第三届“认证杯”数学中国
数学建模国际赛
承诺书
我们仔细阅读了第三届“认证杯”数学中国数学建模国际赛的竞赛规则。
我们完全明白,在竞赛开始后参赛队员不能以任何方式(包括电话、电子邮
件、网上咨询等)与队外的任何人(包括指导教师)研究、讨论与赛题有关的问
题。
我们知道,抄袭别人的成果是违反竞赛规则的,如果引用别人的成果或其他
公开的资料(包括网上查到的资料),必须按照规定的参考文献的表述方式在正
文引用处和参考文献中明确列出。
我们郑重承诺,严格遵守竞赛规则,以保证竞赛的公正、公平性。如有违反
竞赛规则的行为,我们将受到严肃处理。
我们允许数学中国网站(www.madio.net)公布论文,以供网友之间学习交流,
数学中国网站以非商业目的的论文交流不需要提前取得我们的同意。
我们的参赛队号为:1396
我们选择的题目是:Camouflagepatternandcolorchangeassessment
参赛队员(签名):
队员1:王军
队员2:侯红霞
队员3:冯光娟
参赛队教练员(签名):王鸿杰
Team#1396Page2of29
第三届“认证杯”数学中国
数学建模国际赛
编号专用页
参赛队伍的参赛队号:1396队
竞赛统一编号(由竞赛组委会送至评委团前编号):
竞赛评阅编号(由竞赛评委团评阅前进行编号):
Team#1396Page3of29
Camouflagepatternandcolorchangeassessment
Abstract:Duetotherequirementsofthemodernwaronthecamouflagetechnique,
thecamouflagegoesintoanewdesignperiodofR&Dandproduction.Camouflage
technologypromotionisoneoftheimportantindicatorstomeasureacountry''s
militarylevel,butalsotoadapttothetrendanddirectionofdevelopmentofwar.In
recentyears,theworldmilitarycamouflagetechnologicaldevelopmentquicklyin
othercountries.Althoughtherearedifferences,becausethetechnologyisnotperfect,
leadingtocamouflagetheexposedareaistoolarge,makesitimpossibletoplaythe
roleofeyedeceiveeachotherinaspecificenvironment.Basedontheanalysis,this
papergivesasetofoptimalcamouflage,camouflagecolorvalueandsimulationvalue
schemeoftexture.
Inthispaper,throughthecollectionofvarious(ocean,desert,forest,grassland,
snowandspecialregion)andthecamouflagepicture(eachoftheregionalpicture10),
werecalculatedusingMATLABoriginalpictureofthetargetpatternarea,thenOtsu
methodanditerativemethod(modela)onthepicturefortwovalueimageprocessing,
makestheimagetargetpatternareaasminimalaspossible,targetpatternfurtherarea
calculation,thusobtainssixkindsofcamouflagepictureofregionalenvironment
undertherespectiveideal,imagepixelvaluecorrespondingtothetargetareaofthe
patterncanbeaideddesignidealcamouflagepatterntexture.Inordertosolvethe
single,evaluationindexoftraditionalcamouflageassessmentmethodresultsin
problemssuchaslackofobjectivity.Thispaperpresentsacomprehensiveevaluation
oftheEuclideandistanceandclusteringtheorymethod(modeltwo).Themethod
comprehensiveselectthecolorvaluesofparametersofsixregionalenvironmentsas
anindicatorevaluationsystemaccordingtotheevaluationofthepurposeandobject
ofthecolorvalues.Itsprimaryvalueisdividedinto[25,58]、[10,102]and[327,350]
byclusteranalysis,subjectivedifferenceclassificationobtainedbyclusteringmethod,
classificationstandardisnotaccurateenoughandfurtherdividedbytheEuclidean
distance.TogettheEuclideandistanceandtheresultofclusteringanalysisandthree
categories,furtherexpandthelowestvisibilityincertainenvironmentscamouflage.
