).Thus,foreground information from the frame incident at time t contains substantial evidence for the.The authors are with the School of Computer Science,University of CentralFlorida,4000Central Florida Blvd.,Orlando,FL32816.E-mail:{yaser,shah}@.Manuscript received22July2004;revised30Nov.2004;accepted3Mar.2005;published online14Sept.2005.Recommended for acceptance by S.Baker.For information on obtaining reprints of this article,please send e-mail to:tpami@,and reference IEEECS Log Number TPAMI-0375-0704.0162-8828/05/$20.00ß2005IEEE Published by the IEEE Computer Societydetection of foreground objects at time tþ1.In this paper, this fact is exploited by maintaining both background and foreground models to be used competitively for object detection in stationary cameras,without explicit tracking.Finally,once pixel-wise probabilities are obtained for belonging to the background,decisions are usually made by direct thresholding.Instead,we assert that spatial context is an important constraint when making decisions about a pixel label,i.e.,a pixel’s label is not independent of the pixel’s neighborhood labels(this can be justified on Bayesian grounds using Markov Random Fields[11], [23]).We introduce a MAP-MRF framework,that competi-tively uses both the background and the foreground models to make decisions based on spatial context.We demonstrate that the maximum a posteriori solution can be efficiently computed by finding the minimum cut of a capacitated graph,to make an optimal inference based on neighbor-hood information at each pixel.The rest of the paper is organized as follows:Section1.1 reviews related work in the field and discusses the proposed approach in the context of previous work.A description of the proposed approach is presented in Section1.2.In Section2,a discussion on modeling spatial uncertainty(Section2.1)and on utilizing the foreground model for object detection(Section2.2)and a description of the overall MAP-MRF framework is included(Section2.3). In Section2.3,we provide an algorithmic description of the proposed approach as well.Qualitative and quantitative experimental results are shown in Section3,followed by conclusions in Section4.1.1Previous WorkSince the late1970s,differencing of adjacent frames in a video sequence has been used for object detection in stationary cameras,[17].However,it was realized that straightforward background subtraction was unsuited to surveillance of real-world situations and statistical techni-ques were introduced to model the uncertainties of back-ground pixel colors.In the context of this work,these background modeling methods can be classified into two categories:1)Methods that employ local(pixel-wise)models of intensity and2)Methods that have regional models of intensity.Most background modeling approaches tend to fall into the first category of pixel-wise models.Early approaches operated on the premise that the color of a pixel over time in a static scene could be modeled by a single Gaussian distribution,Nð ;ÆÞ.In their seminal work,Wren et al.[39] modeled the color of each pixel,Iðx;yÞ,with a single three-dimensional Gaussian,Iðx;yÞ$Nð ðx;yÞ;Æðx;yÞÞ.The mean ðx;yÞand the covarianceÆðx;yÞ,were learned from color observations in consecutive frames.Once the pixel-wise background model was derived,the likelihood of each incident pixel color could be computed and labeled as belonging to the background or not.Similar approaches that used Kalman Filtering for updating were proposed in[20] and[21].A robust detection algorithm was also proposed in [14].While these methods were among the first to principally model the uncertainty of each pixel color,it was quickly found that the single Gaussian pdf was ill-suited to most outdoor situations since repetitive object motion,shadows or reflectance often caused multiple pixel colors to belong to the background at each pixel.To address some of these issues,Friedman and Russell,and indepen-dently Stauffer and Grimson,[9],[33]proposed modeling each pixel intensity as a mixture of Gaussians,instead,to account for the multimodality of the“underlying”like-lihood function of the background color.An incident pixel was compared to every Gaussian density in the pixel’s model and,if a match(defined by threshold)was found,the mean and variance of the matched Gaussian density was updated,or otherwise a new Gaussian density with the mean equal to the current pixel color and some initial variance was introduced into the mixture.Thus,each pixel was classified depending on whether the matched distribu-tion represented the background process.While the use of Gaussian mixture models was tested extensively,it did not explicitly model the spatial dependencies of neighboring pixel colors that may be caused by a variety of real nominal motion.Since most of these phenomenon are“periodic,”the presence of multiple models describing each pixel mitigates this effect somewhat by allowing a mode for each periodically observed pixel intensity,however performance notably deteriorates since dynamic textures usually do not repeat exactly(see experiments in Section3).AnotherFig.1.Various sources of dynamic behavior.The flow vectors represent the motion in the scene.