Wilamowski B.M., Irwin J.D. The Industrial Electronics Handbook. Second Edition: Industrial Communication Systems

Подождите немного. Документ загружается.

26-14 Industrial Communication Systems

references

. 1.. ANSI/EIA/CEA-709.1-A,. Control. network. protocol. specification,. ANSI,. Washington. DC,.

199

. 2.. ANSI/EIA/CEA.852,.Tunne

lling

.comp

onent

.netw

ork

.pro

tocols

.over.Int

ernet

.pro

tocol

.cha

nnels,

ANSI,.Wa

shington

.DC,.2002.

. 3.. EN.133

21-2,

.Ope

.dat

.com

munication

.in.bui

lding

.aut

omation,

.con

trols

.and.bui

lding

.man

agement-

Hom

.and.bui

lding

.ele

ctronic

.sys

tems—Part

.2:.KNX

net/IP

.com

munication,

.CEN

.Bru

ssels,

Belg

ium,

.20

06.

. 4.. EN.1434-3,.He

.met

ers—Data

.exc

hange

.an

.int

erfaces,

.CEN,.Br

ussels,

.Be

lgium,

.1997.

. 5.. EN.14908,.Open.dat

.communic

ation

.in.buildin

.aut

omation,

.cont

rols

.and.buildin

.man

agement.

Con

trol

.net

work

.pr

otocol,

.CEN,.Br

ussels,

.Be

lgium,

.2005.

. 6.. EN. 500

90,

. Hom

. and. bui

lding

. ele

ctronic

. sys

tems

. (HB

ES),

. CEN

ELEC,

. Bru

ssels,

. Bel

gium,

199

4–2009.

. 7.. IEC. 609

29,

. AC-

supplied

. ele

ctronic

. bal

lasts

. for. tubu

lar

. uo

rescent

. lam

ps,

. IEC

. Gen

eva,

Swi

tzerland.

. 8.. ISO. 16484-2,. Buildin

. aut

omation

. and. cont

rol

. system

. (BACS)—

Part

. 2:. Har

dware,

. Gene

va,

Swi

tzerland,

.2004.

. 9.. ISO.16484-5,.Buildin

.aut

omation

.and.cont

rol

.system

s—Part

.5:.Dat

.communic

ation

.pro

tocol,

ISO,.Gen

eva,

.Sw

itzerland,

.2007.

. 10.. ISO/IEC. 14543,. Info

rmation

. tec

hnology—Home

. elec

tronic

. system. (HES). arc

hitecture,

. ISO,.

Geneva,.Switzerland,.2006-2010.

. 11.. ASHR

.St

anding

.St

andard

.Pr

oject

.Co

mmittee

.135,.ww

w.bacnet.org

. 12.. Pro

le

.fo

.We

.Ser

vices,

.sc

hemas.xmlsoap.org/ws/2006/02/devprof

. 13.. Digit

alstrom.org,

.ww

w.digitalstrom.org

. 14.. Digit

.Li

ving

.Net

work

.Al

liance,

.ww

w.dlna.org/home

. 15.. Ech

onet

.Co

nsortium,

.ww

w.echonet.gr.jp

. 16.. EnOcea

.Gm

bH,

.ww

w.enocean.com

. 17.. Hom

ePlug

.Al

liance,

.ww

w.homeplug.org

. 18.. IGRS,.ww

w.igrs.org

. 19.. OPC.Fo

undation,

.ww

w.opcfoundation.org

. 20.. Konn

.As

sociation,

.ww

w.knx.org

. 21.. LonM

ark

.In

ternational,

.ww

w.lonmark.org

. 22.. OSGi.Al

liance,

.ww

w.osgi.org

. 23.. e.SMI.Gr

oup,

.ww

w.smi-group.com

. 24.. UPnP.Fo

rum,

.ww

w.upnp.org

. 25.. ZigBe

.Al

liance,

.ww

w.zigbee.org

. 26.. Z-Wave.Alliance,.www.z-wavealliance.org

. 27.. W..Gr

anzer,

.W..Ka

stner,

.G..Ne

ugschwandtner,

.an

.F..Pr

aus,

.Se

curity

.in.ne

tworked

.bu

ilding

.au

tomation

system

.Pro

ceedings of the Sixth IEEE International Workshop on Factory Communication Systems,

pp..28

3–292,

.Tor

ino,

.Ita

ly,

.Jun

.200

. 28.. W..Kas

tner,

.G..Neugsc

hwandtner,

.S..Souce

.and.H.. M.. New

man,

.Comm

unication

.system

.for.

buildin

.au

tomation

.an

.con

trol,

.Pr

oceedings of the IEEE,

.93(6):.1178–1203,.2005.