Consideringtheinfluenceofregional,color,temperature,illumination,textureand
otherfactorsonthecamouflageeffect,thispaperusesentropymethod(modelthree)
todeterminetheeffectofeachindexweight.Finally,theapplicationofneural
network,simulationintegratedcomputeronthemainfactor,thefinalparameters
obtainedthreetypesofcamouflagecolorandtexturevaluesandthesimulatedimages,
quantitativeassessmentofthebestonthecamouflageeffect.Finally,cconcerningthe
battlefieldinfuturewarsmayturntothetown,therefore,thispaperdesignsthe
simulationforurbancombatcamouflage.
Keywords:Digitalimageprocessing;Clusteringmethod;Entropy-weightmethod;
Factoranalysis;Emulationtechnique;Neuralnetwork
Team#1396Page4of29
Contents
Camouflagepatternandcolorchangeassessment.........................................................3
Contents.........................................................................................................................4
I.Introduction................................................................................................................1
1.1questionorigin.........................................................................................................1
1.2Theproblembackground.........................................................................................1
1.3predecessorsinthecamotechnologyassessmentandshortcomings.......................1
1.4changegraincolorcamouflageeffectevaluationmethodisexpectedtoachieve
goals...............................................................................................................................2
II.TheDescriptionoftheProblem................................................................................2
2.1Background..............................................................................................................2
2.2Problemsneedtobesolved......................................................................................3
III.Models......................................................................................................................3
3.1Basicmodel..............................................................................................................3
3.1.1Symbolsanddefinitions........................................................................................3
3.1.2Modelhypothesis..................................................................................................3
3.1.3Theestablishmentofmodel..................................................................................4
3.1.4Methodandtheresults..........................................................................................6
3.1.5Theresultsofanalysis...........................................................................................9
3.1.6Theadvantagesandthedisadvantagesofmodel..................................................9
3.2ImprovedModel.......................................................................................................9
3.2.1ExtraSymbols.......................................................................................................9
3.2.2AdditionalAssumptions......................................................................................10
3.2.3TheFoundationofModel...................................................................................10
3.2.4SolutionandResult.............................................................................................11
3.2.5AnalysisoftheResult.........................................................................................15
3.2.6StrengthandWeakness.......................................................................................15
IV.Conclusions............................................................................................................15
V.FutureWork............................................................................................................16
5.1pairsofelementsapplycamouflagefactoranalysis..............................................16
5.2Predictionandsimulationofneuralnetworkmodel..............................................18
5.3Urbancamouflageoutlook.....................................................................................20
VI.References..............................................................................................................22
Team#1396Page5of29
VII.Appendix..............................................................................................................23
I.Introduction
InordertoindicatetheoriginofColorcamouflagepatternchangeimpact
assessmentproblems,thefollowingbackgroundisworthmentioning.
1.1Questionorigin
Camouflagereferstotakingvariousmeasurestodeceiveorconfuse
enemies.CamouflageoriginatedinScotland,localhuntersputtheclothandropein
theirbodywhentheyhunting,whichcanfooledpreyandgethigherchanceofsuccess,
whenthefirstworldwarsomearmieshasbeenfollowthiswaytoconfusethe
enemy,inthesecondworldwar,Germansbegantowidespreadthisuseand
simplifiedittobecomeinthemodernsenseofcamouflageinthewar,thecamouflage
effectisverysimple:hideyourselfandyourequipment,soastoavoidtobefoundby
theenemy.Inmodernwar,camouflageandstealthtechnologyasahighlyskilledanti
reconnaissancemeanshasbecomeanimportantpartofthebattlefield.Camouflage
usepaint,dyesandothermaterialstochangetargets,screenandbackgroundcolorand
pattern,inordertoeliminatethelusterofthetarget,reducethesignificanceoftargets
andchangegold’sshape.