(a)The lake-side water ripples and shimmers.(b)The fountain,like the lake-side water,is a temporal texture and does not have exactly repeating motion.(c)A strong breeze can cause nominal motion(camera jitter)of up to25pixels between consecutive frames.limitation of this approach is the need to specify the number of Gaussians(models),for the E-M algorithm or the K-means approximation.Still,the mixture of Gaussian approach has been widely adopted,becoming something of a standard in background subtraction,as well as a basis for other approaches([18],[15]).Methods that address the uncertainty of spatial location using local models have also been proposed.In[7], El Gammal et al.proposed nonparametric estimation methods for per-pixel background modeling.Kernel den-sity estimation(KDE)was used to establish membership and,since KDE is a data-driven process,multiple modes in the intensity of the background were also handled.They addressed the issue of nominally moving cameras with a local search for the best match for each incident pixel in neighboring models.Ren et al.as well explicitly addressed the issue of background subtraction in a nonstationary scene by introducing the concept of a spatial distribution of Gaussians(SDG)[29].After affine motion compensation,a MAP decision criteria is used to label a pixel based on its intensity and spatial membership probabilities(both mod-eled as Gaussian pdf s).There are two primary points of interest in[29].First,the authors modeled the spatial position as a single Gaussian,negating the possibility of bimodal or multimodal spatial probabilities,i.e.,that a certain background element model may be expected to occur in more than one position.Although,not within the scope of their problem definition,this is,in fact,a definitive feature of a temporal texture.Analogous to the need for a mixture model to describe intensity distributions,unimodal distributions are limited in their ability to model spatial uncertainty.“Nonstationary”backgrounds have most re-cently been addressed by Pless et al.[28]and Mittal and Paragios[24].Pless et al.proposed several pixel-wise models based on the distributions of the image intensities and spatio-temporal derivatives.Mittal et al.proposed an adaptive kernel density estimation scheme with a joint pixel-wise model of color(for a normalized color space)and optical flow at each pixel.Other notable pixel-wise detection schemes include[34],where topology free HMMs are described and several state splitting criteria are compared in context of background modeling,and[30], where a(practically)nonadaptive three-state HMM is used to model the background.The second category of methods use region models of the background.In[35],Toyama et al.proposed a three tiered algorithm that used region based(spatial)scene information in addition to per-pixel background model: region and frame-level information served to verify pixel-level inferences.Another global method proposed by Oliver et al.[26]used eigenspace decomposition to detect objects. For k input frames of size NÂM a matrix B of size kÂðNMÞwas formed by row-major vectorization of each frame and eigenvalue decomposition was applied to C¼ðBÀ ÞTðBÀ Þ.The background was modeled by the eigenvectors corresponding to the largest eigenvalues, u i,that encompass possible illuminations in the field of view(FOV).Thus,this approach is less sensitive to illumination.The foreground objects are detected by projecting the current image in the eigenspace and finding the difference between the reconstructed and actual images.The most recent region-based approaches are by Monnet et al.[25]and Zhong and Sclaroff[40].Monnet and Scarloff and Zhong et al.simultaneously proposed models of image regions as an autoregressive moving average(ARMA) process,which is used to incrementally learn(using PCA) and then predict motion patterns in the scene.The foremost assumption made in background modeling is the assumption of a stationary scene.However,this assumption is violated fairly regularly,through common real-world phenomenon like swaying trees,water ripples, fountains,escalators,etc.The local search proposed in[7], the SDG of[29],the time series models of[25],[40],and KDEs over color and optical flow in[24]are several formulations proposed for detection nonstationary back-grounds.While each method demonstrated degrees of success,the issue of spatial dependencies has not been addressed in a principled manner.In context of earlier work (in particular,[24]),our approach falls under the category of methods that employ regional models of the background. We