. 29.. W..Kas

tner

.and.G..Neugsc

hwandtner,

.Dat

a Communication for Distributed Building Automation,

CRC.Pr

ess,

.Bo

.Ra

ton,

.FL,.2009.

. 30.. H..M..New

man,

.Dir

ect Digital Control of Building Systems: eory and Practice,

.John.Wile

.&.Son

New.Yo

rk,

.1994.

Building and Home Automation 26-15

. 31.. T..Novak,.A..Treytl,.and.P..Palensky,.Common.approach.to.functional.safety.and.system.security.

in.buildin

.aut

omation

.and.cont

rol

.system

.12th IEEE

Conference on Emerging Technologies and

Factory Automation,

.pp

.1141–1148,.Pa

tras,

.Gr

eece,

.Sep

tember

.2007.

. 32.. C..Reinis

ch,

.W..Kas

tner,

.G..Neugsc

hwandtner,

.and.W..Granzer

.Wire

less

.tec

hnologies

.in.hom

.and.

buildin

.aut

omation,

.Pro

ceedings of the Fih IEEE International Conference on Industrial Informatics,

pp..93–98,.Pra

gue,

.Czec

.Rep

ublic,

.Jun

.23–27,.2007.

. 33.. S..Souce

.and.T..Saut

er,

.Qua

lity

.of.serv

ice

.concer

.in.IP-ba

sed

.cont

rol

.system

.IEEE Tr

ansactions

on Industrial Electronics,

.51(6):.1249–1258,.2004.

27-1

27.1 Introduction

It.is.commonly.accepted.that.the.multimedia.applications.that.appeared.in.the.mid-1990s.are.the.third.

gene

ration

.of.comp

uter

.appl

ications

.[P98

. e.rst.gene

ration

.was.char

acterized

.by.its.abil

ity

.to.man-

.data

.whil

.the.seco

.gene

ration

.was.consi

dered

.to.be.prin

cipally

.for.comm

unication.

. is.has.had.

an.inu

ence

.on.comp

uter

.inst

ruction

.set.arch

itecture

.(ISA

)

.in.curr

ent

.proc

essors

.with.the.intr

oduction

of.mult

imedia

.exte

nsions

.(e.g

.MMX.and.SSE.in.Inte

.or.in.the.deve

lopment

.and.den

ition

.of.LAN.and.

WAN.netw

orks

.that.were.able.to.prov

ide

.qual

ity

.of.serv

ice

.(QoS

)

.to.this.tra

.(suc

.as.IEEE.802.

11e).

 e.area.of.indu

stry

.has.not.been.an.exce

ption

.to.this.phen

omenon

.and,.from.the.begi

nning,

.the.deve

opment

.of.new.indu

strial

.appl

ications

.[P98

,RSB99,WIF01]

.was.pred

icted

.both.for.the.high.degr

.of.

inte

raction

.betw

een

.huma

.and.indu

strial

.proc

ess

.that.they.allo

.and.for.thei

.part.in.the.deve

lopments

in.moni

toring

.and.cont

rol,

.qual

ity

.cont

rol,

.fact

ory

.auto

mation,

.and.fact

ory

.comm

unication.

.Nowa

days

.it.is.

poss

ible

.to.nd.mult

imedia

.appl

ications

.in.near

.all.area

.of.indu

strial

.comm

unications—automation,

moni

toring,

.ima

.pro

cessing,

.rob

otics,

.rem

ote

.con

trol,

.etc

 ese

.appl

ications

.have.a.wide.rang

.of.requ

irements,

.and.ther

.are.nume

rous

.tech

nologies

.that.sup-

ort

.them

.Data.comp

ression

.tech

niques

.and.tech

niques

.for.tran

smission

.of.this.kind.of.info

rmation

are.amon

.the.most.impo

rtant

.of.thes

.tech

nologies.

.Howe

ver,

.it.is.nece

ssary

.to.rst.consi

der

.the.stru

ture

.of.the.data.mana

ged

.in.this.kind.of.appl

ication.

.Imag

.or.imag

.sequ

ences

.have.a.heig

.and.widt

call

.reso

lution,

.whic

.is.meas

ured

.in.pixe

ls.