1.2Theproblembackground
Camouflagewithitsresultsfast,lowcost,convenientconstructionandother
advantagesiswidelyusedinmilitarytargets,disguiseasaweaponagainstmilitary
reconnaissanceandattacksystemcommonlyusedmethod,itsadvantagesand
disadvantagesofcamouflageeffectisrelatedtothebattleeffectivenesscanbe
effectivelypreserved.Bothinpeacetimeandwartime,camouflagetoprotectand
improvethebattleeffectivenessandwarreadinessmaterialfieldsurvivalabilityhave
animportantrole.Themodernwar,withthecontinuousdevelopmentof
reconnaissancesurveillancetechnologyandthewideuseofprecisionguidedweapons,
makesthebattlefieldtargetfacesagrowingthreattosurvival.Disguisedasakindof
importantsafeguards,facingtheaspectssuchasreconnaissance,surveillanceand
targetacquisitionisbecomingmoreandmoreserious.Developnewcamouflage
technology,materials,equipment,measuresandtacticstousemethodssuchastasks
becomeveryurgent.Andallthesenewcamouflagemethodsneedtobetestto
evaluatetheeffect,toseeiftheyachievecamouflageisvalid.
1.3Predecessorsinthecamotechnologyassessmentandshortcomings
Traditionalcamouflageevaluationmethodusingthefieldobservationmethod,this
methodisaffectedbysubjectiveobservationpersonnelisverybig,theresultsofthe
assessmentisnotscientific,andwastealotofmanpower,materialresourcesandtime.
Tosolvetheseproblemsofmanyresearchersintherelatedresearch:WangDong,
etc.Throughtheestablishmentofgrayhistogramanalysismodeltoevaluate
camouflageeffect:Mulleretc.Basedonstructureoftemplatematchingmethodto
evaluatecamouflageeffect;TheSchemeetc.Throughthecomparisonoftargetand
Team#1396Page2of29
backgroundsignificantlytoevaluatecamouflageeffect.Thesemethodscommon
problemisthesingleevaluationindex,soitisonlyeffectivefortheevaluationof
specificbackground,andtheresultsofmostofthebackgroundisnotcorrect.Andon
thespecificevaluationmethods,suchasZhangDeyangusingfuzzyevaluation
methodtoevaluatecamouflageeffectiveness,butthismethoddoesnotgetridofthe
expertsubjectiveuncertainty;IngoldandsoonbasedonBPneuralnetworkmethod
toestablishaevaluationmodelofdisguise,thismethodnotonlyneedsalargenumber
ofsampledata,butalsoneedtosampledatawithhighaccuracy,theerrorrateis
greatlyimproved,soseriousimpactassessmentofeffectiveness;Blotchanddynamic
camouflageevaluationbasedonexperience,butthismethodisvalidonlytothe
specificregionenvironmentbackground,lackofuniversality.
Accordingtopreviousresearchinthesingleevaluationindex,evaluation
subjectivityisstrong,andtheproblemoflowefficiency,SunHongzhuputforward
theanalytichierarchyprocess(ahp)andfuzzycomprehensiveevaluationmethodof
camouflageeffectevaluationbutthetwokindsofevaluationmethodsrequireexperts
toparticipateintheevaluation,thedecisionresultscontaincertainsubjectivity,cutthe
lackoftheabilityofautonomouslearning,itishardtogetridoftheexpertsubjective
uncertaintyindecision-making.SothisarticleUSESthemethodofiterationmethod,
thevariancebetweentheclusteringanalysismethodtochoosemorefactorssuchas
factoranalysisandneuralnetworkvariablesandquantitativeanalysistoassessthe
reasonablecamo.
1.4Changegraincolorcamouflageeffectevaluationmethodisexpectedto
achievegoals
Quantitativecamoevaluationmethodmodelmakesthemodeltoachievethebest
effectofoptimization,designaappliestobothforest,desert,mountains,ocean,such
asthemostidealurbancamouflagepattern,andthispatternisappliedtotheactual
camouflageeffectevaluation.
II.TheDescriptionoftheProblem
2.1Background
Camouflageisanimportantmeanstoimprovethesoldiersandequipment
battlefieldsurvivabilityinmodernwar.Buthowtoevaluatethequalityofa
camouflagepatternasthebackground,howtochoosethemostappropriate
camouflagepattern,atpresentthereisnouniformstandard.Inordertosolvethis
problem,inthispaper,throughtheestablishmentofmathematicalmodel,thekey
factorsindifferentgeographicalenvironmentcamouflagesimulation,effectsonthe
regionalenvironment,camouflagecolorcamouflagevaluesandripplesizedatais
simulatedunderthespecificenvironmentcamouflagetheidealeffectofcamouflage
patternandwaveparametersthroughthekeyfactor.