assert that useful correlation exists in the intensities of spatially proximal pixels and this correlation can be used to allow high levels of detection accuracy in the presence of general nonstationary phenomenon.1.2Proposed FormulationThe proposed work has three novel contributions.First,the method proposed here provides a principled means of modeling the spatial dependencies of observed intensities. The model of image pixels as independent random variables,an assumption almost ubiquitous in background subtraction methods,is challenged and it is further asserted that there exists useful structure in the spatial proximity of pixels.This structure is exploited to sustain high levels of detection accuracy in the presence of nominal camera motion and dynamic textures.By using nonpara-metric density estimation methods over a joint domain-range representation,the background data is modeled as a single distribution and multimodal spatial uncertainties can be directly handled.Second,unlike previous ap-proaches,the foreground is explicitly modeled to augment the detection of objects without using tracking information. The criterion of temporal persistence is proposed for simultaneous use with the conventional criterion of back-ground difference.Third,instead of directly applying a threshold to membership probabilities,which implicitly assumes independence of labels,we propose a MAP-MRF framework that competitively uses the foreground and background models for object detection,while enforcing spatial context in the process.2O BJECT D ETECTIONIn this section,we describe the novel representation of the background,the use of temporal persistence to pose object detection as a genuine binary classification pro-blem,and the overall MAP-MRF decision framework.For an image of size MÂN,let S discretely and regularly index the image lattice,S¼fði;jÞj1i N;1j M g: In context of object detection in a stationary camera,the objective is to assign a binary label from the set L¼f background;foreground g to each of the sites in S.2.1Joint Domain-Range Background ModelIf the primary source of spatial uncertainty of a pixel is image misalignment,a Gaussian density would be an adequate model since the corresponding point in the subsequent frame is equally likely to lie in any direction.However,in the presence of dynamic textures,cyclic motion,and nonstationary backgrounds in general,the “correct”model of spatial uncertainty often has an arbitrary shape and may be bimodal or multimodal,but structure exists because by definition,the motion follows a certain repetitive pattern.Such arbitrarily structured data can be best analyzed using nonparametric methods since these methods make no underlying assumptions on the shape of the density.Nonparametric estimation methods operate on the principle that dense regions in a given feature space,populated by feature points from a class,correspond to the modes of the “true”pdf .In this work,analysis is performed on a feature space where the p pixels are represented by x i 2I R 5,i ¼1;2;...p .The feature vector,x ,is a joint domain-range representation,where the space of the image lattice is the domain ,ðx;y Þand some color space,for instance ðr;g;b Þ,is the range [4].Using this representation allows a single model of the entire background,f R;G;B;X;Y ðr;g;b;x;y Þ,rather than a collection of pixel-wise models.Pixel-wise models ignore the dependencies between proximal pixels and it is asserted here that these dependencies are important.The joint representation provides a direct means to model and exploit this dependency.In order to build a background model,consider the situation at time t ,before which all pixels,represented in 5-space,form the set b ¼f y 1;y 2...y n g of the background.Given this sample set,at the observation of the frame at time t ,the probability of each pixel-vector belonging to the background can be computed using the kernel density estimator ([27],[31]).The kernel density estimator is a nonparametric estimator and under appropriate conditions the estimate it produces is a valid probability itself.Thus,to find the probability that a candidate point,x ,belongs to the background, b ,an estimate can be computed,P ðx j b Þ¼nÀ1X n i ¼1’H x Ày i;ð1Þwhere H is a symmetric positive definite d Âd bandwidthmatrix,and’H ðx Þ¼j H j À1=2’ðH À1=2x Þ;ð2Þwhere ’is a d -variate kernel function usually satisfying R ’ðx Þdx ¼1,’ðx Þ¼’ðÀx Þ,R x ’ðx Þdx ¼0,R xx T’ðx Þdx ¼I d and is also usually compactly supported.The d -variate Gaussian density is a common choice as the kernel ’,’ðN ÞH ðx Þ¼j H j À1=2ð2 ÞÀd=2expÀ12x T H À1x:ð3ÞIt is stressed here,that using a Gaussian kernel does not make any assumption on the scatter of data in the feature space.The kernel function only defines the effective regionof influence of each data point while computing the finalprobability estimate.Any function that satisfies the