. ese.pixe

.repr

esent

.info

rmation

.for.whic

.a.spec

ic

numbe

.of.bits.per.pixe

.(bpp

)

.is.requ

ired,

.the.most.comm

.bein

.8.bits.in.gray

scale

.and.15.(5.bits.for.

each.colo

.plan

e),

.24,.or.32.in.colo

.Reso

lution,

.toget

her

.with.the.nece

ssary

.rate.of.fram

.per.seco

(fps).provides.the.bit.rate.necessary.for.the.application,.normally.measured.in.bits.per.second.(bps)..

 e.most.comm

only

.used.form

.in.indu

strial

.and.moni

toring

.appl

ications

.is.8.bpp.in.gray

scale,

.and.

when.work

ing

.in.colo

.the.RGB.(red

.gree

.blue

)

.form

at,

.using.8.bits.for.each.colo

.plan

. is.mean

24.bpp.are.used.for.colo

.as.can.be.seen.in.Figu

.27.1

.Howe

ver,

.the.use.of.the.YUV.colo

.form

.with.

subs

ampling

.is.comm

on,

.and.this.is.gene

rally

.used.in.the.comp

ression

.of.colo

.imag

es.

. is.form

at,

Industrial Multimedia

27.1. Introduction.....................................................................................27-1

27.2

. Mult

imedia

.Com

pression:

.A.Rev

iew............................................27-4

Image.Compressors. •. Video.Compressors. •. Quality.Evaluation

27.3. Industrial.Multimedia.Applications.............................................27-8

Monitoring.Applications. •. Computer.Vision.Applications

27.4. Image.Transmission........................................................................27-9

IEEE.1394. •. IP-Based.Networks

27.5. Conclusions.................................................................................... 27-11

Acknowledgment....................................................................................... 27-11

References................................................................................................... 27-11

Javier

Silvestre-Blanes

Universidad Politécnica

de Val

encia

Manfred Weihs

TTTech

Comp

utertechnik

Víctor-M.

Sempere-Payá

Universidad Politécnica

de Val

encia

27-2 Industrial Communication Systems

used.in.analog.signals,.has.one.plane.used.for.luminance.(Y.or.brightness).and.the.other.two.used.by.the.

chrominance.component.(U.and.V,.the.information.on.color)..It.is.known.that.the.human.vision.system.

(HVS).is.more.sensitive.to.the.quality.of.the.Y.plane.than.the.U.and.V.planes,.so.subsampling.is.applied.

to.reduce.the.amount.of.information.in.analog.signals.and.can.be.seen.as.a.rst.stage.of.compression.

.resolution.and.the.fps.rate.are.established.by.the.type.of.application..Once.established,.these.fac-

tors

.inuence.the.requirements.of.the.image’s.characteristics,.the.spatial.relationship.of.the.equipment.

participating.in.the.application,.and.the.implications.for.the.temporal.requirements.that.must.be.

satised..Although.there.is.a.wide.range.of.multimedia.applications.being.used.in.factory.automa-

tion,

.such.as.virtual.manufacturing.[WIF01].or.online.interactive.training,.we.consider.[SS07].process.

monitoring.and.control.through.image.processing.to.be.the.most.relevant..In.the.case.of.monitoring.

processes,.where.an.application’s.temporal.requirements.are.not.critical,.the.sensor.and.the.visualiza-

tion

.node.can.be.in.the.same.machine,.in.the.same.plant,.or.in.any.part.of.the.world..In.the.case.of.auto-

mated

.control,.where.applications.do.have.critical.temporal.requirements,.sensor.and.processing.node.

are.normally.in.the.same.machine.(personal.area.network,.distances.of.a.few.meters),.but.other.solutions.

are.possible.as.well..Reductions.in.costs.and.the.increases.in.processing.capacity.are.giving.rise.to.a.new.

generation.of.intelligent.cameras,.where.capture.and.image.processing.are.integrated.in.the.same.device.

(distances.in.centimeters)..On.the.other.hand,.this.area.is.beginning.to.see.the.inuence.of.the.expand-

ing

.use.of.Ethernet.to.areas.outside.the.oce,.and.this.is.leading.to.the.use.of.IP.technology.in.image.

transmission.over.distances.of.hundreds.meters.normally,.but.also.over.greater.distances,.when.there.

are.less.strict.temporal.requirements.

.Figure.27.2,.the.use.of.different.types.of.communication.networks.used.in.multimedia.appli-

cations

.depending.on.communication.distances.between.sensor.and.processing.node.can.be.seen.