Team#1396Page3of29
2.2Problemsneedtobesolved
Inordertodesignaspecificgeographicalenvironmentinthemostideal
camouflagescheme,thedifferentgeographicalenvironment;thedifferentinfluencing
factors;researchpatterncolordifferentvaluesandchangewiththeenvironmentofthe
camouflagepatterncorrugatedthesefourquestions:
(1)Inconjunctionwiththeworldregionalenvironment,thispaperwillbedivided
intogeographicalenvironmentofmarine,forest,desert,grasslands,snowandsome
specialgeographicalenvironment(suchastheDanxialandform,Karst
landform),becauseoffuturewarbrokeoutinthecity,urbancamouflageisalso
discussedinthispaper.
(2)Takingintoaccountthedifferentgeographicalenvironmentcolordifference,
accordingtotheprinciplesofmatchingreferencesbetweenaestheticscamouflage
pigmentinmutualmatchingandcolormatchingprinciple,andapplicationofcolor
aestheticvalueinthequalitativeintoquantitativecolorvaluetomeasurethevalueof
camouflage.
(3)Animportantpartofcamouflage,camouflagepatterndesignofcorrugatedis,
camouflagecorrugatedisbasedonlinedistributionindifferentgeographical
environmentaredepicted,calculation,analysisandresearchinthispaper,specific
regionsoftheripplesizeandcorrugatedopticalindex.
(4)Camouflageeffectdecidedmanycolor,andtemperature,illuminationandother
factorshaveacertainrelationship,throughtheeffectofeachfactorisendowedwith
weight,makingtheexternalfactorquantification,whichcanmakeinaparticular
environmentcamouflagetotendtotheideal.
III.Models
3.1Basicmodel
3.1.1Symbolsanddefinitions
SymbolsExplanation&definitions
1QArea1arearatio
2QRegionalarearatioof2
1utheaveragegraylevelarea
2??
Describetheregionaldifferencesofvarianceofparameters
effectively
(0)DDistancematrix
Team#1396Page4of29
SymbolsExplanation&definitions
SAcollectionofimagegrayscale
NAllpixelsoftheimage
pAfterstandardizationofimagepixels
uWholeimagegrayscaleaverage
BaOceanblue
GagGrassgreen
GBGrassgreen
YakKhaki
WsSnowywhite
GImagegrayseries
CMYKMixSpecial
3.1.2Modelhypothesis
(1)Assumption:theaccuracyofthecollecteddataoftrueandreliableinformation,
colorcanbeobtainedcorrespondingregionalenvironment,camouflagepicture
throughthedatavaluesandrelatedfactorsanalysisvalue.
(2)Exclusivehypothesis:hypothesiscamouflageeffectonlyforconfusedeyesset.
(3)Therationalityassumption:inthesamegeographicalenvironment,camouflage
clothing,theprincipleofarmedequipmentandvehiclesofthesame.
(4)Thefeasibilityofhypothesis:camouflagematerialdesignsimulationhypothesis
isavailable.
3.1.3Theestablishmentofmodel
Forthedesignofmatchingbattleterrainmilitaryclothing,firstofallshouldfirst
understandwhatterrain,belowisthetablelookupfinishingoftheearth''ssurfaceand
variousterrainoftheearth''stotalarea(Table3.1).
Table3.1theearthsurfaceareaandthetotalareaofallkindsoflandform
Theprojectarea/millionhectares
Thesurfaceoftheearthatotalareaabout451.58
Thetotallandareaabout187.50
Thetotalareaoftheoceanabout264.08
Thetotalareaofgrasslandabout45
Thetotalforestareaabout59.44
Thetotalareaofthedesertabout31.40
Thesnowcoveredatotalareaabout22
Thetotalareaspeciallandformabout29.66
Team#1396Page5of29
Figure3.1colors,finenessandtheprincipleofcomplementarycolormap
Themainfunctionofcamouflageistoconfusethepeople’stheeye,theso-called
confusethepeople’stheeyeisthetargetweseeisascloseasbackgroundgrayvalue
and,sowemustfirstunderstandthedesignwiththecolor,Figure3.1isthe
schematicsoftoning,finenessandcomplementarycolor,table3.2istherelational
tablesbetweentoningandcomplementary.