con-straints specified after (2),i.e.,a valid pdf,symmetric,zero-mean,with identity covariance,can be used as a kernel.There are other functions that are commonly used,some popular alternatives to the Gaussian kernel are the Epanechnikov kernel,the Triangular kernel,the Biweight kernel,and the Uniform kernel,each with their merits and demerits (see [38]for more details).Within the joint domain-range feature space,the kernel density estimator explicitly models spatial dependencies,without running into difficulties of parametric modeling.Furthermore,since it is well known that the rgb axes are correlated,it is worth noting that kernel density estimation also accounts for this correlation.The result is a single model of the background.Last,in order to ensure that the algorithm remains adaptive to slower changes (such as illumination change or relocation)a sliding window of length b frames is maintained.This parameter corresponds to the learning rate of the system.2.1.1Bandwidth EstimationAsymptotically,the selected bandwidth H does not affect the kernel density estimate but in practice sample sizes are limited.Too small a choice of H and the estimate begins to show spurious features,too large a choice of H leads to an over-smoothed estimate,losing important structural fea-tures like multimodality.In general,rules for choosing bandwidths are based on balancing bias and variance globally.Theoretically,the ideal or optimal H can be found by minimizing the mean-squared error,MSE f ^fH ðx Þg ¼E f½^f H ðx ÞÀf H ðx Þ 2g ;ð4Þwhere ^fis the estimated density and f is the true density.Evidently,the optimal value of H is data dependent since the MSE value depends on x .However,in practice,one does not have access to the true density function which is required to estimate the optimal bandwidth.Instead,a fairly large number of heuristic approaches have been proposed for finding H ,a survey is provided in [36].Adaptive estimators have been shown to considerably outperform (in terms of the mean squared error)the fixed bandwidth estimator,particularly in higher dimensional spaces [32].In general,two formulations of adaptive or variable bandwidth estimators have been considered [19].The first varies the bandwidth with the estimation point and is called the balloon estimator given byf ðx Þ¼1Xn i ¼1’H ðx Þðx Àx i ÞÞ;ð5Þwhere H ðx Þis the bandwidth matrix at x .The second approach,called the sample-point estimator,varies the bandwidth matrix depending on the sample pointf ðx Þ¼1Xn i ¼1’H ðx i Þðx Àx i ÞÞ;ð6Þwhere Hðx iÞis the bandwidth matrix at x i.However, developing variable bandwidth schemes for kernel density estimation is still research in progress,both in terms of theoretical understanding and in terms of practical algorithms[32].In the given application,the sample size is large,and although it populates a five-dimensional feature space,the estimate was found to be reasonably robust to the selection of bandwidth.Furthermore,choosing an optimal bandwidth in the MSE sense is usually highly computationally expensive. Thus,the balance between accuracy required(for matting, object recognition,or action recognition)and computational speed(for real-time surveillance systems)is application specific.To reduce the computational load,the Binned kernel density estimator provides a practical means of dramatically increasing computational speeds while closely approximat-ing the kernel density estimate of(1)([38],Appendix D).With appropriate binning rules and kernel functions the accuracy of the the Binned KDE is shown to approximate the kernel density estimate in[13].Binned versions of the adaptive kernel density estimate have also been provided in[32].To further reduce computation,the bandwidth matrix H is usually either assumed to be of the form H¼h2I or H¼diagðh21;h22;...h2dÞ.Thus,rather than selecting a fully parameterized bandwidth matrix,only two parameters can be defined,one for the variance in the spatial dimensions ðx;yÞand and one for the color channels,reducing computa-tional load.2.2Modeling the ForegroundThe intensity difference of interesting objects from the background has been,by far,the most widely used criterion for object detection.In this paper,temporal persistence is proposed as a property of real foreground objects,i.e., interesting objects tend to remain in the same spatial vicinity and tend to maintain consistent colors from frame to frame.The joint representation used here allows competitive classification between the foreground and background.To that end, models for both the background and the foreground are maintained.An appealing feature of this representation is that the foreground model can be constructed in a consistent fashion with the background model:a joint domain-range nonparametric density f¼f z1;z2...z m g. Just as there was a learning rate parameter b for the background model,a parameter f is defined for