RGB 4:2:2

4:2:0 4:1:1

pixel (0,0) pixel (1,0) pixel (2,0) pixel (3,0) pixel (0,0) pixel (1,0) pixel (2,0) pixel (3,0)

pixel (0,1) pixel (1,1) pixel (2,1) pixel (3,1) pixel (0,1)

pixel (1,1)

pixel (2,1) pixel (3,1)

pixel (0,2) pixel (1,2) pixel (2,2) pixel (3,3) pixel (0,2) pixel (1,2) pixel (2,2) pixel (3,3)

pixel (0,3)

pixel (1,3)

pixel (2,3)

pixel (3,3)

pixel (0,3)

pixel (1,3)

pixel (2,3)

pixel (3,3)

pixel (0,0) pixel (1,0) pixel (2,0) pixel (3,0) pixel (0,0) pixel (1,0) pixel (2,0) pixel (3,0)

pixel (0,1)

pixel (1,1)

pixel (2,2)

pixel (3,1)

pixel (0,1) pixel (1,1) pixel (2,1)

pixel (3,1)

pixel (0,2) pixel (1,2) pixel (2,1) pixel (3,3) pixel (0,2) pixel (1,2) pixel (2,2) pixel (3,3)

pixel (0,3)

pixel (1,3)

pixel (2,3)

pixel (3,3)

pixel (0,3) pixel (1,3) pixel (2,3)

pixel (3,3)

B B

FIGURE 27.1 Multimedia.information.

Industrial Multimedia 27-3

These.networks.have.transmission.capacities.ranging.from.kbps.to.Gbps..Image.resolutions.can.

also.be.wide.ranging,.as.can.be.seen.in.Table.27.1,.where.some.of.the.currently.existing.resolutions.

are.enumerated.and.where.the.bandwidth.necessary.for.raw.transmission.with.a.rate.of.30.images/s,.

expressed.in.megabit.per.second.(Mbps),.is.shown.

.monitoring.applications,.there.is.more.freedom.to.choose.resolutions.and.capture.rates,.but.with.

some.restrictions..For.example,.in.a.security.application.in.which.the.movement.of.people.is.being.

monitored,.the.capture.speed.must.ensure.that.someone’s.passing.is.captured.in.at.least.one.image..

Further,.the.resolution.used.will.determine.if.their.face.can.be.identied.or.not..In.control.applications.

there.is.less.freedom,.and.these.parameters.are.strictly.dened.by.the.application.requirements..If.the.

objects.being.monitored.move.at.a.speed.of.600.objects/min,.at.least.10.images/s.must.be.captured..If.

the.objects.have.a.dimension.of.500.×.500.mm.and.4.pixels/mm.are.needed.to.guarantee.the.target.of.

TABLE 27.1 Image.Resolutions.in.Multimedia.Applications

Name Resolution Mbps Name Resolution Mbps

SQCIF 128

.×.96 8.84 QCIF 176.×.144 18.2

CIF 352

.×.288 72.9 4CIF 704.×.576 291.9

QQVGA 160

.×.120 13.8 QVGA 320.×.240 55.2

VGA 640

.×.480 221.1 SVGA 800.×.600 345.6

XGA 1024

.×.768 556.2 UXGA 1600.×.1200 1,382.4

QXGA 2048

.×.1536 2264.9 SXGA 1280.×.1024 943.7

QSXGA 2560

.×.2048 2831.1 HSXGA 5120.×.4096 15,099.5

WVGA 852

.×.480 294.4 WXGA 1366.×.768 755.34

WSXGA 1600

.×.1024 1179.6 WUXGA 1920.×.1200 1,658.8

WOXGA 2560

.×.1600 2949.1 WQSXGA 3200.×.2048 4,718.6

WQUXGA 3840

.×.2400 6635.5 WHSXGA 6400.×.4096 18,874.36

WHUXGA 7680

.×.4800 26542.0 HD480 852.×.480 294.4

HD720 1280

.×.720 663.5 HD1080 1920.×.1080 1,492.9

Multimedia

source

Multimedia

processing node

PAN

....................

RDSI

ADSL

SDSL

WiMAX

MAN/WAN

LAN

Ethernet

Fast-Ethernet

WiFi

WiMAX

Analog:

RS-4xx

Digital

USB

Firewire

Cameralink

Bluetooth

FIGURE 27.2 Multimedia.communications.in.industrial.environment.

27-4 Industrial Communication Systems

the.image.processing.application,.then.2000.×.2000.pixels.will.be.necessary,.captured.by.a.sensor.with.

a.sucient.resolution.. is.means.that.a.transmission.capacity.of.at.least.approximately.960.Mbps.is.

required..In.the.case.of.industrial.applications.where.3D.information.is.needed,.this.value.has.to.be.mul-

tiplied

.by.the.number.of.cameras.needed..Due.to.these.high.data.rates,.it.is.sometimes.necessary.to.use.

image.compression/decompression,.although.these.operations.can.introduce.latency.into.the.process.