Table3.2colorandcomplementaryrelationshiptable
ComplementaryColorcouplers
Redandbluepurpleyellow
Blueandyellowgreenred
Yellowandredorangeblue
Purpleandgreengreyorange
Greenandorangeolivepurple
Orangeandpurplebrowngreen
Themethodofclusteranalysisofcolorpalette,finenessandcomplementary
principleofmilitarycamouflagepatternbasedonclothing;thatmakessixdifferent
environmentsofthecamouflagepatternclusteringintothreetypesofcamouflage
pattern,therebygreatlyincreasingthecombatfatiguesgeographictypesofawide
rangeofrequirements,moretomeetfuturedemandfactorssuchasstation.
Collectallkindsof(ocean,desert,forest,grassland,snowandspecialregion)
camouflagepicturelandforms(eachoftheregionalpicture10),werecalculatedusing
MATLABoriginalpictureofthetargetpatternarea,andthenthemaximumbetween
classvariancemethodandtheiterativemethodoftwovalueimageprocessingtothe
picture,makingtheimagetargetpatternareatheminimumpossible,targetpattern
furtherareacalculation,thusobtainssixkindsofcamouflagepictureofregional
environmentundertherespectiveideal.
Team#1396Page6of29
3.1.4Methodandtheresults
ThecalculationoftheoriginalpicturesrespectivelybyMATLABtargetpattern
areaandusetheOtsumethodanditerativemethodprocessingpictureofthetarget
patternarea,getthetargetcamouflageareadatatable3.3,fromtable3.3canbeseen
clearlybyusingthemaximumbetweenclassvariancemethodtheobtainedtargetmap
arealargerthantheoriginaltargetarea
Table3.3theoriginalpictureafterpicturepatternandtargetarea
ConditionsOriginalAfterprocessing
Marine54853.5252412
Forest70188.5225070.2
Desert108544.8209705.9
Snow124479.7241926.5
Grassland233696.4227397.6
City38383182816
Speciallandform106689.3223221.4
Thefollowingconcretecalculationmethod:
SettheimagegraylevelissettoS=X(X=1,2,3......).Thenumberofpixels
grayscale,settothenumberofthepixel,imagefor:
12...LiiSNnnnn???????(3.1)
Thestandardization,pixelnumber:
/iPnN?(3.2)
Amongthem,,0,1
iiiSiSpp?????
。
Withacertainimagegrayhistogram,thereistheseparationoftworegional
thresholds.Bythehistogramstatisticstheaveragegraylevelisaftertheseparationof
theregion1,regionaccountedfor2ofthewholeimagearearatioandthewhole
image,region1,region2isshownasfollows.
Thearearatioof1:
10
tj
j
nn?
???
Thearearatioof2:
1
21
Gj
jt
nn??
????
Team#1396Page7of29
Theaveragegraylevelimage:
1
0()
Gj
jjnufn
?
????
Theaveragegrayregion1:
101()
tj
jjnufn?????
Theaveragegrayregion2:1
201()
Gj
jjnufn?
?
????
Thewholeimagemeansgrayandarea1,area2Relationshipbetweentheaverage
grayvaluesas:
1122uuu????(3.3)
Thesameregionoftenhasthesimilarpropertiesofgray,andbetweendifferent
regionsisshownasthegraydifferenceisobvious,whenthedifferenceislargergray
levelbetweenthetworegionsarethreshold,theaveragegrayoftworegions,andthe
overallimageoftheaveragegraydifference,variancebetweenregionsistodescribe
theeffectiveparametersofthisdifference.Itsexpressionis:
2221122()(())Buuutu???????(3.4)
Intheformula,2B?saidtheimagesarevariancebetweenthetwothreshold
segmentationofthe.Obviously,differentvalues,willbetheregionalvarianceof
different,thatistosay,theregionvarianceandmeanarea1,area2,arearatiois
averagethresholdfunction,sotheformulacanbewrittenas:
2221122()()()(())Btuututu???????(3.5)
Thevariancebetweentheregionalmathematicalderivationscanbeexpressedas:
221212()()(()())Bttutut??????(3.6)
Thevarianceoftworegionshavebeendividedbetweenthemaximumvalues,isconsideredthe
bestseparationstateintworegions,therebydeterminingthreshold.