theFig.2.Foreground ing kernel density estimates on a model built from recent frames,the foreground can be detected in subsequent frames using the property of temporal persistence,(a)Current frame and(b)the X;Y-marginal,f X;Yðx;yÞ.High membership probabilities are seen in regions where foreground in the current frame matches the recently detected foreground.The nonparametric nature of the model allows the arbitrary shape of the foreground to be captured accurately(c)the B;G-marginal,f B;Gðb;gÞ,(d)the B;R-marginal,f B;Rðb;rÞ,and(e)the G;R-marginal, f G;Rðg;rÞ.foreground frames.However,since the foreground changes far more rapidly than the background,the learning rate of the foreground is typically much higher than that of the background.At any time instant the probability of observing a foreground pixel at any locationði;jÞof any color is uniform.Then,once a foreground region is been detected at time t,there is an increased probability of observing a foreground region at time tþ1in the same proximity with a similar color distribution.Thus,foreground probability is expressed as a mixture of a uniform function and the kernel density function,Pðx j fÞ¼ þð1À ÞmÀ1X mi¼1’HxÀz i;ð7Þwhere (1is the mixture weight,and is a random variable with uniform probability,that is R;G;B;X;Yðr;g;b;x;yÞ¼1RÂGÂBÂMÂN,where0r R,0g G,0b B,0x M, 0y N.This mixture is illustrated in Fig.3.If an object is detected in the preceding frame,the probability of observing the colors of that object in the same proximity increases according to the second term in(7).Therefore,as objects of interest are detected(the detection method will be explained presently)all pixels that are classified as“interesting”are used to update the foreground model f.In this way, simultaneous models are maintained of both the background and the foreground,which are then used competitively to estimate interesting regions.Finally,to allow objects to become part of the background(e.g.,a car having been parked or new construction in an environment),all pixels are used to update b.Fig.2shows plots of some marginals of the foreground model.At this point,whether a pixel vector x is“interesting”or not can be competitively estimated using a simple likelihood ratio classifier(or a Parzen Classifier since likelihoods are computed using Parzen density estimates,[10]),¼Àln Pðx j bÞPðx j fÞ¼ÀlnnÀ1P ni¼1’HxÀy iþð1À ÞmÀ1P mi¼1’HxÀz i:ð8ÞThus,the classifier is,ðxÞ¼À1ifÀlnPðx j bÞPðx j fÞ>1otherwise;where is a threshold which balances the trade-off betweensensitivity to change and robustness to noise.The utility inusing the foreground model for detection can be clearlyseen in Fig.4.Fig.4e shows the likelihood values based onlyon the background model and Fig.4f shows the likelihoodratio based on both the foreground and the backgroundmodels.In both histograms,two processes can be roughlydiscerned,a major one corresponding to the backgroundpixels and a minor one corresponding to the foregroundpixels.The variance between the clusters increases with theuse of the foreground model.Visually,the areas corre-sponding to the tires of the cars are positively affected,inparticular.The final detection for this frame is shown inFig.8c.Evidently,the higher the likelihood of belonging tothe foreground,the lower the overall likelihood ratio.Fig. 3.Foreground likelihood function.The foreground likelihood estimate is a mixture of the kernel density estimate and a uniform likelihood across the five-spaceof features.This figure shows aconceptualization as a1D function.Fig.4.Improvement in discrimination using temporal persistence.Whitervalues correspond to higher likelihoods of foreground membership.(a)Video Frame410of the Nominal Motion Sequence(b)Log-Likelihood Ratio values obtained using(8).(c)Foreground likelihoodmap.(d)Background negative log-likelihood map.(e)Histogrammednegative log-likelihood values for background membership.The dottedline represents the“natural”threshold for the background likelihood,i.e.,logð Þ.(f)Histogrammed log-likelihood ratio values.Clearly,the variancebetween clusters is decidedly enhanced.The dotted line represents the“natural”threshold for the log-likelihood ratio,i.e.,zero.。
E第四节-翻译的功能学派

The first to introduce into TS a consideration of the communicative purpose of translation.
The TT of an expressive tex“t should transmit the aesthetic and artistic form of the ST.
The TT of an operative text should produce the desired response in the TT receiver.
(Newmark, A Textbook of Translation, 1988)
Newmark admits that few texts are purely expressive, informative or vocative (1988: 47)
Reiss & Vermeer: Groundwork for a General Theory of Translation (《翻译理论基础》1984):
Audio-medial texts require the “supplementary” method.
Text types and varieties (genres)
cf. New Mark’s theory on text types
Three main types of texts:
An adequate translation should be “adequate to” the requirements of the transaltion brief.