Depending

.on.the.eectiveness.of.the.compression.and.also.the.limits.imposed.by.the.type.of.applica-

tion,

.this.process.will.produce.a.compression.ratio,.normally.expressed.in.X:1,.being.X.the.number.of.bytes.

that.can.be.represented.using.only.1.byte,.or.in.bpp,.thus.expressing.the.reduction.as.well..For.example,.an.

image.of.800.×.600.pixels.with.24.bpp.requires.11,520,000.bits,.which.is.1.44.Mbyte..If.we.compress.using.

a.5:1.ratio,.there.will.be.2,304,000.bits,.giving.an.equivalent.compression.ratio.of.4.8.bpp.

27.2 Multimedia Compression: a review

 e.aim.of.image.compression.is.to.reduce.the.application’s.bandwidth.requirements.(or.storage.in.other.

application.types).without.distorting.signicantly.the.original.information,.although.this.causes.laten-

cies,

.as.it.can.be.seen.in.Figure.27.3..Distortion.or.errors.generated.by.the.compression.and.decompres-

sion

.process.of.the.image.must.be.limited.and.have.to.be.measured:

e x y f x y g x y

( , ) ( , ) ( , )

= −

(27.1)

.Figure.27.3,.we.can.see.how.the.original.image.must.go.through.a.series.of.steps.to.achieve.a.reduction.

of.the.information.to.be.transmitted.. is.will.cause.a.latency.T

.(compression.time).that.will.depend.on.

the.complexity.of.the.algorithm.used.. e.compressed.information.must.then.be.transmitted.through.

the.network,.which.means.it.is.necessary.to.packetize.the.information.and.wait.a.period.of.time.that.

will.depend.on.the.scheduling.policy.of.the.source.and.the.priority.of.the.multimedia.information.with.

regard.to.other.information.that.is.using.the.same.medium.(e.g.,.an.alarm.could.have.more.priority)..

Finally,.we.have.to.access.the.medium.and.carry.out.the.transmission.of.the.information.. is.process.

will.introduce.a.latency.T

.(transmission.time).that.will.be.smaller.the.more.ecient.the.compression.

process.has.been.. ere.is.also.an.unpack/reception.time,.but.this.is.almost.insignicant,.and.above.all,.

a.decompression.time.(T

).depending.on.the.chosen.compressor.

 ere

.are.several.classications.of.compressors,.depending.on.the.distortion.that.they.introduce.in.

the.source,.the.complexity.of.the.transmitter.and.receiver,.and.the.type.of.redundancy.employed..In.

the.rst.place,.there.are.compressors.with.or.without.losses..Lossless.compressors.do.not.distort.the.

image,.that.is,.f(x,.y).=.g(x,.y).. ese.compressors.are.not.used.in.excess.because.the.rate.of.compression.

they.can.achieve.is.approximately.2:1,.which.restricts.their.use.in.the.industrial.area..However,.they.are.

used.in.other.types.of.applications,.such.as.medical.applications,.where.degradation.that.can.distort.the.

medical.diagnosis.is.not.tolerated.and.where.transmission.latencies.are.not.critical..In.other.types.of.

Image

preparation

Access

medium

Packetization

Transmission

Scheduling

Reverse

process

f (x, y)

Quantization

Encoding

image

Motion

analysis

Compression Transmission

Uncompression

Received image g(x, y)

Original image

Transformation

domain

FIGURE 27.3 General.multimedia.application.scheme.

Industrial Multimedia 27-5

application.where.loss.of.information.is.tolerated,.lossy.compressors.are.used,.which.allow.a.signicant.

inc

rease

.in.the.com

pression

.rat

.Con

cerning

.the.com

plexity

.of.the.pro

cess,

.thr

ough

.an.asy

mmetric

app

roach,

.it.is.pos

sible

.to.use.qui

.com

plex

.alg

orithms

.in.the.sou

rce

.to.opti

mize

.the.qua

lity

.and.bit.

rat

.(th

.can.inv

olve

.ful

.two.pas

ses),

.whi

.the.dec

ompression

.is.per

formed

.in.rea

.tim

.How

ever,

thi

.app

roach

.can

not

.be.use

.in.liv

.rea

l-time

.app

lications,

.whi

.are.the.mor

.com

mon,

.so.usu

ally

sym

metric

.app

roaches

.are.use

.whi

.hav

.a.simi

lar

.com

plexity

.in.com

pression

.and.unc

ompression.

Fin

ally,

.the

.are.com

pressors

.tha

.onl

.use.the.spa

tial

.redu

ndancy

.of.ima

ges,

.tha

.is,.of.the.self

-contained

inf

ormation

.in.an.ima

ge,

.and.com

pressors

.tha

.also.use.the.tem

poral

.redu

ndancy,

.tha

.is,.the.redu

ndancy

cre

ated

.by.sim

ilarities

.be

tween

.co

nsecutive

.fr

ames.

Spatial

.red

undancy

.exp

loits

.the.pro

ven

.sta

tistical

.cor

relation

.bet

ween

.the.pix

els

.of.an.ima

ge,

.so.it.

all

ows

.us.to.ext

ract

.a.pix

el’s

.val

.fro

.the.nei

ghboring

.pix

els.

.With.thi

.tec

hnique,

.it.is.not.nec

essary

to.rep

roduce

.eac

.pix

.of.an.ima

.ind

ependently.

.In.Fig

ure

.27.

.we.can.see.two.sce

nes

.of.ind

ustrial

mon

itoring

.[SS

A04].

.On.the.le

.the

.are.wid

.are

.in.the.ima

.tha

.hav

.a.simi

lar

.val

ue,

.whi

.is.

use

.by.com

pressors

.to.red

uce

.the.nec

essary

.bit

.to.rep

roduce

.the.inf

ormation.

.On.the.le

.all.pix

els

.in.

the.int

erval

.[28

–32]

.hav

.bee

.rep

resented

.in.whi

te,

.whi

.giv

.an.ide

.of.the.spa

tial

.red

undancy

.and.

its.po

tential

.to.re

duce

.th

.nu

mber

.of.bi

.to.re

present

.th

.mu

ltimedia

.in

formation.

.tem

poral

.red

undancy

.com

.fro

.the.fac

.tha

.a.cor

relation

.bet

ween

.pix

els

.of.con

secutive

.ima

ges

in.a.seq

uence

.exi

sts.

.Bet

ween

.con

secutive

.fra

mes,

.les

.tha

.10%.of.pix

els

.cha

nge

.the

.val

.in.abo

1%.of.the.pea

.sign

al.

.In.thi

.way

.ima

ges

.pix

els

.can.be.pre

dicted

.fro

.the.pix

els

.of.an.ima

.nea

rby

.in.

the.sequence..On.the.other.hand,.the.fact.that.changes.between.consecutive.images.are.caused.by.the.

mov

ement

.of.som

.obje

cts

.in.the.sce

.has.cau

sed

.the.dev

elopment

.of.mot

ion

.com

pensation

.cod

ing

tec

hniques.

.Fig

ure

.27.

.sho

.thi

.in.an.ind

ustrial

.mon

itoring

.seq

uence

.capt

ured

.at.25.fra

mes/s.

.As.we.

can.see

.the.di

erence

.bet

ween

.con

secutive

.ima

ges

.is.ver

.lit

tle

.in.thi

.cas

.and.for.thi

.rea

son,

.in.thi

typ

.of.se

quence,

.th

.us

.of.te

mporal

.re

dundancy

.al

lows

.us.to.ob

tain

.hi

.ra

tes

.of.co

mpression.

FIGURE 27.4 Spatial.redundancy.examples.

27-6 Industrial Communication Systems

Whichever.method.the.compressor.uses,.it.is.still.possible.to.make.use.of.another.redundancy.of.the.data.

extracted—the.code.redundancy.. is.is.related.to.the.reproduction.of.information.with.the.same.coding.

and.for.which.there.are.techniques.that.allow.us.to.reduce.the.number.of.bits.per.symbol,.based.on.tech-

niques

.that.allow.the.use.of.very.few.bits.to.reproduce.the.most.common.symbols,.and.more.bits.for.those.

symbols.that.appear.less.frequently,.such.as.Human.coding,.arithmetic.code,.and.variable-length.coding.

27.2.1 Image Compressors

Still.image.compressors.were.developed.in.mid-1980s,.with.International.Telecommunication.Union.(ITU).

and.International.Organization.for.Standardization.(ISO),.followed.by.ISO/IEC.international.standard.

JPEG.(Joint.Photographic.Experts.Group).in.1992.[ISO94].. is.standard.takes.in.four.compression.modes,.

although.not.all.of.them.are.always.supported.. ese.modes.are.sequential.discrete.cosine.transformation.

(DCT)-based.mode,.progressive.DCT-based.mode,.hierarchical.mode,.and.lossless.mode.. e.rst.is.the.

baseline.of.JPEG.and.the.most.common.one..Progressive.and.hierarchical.modes.enable.a.progressive.pre-

sentation

.of.the.image,.and.the.latter.is.also.useful.in.applications.with.multiresolution.requirements.. e.

standard.JPEG.prepares.the.image.by.carrying.out.a.color.space.transformation.from.RGB.to.YUV.4:2:2.or.