- 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)。
滨江学院毕业论文题目On the Features, Functions and Translation Methodsof Public Signs院系外语系专业英语学生姓名钱雯雯学号20072310935指导老师夏杨称职讲师二O一一年五月二十日On the Features, Functions and Translation Methods ofPublic SignsBy Qian W enwenIn partial fulfillment of the requirementFor the B. A. degreeDepartment of EnglishBinjiang CollegeNanjing University of Information Science and TechnologyMay, 2011AcknowledgementsI would like to express my gratitude to all those who helped me during the writing of this thesis.My deepest gratitude goes first and foremost to my supervisor Xia Y ang, for his valuable suggestions and constant encouragement in the academic studies. I do appreciate her patience and professional instructions during my thesis writing.Second, I am also greatly indebted to the professors and teachers at the Department of English, from whose devoted teaching and enlightening lectures I have benefited a lot and academically prepared for the thesis.Third, I owe much to my friends and classmates for their valuable suggestions and critiques which are of help and importance in making the thesis a reality.Last but not least, I would like to ex press my gratitude to my beloved parents who have always been helping me out of difficulties and supporting without a word of complaint.Contents Abstract (1)摘要 (1)1.Introduction (1)1.1 Definition of Public Signs (2)1.2 Applied range and significance of Public Signs (2)1.3 Current Situation and Existent Problems of Public Signs (3)2. The Functions and Features of Public Signs (4)2.1 Functions (4)2.1.1 Function of Directing (4)2.1.2 Function of Prompting (4)2.1.3 Function of Restricting (4)2.1.4 Function of Compelling (5)2.2 Features (5)2.2.1 Extensive Use of Nouns, V erbs and Gerunds (5)2.2.2 Use of Phrases and Abbreviations (6)2.2.3 Combination of Written Words with Pictures (6)2.2.4 Use of Present Tense (7)2.2.5 Use of Imperative Sentences (8)2.2.6 No Uncommon Words (8)2.2.7 Conciseness and Accuracy of Public Signs (8)3. Translation Methods of Public Signs (8)3.1 Following Established Translation (8)3.1.1 Established Translation of Road and Street (9)3.1.2 Established Translation of Business and Service Industry (9)3.1.3 The Principle of Consistency (9)3.1.3.1 Consisting with the International Idiomatic Public Signs ofEnglish (9)3.1.3.2 Consisting with the Fixed Translation (10)3.1.3.3 Consisting with the Official Websites and AuthoritativeDictionaries or Newspapers (10)3.2 Adding and Omitting (11)3.3 Literal and Free Translation (11)4. Typical Errors and Solutions (12)4.1 Collection of some Typical Errors (12)4.1.1 Errors of non-standard writing (12)4.1.2 Errors of non-standard grammar (12)4.1.3 Errors of “Interpret without Real U nderstanding” (13)4.1.4 Errors of ignoring the cultural differences (13)4.1.5 Errors of violating the established usage (14)4.2 Causes of Errors in Translation (14)4.2.1 Chinglish (14)4.2.2 Inappropriate Wording (15)4.2.3 Lack in the Knowledge of Public Signs (15)4.3 Possible Solutions (15)4.3.1 Pay attention to cross-cultural awareness (15)4.3.2 Analyze the context of words and avoid Chinglish (16)4.3.3 Strive to be easy-to-understand (16)5. Conclusion (16)References (17)On the Features, Functions and Translation Methods ofPublic SignsQian WenwenDepartment of EnglishSchool of Language and CultureNanjing University of Information Science and TechnologyAbstract:Public signs are widely used in our daily lives and essential to the public and tourists. However, there are many wrongly used public signs around us. Any abuse or ambiguity of public signs will cause undesirable consequences. This paper discusses public signs from Definition, Applied Range, Significance, Current Situation and Existent Problems. Directing, Prompting, Restricting and Compelling are four of their most important functions; Extensive use of Noun, V erb and Gerund, Use of Phrase and Abbreviation, Combination of Written Words with Pictures, Use of Present Tense, Use of Imperative Sentence, No Uncommon Words, Conciseness and Accuracy of Public Signs are their outstanding features of public signs. This paper also probes into its translation methods after analyzing its functions and features, in order to shed light on our understanding and translation of public signs.Key Words: public signs; functions; features; translation methods摘要:公示语广泛应用于我们生活的方方面面,对公众和游客至关重要。
然而,我们周围存在许多错误使用的公示语。
任何滥用或歧义都会导致不良的后果。
本文根据公示语的定义、应用范围和意义及现状和存在的问题对其做介绍。