4:1:1,.and.the.data.unit.is.a.block.of.8.×.8.pixels.. ese.units.are.transformed.through.a.DCT.obtaining.64.

DCT.coecients.. ese.coecients.are.mapped.to.integer.numbers.and,.depending.on.the.quality.speci-

ed,

.the.coecients.lower.than.a.given.value.are.set.to.0.. ese.coecients.are.ordered.in.a.“zig-zag”.form.

and.coded.using.Human.followed.by.a.runlength.coding.. e.standard.JPEG2000.[CSE00].tries.to.solve.

most.of.JPEG’s.limitations.while.at.the.same.time.providing.a.signicant.quality.improvement.for.the.same.

compression.rates..To.achieve.this.objective,.it.uses.more.advanced.techniques.in.each.stage.of.the.compres-

sion.

.In.this.way,.instead.of.using.the.DCT.transform,.it.uses.wavelet.transformation.on.the.tile.compo-

nents,

.which.gives.dierent.decomposition.levels,.each.of.them.containing.a.number.of.subbands.with.the.

transformation.coecients.. e.transformed.coecients.of.each.subband.are.arranged.in.code.blocks.and.

the.symbols.produced.by.a.bit-plane.coding.technique.are.coded.through.an.adaptative.binary.arithmetic.

code..Today,.it.is.generally.accepted.that.this.standard.only.outperforms.JPG.signicantly.when.very.high.

compression.rates.have.to.be.used,.since.the.computation.complexity.is.quite.high.compared.to.JPEG.

27.2.2 Video Compressors

 ere.are.two.main.families.of.video.compressors,.the.ITU.standards,.denominated.H.26x,.and.the.

ISO.Motion.Picture.Expert.Group.standards,.denominated.MPEGx.. e.rst.ITU.standard.was.H.261,.

i2-i1 i9-i1

i6 i7 i8 i9

Time

t = 320t = 280t = 240t = 200t = 160t = 120t = 80t = 40t = 0 ms . . . . . . . .

i4i3i2i1

FIGURE 27.5 Temporal.redundancy.example.

Industrial Multimedia 27-7

which.works.only.with.QCIF.and.CIF.resolutions.(see.Table.27.1)..It.works.with.blocks.of.8.×.8.pixels.

and.in.the.colo

.spac

.YUV.with.subs

ampling

.4:1:

. e.intr

aframe

.codi

.(den

ominated

.I.fram

using.MPEG.nome

nclature)

.is.base

.on.JPEG.tech

nology,

.and.the.inte

rframe

.codi

.(den

ominated

P,.from.predi

ctive)

.is.based.on.moti

.esti

mation

.with.resp

ect

.to.prev

ious

.fram

es,

.with.a.searc

.rang

of±

.pixe

ls,

.and.codi

.the.die

rences

.betw

een

.macr

oblocks

.when.these.are.high

.than.a.thre

shold

with.DCT.. e.stan

dard

.H.26

.is.an.exte

nsion

.of.H.26

.reac

hing

.high

.bit.rate

.for.the.same.qual

ity.

 e.mor

.rel

evant

.die

rences

.are.that.the.moti

.vecto

.may.be.based.also.on.futu

.fram

.(if.the.fram

is.coded.usin

.past.and.futu

.fram

es,

.it.is.deno

minated

.B,.from.bidi

rectional),

.an.unli

mited

.searc

.spac

for.moti

.vecto

.and.the.poss

ibility

.of.work.with.more.resol

utions

.such.as.4CIF.and.16CI

.othe

.most.impo

rtant

.fami

.of.comp

ressors

.are.thos

.gene

rated

.by.MPEG.(Mot

ion

.Pict

ure

Expe

.Grou

p),

.from.ISO/

IEC.

. e.rst.of.thes

.MPEG

-1

.(ISO.1117

.1991

)

.is.very.simil

.to.H.26

.and.

was.deve

loped

.to.send.audi

o/video

.at.abou

.1.5.Mbps

.using.a.subs

ampled

.4:2:

.and.the.high

.reso

lu-

tion

.of.786.×.576.pixe

ls.

.Anot

her

.die

rence

.is.that.it.supp

orts

.VCR-

.oper

ations

.(fas

.forw

ard,

.rewi

nd,

dire

.rand

.acce

ss,

.etc.

.MPEG

-2

.(ISO.1381

.1993

)

.was.a.stan

dard

.deve

loped

.for.the.broa

dcasting

indu

stry,

.with.a.bit.rate.betw

een

.4.and.6.Mbps

.MPEG

-4

.(ISO.1449

.1999

)

.was.deve

loped

.orig

inally

for.tele

conferencing

.appl

ications

.but.soon.it.show

.adeq

uate

.prop

erties

.to.be.used.in.dist

ributed

.mul-

imedia

.appl

ications.

.Virt

ual

.scen

arios

.can.be.comp

osed

.of.die

rent

.objec

ts,

.whic

.can.be.mana

ged

by.the.user.inde

pendently

.of.othe

rs.

.ITU.intr

oduced

.some.inte

resting

.impr

ovements

.in.the.stan

dards

that.foll

owed

.H.26

.pres

ented

.SNR.scal

ability

.(spa

tial

.and.temp

oral)

.and.H.26

.pres

ents

.some.

improvements.in.information.coding.. is.work.coincided.with.the.developments.made.by.MPEG,.

deve

lopments

.carr

ied

.out.in.unis

.with.deve

lopments

.on.the.stan

dard

.H.26

.(JVT

.Join

.Vide

.Team

bett

.know

.as.MPEG

-4

.AVC.(adv

anced

.vide

.codi

ng)

.(MPE

.part.10).. e.targ

.of.new.comp

res-

sors

.was.to.get.a.bit.rate.redu

ction

.of.50%.with.resp

ect

.to.prev

ious

.stan

dards

.and.achi

eve

.a.more.reli

able

code

.in.the.pre

sence

.of.err

ors.

27.2.3 Quality Evaluation

 e.measurement.of.distortion.produced.by.the.compression/decompression.process.is.still.open.to.

deba

te,

.and.has.been.appr

oached

.from.vari

ous

.angl

es.

.Firs

.the.dest

ination

.of.the.info

rmation

.is.usu-

lly

.the.huma

.eye,.so.the.degr

adation

.meas

urement

.must.take.this.into.acco

unt.

. is.has.prod

uce

mech

anisms

.to.get.subje

ctive

.meas

urement,

.bein

.the.mean.opin

ion

.scor

.(MOS

)

.from.expe

riments

with.huma

.subje

cts

.the.most.comm

on.

.Howe

ver,

.this.eval

uation

.is.time

-consuming

.and.expe

nsive,

.so.

die

rent

.objec

tive

.meth

odologies

.desig

ned

.to.get.quan

titative

.meas

urement

.has.been.deve

loped.

. ese.

can.be.clas

sied

.as.full

-reference

.and.no-r

eference,

.depe

nding

.on.the.avai

lability

.of.an.orig

inal

.imag

alth

ough

.her

.we.are.goi

.to.tal

.abo

.ful

l-reference

.met

hods,

.as.it.is.the.mos

.com

mon

.app

roach.

.meas

ure

.the.erro

.betw

een

.two.signa

ls,

.the.simpl

est

.metr

.is.the.mean.squa

.erro

.(MSE

)

.whic

in.mult

imedia

.is.calc

ulated

.by.aver

aging

.the.squa

red

.inte

nsity

.die

rences

.of.disto

rted

.and.refe

rence

imag

.pixe

ls,

.and.the.most.used.is.the.peak.signa

l-to-noise

.rati

.(PSN

R),

.meas

ured

.in.deci

bels.

.Bein

MAX.the.max

imum

.pos

sible

.pix

.val

.of.the.ima

ge,

.it.can.be.cal

culated

.as

PSNR

MAX

MSE













log

(27.

 ese

.metr

ics

.are.quit

.comm

.beca

use

.they.are.simpl

.to.calc

ulate

.and.have.clea

.physi

cal

.mean

ings,

but.they.are.not.well

-matched

.to.the.perc

eived

.visu

.qual

ity.

. is.fact.has.prom

oted

.the.deve

lopment

of.new.metr

ics,

.such.as.the.stru

ctural-similarity

.(SSI

.base

.[WBS

S04]

.for.a.deep

.expl

anation

.on.

the.form

ulas

.that.enab

.thei

.calc

ulation,

.whic

.is.not.base

.on.a.meas

urement

.of.the.erro

.since.it.is.

base

.on.imag

.form

ation

.prop

erties:

.lumi

nance,

.cont

rast,

.and.stru

ctural

.info

rmation.

.Othe

.exam

ples

of.new.metr

ics

.are.UQI,.VQM,.PEVQ

.and.CZD,.anal

yzed

.by.the.Vide

.Qual

ity

.Expe

.Grou

.(VQE

G),

some.of.whi

.wer

.sta

ndardized

.as.ITU

-T

.Rec

.J..24.g.and.J..247.in.200

.[VR

Z07].