2max()BTt??????(3.7)
Thetargetareamapdatamapdataandapplicationoftheoriginaltargetareaisthe
largestclasstocamouflagethevariancebetweenthetreatedandcomparisonof
correlation,Figure3.2,ascanbeseenfromFigure3.2beforeprocessingandtarget
areadatagraphprocessingafterthecross-correlationofthegood,thisshowsthat
maximumbetweenclassvariancemethodisbettersimulationofenvironmental
changeonthepictureandtheeffectofcamouflageinvariousregion,greatlyreduces
thecamouflagetargetexposureareawhichlargelyconfusetheotherpeople''seyes,
hasreachedexcellentresultsonthebattlefield.
Team#1396Page8of29
Figure3.2Comparisonofcorrelation
Iterativelyforapicture,gettheaveragegray0Toftheinitialthreshold,then
dividetheimageintotwopartsaccordingtothepixelsofimage0T?,calculate
respectivelytheaveragegrayoftwoparts,theportionlessthan0TisAT,theportion
largethan0TisBT,thenaveraging1TsumofATandBT,asanewglobal
thresholdinstead,repeattheaboveprocess,soiterateuntilconvergence.Soiteratively
untiltheswitchfunctionisnotchanged,thistimetheresultwegetisthefinal
segmentationofforegroundandbackground.
Forsomeparticularimage,smallchangesinthedatabutitwillcauseahuge
changeinsegmentation,andbothofthedataisonlyaslightchange,buttheresultwas
greatcontrastsegmentation.ImagesobtainedbyiteratingMATLABbinarythreshold
inFigure3.3.
TheoriginalimageIterativeThresholdbinarizedimage
Figure3.3Iterativethresholdbinarizedimage
Team#1396Page9of29
3.1.5Theresultsofanalysis
Thetargetareamapdatamapdataandapplicationoftheoriginaltargetareaisthe
largestclasstocamouflagethevariancebetweentheprocessed,obtainedbettercross
correlationcomparison,whenthecross-correlationisgood,themaximumbetween
classvariancemethodsimulationenvironmentchangeonthepictureandtheeffectof
camouflageinvariousregion,greatlyreducesthecamouflagetargetexposurearea
whichlargelyconfusetheotherperson''seyes.
Withtheiterativemethodofsegmentationfoundoncertainimage,greatchanges
willcausemicrodatasegmentationresults,bothofthedataisonlyalittlechange,but
thesegmentationeffectisagreatcontrast.Thetargetimagepixelvaluecorresponding
totheareaofthepatterncanbeaideddesignidealcamouflagepatterntexture.
3.1.6Theadvantagesandthedisadvantagesofmodel
Advantages:
Themaximumvariancethresholdisdeterminedwithoutsettingtheother
parameters.Itcanautomaticallyselectthethresholdandisnotonlysuitableforthe
singlethresholdselectionofthetworegionscanalsobeextendedtomulti-level
thresholdselectioninmultiregion.Backgroundandobjectivesbetweenclassvariance
isbigger,parttwodifferencesconstitutethegreaterpartoftheimage,whenthetarget
isdividedintothebackgroundorthewrongpartofthebackgrounderrorisdivided
intothetargetwillleadtothetwopartdifferencebecomessmall.
Therefore,thebetweenclassvariancemaximumseparationmeanstheminimum
misclassificationprobability.Iterativemethodissimple,smallamountofcalculation.
Disadvantages:
Iterativemethodisalocalalgorithm,itispossibletoappearconvergence,andit
can’tbesolved.
3.2Improvedmodel
3.2.1Extrasymbolsdefinitions
Team#1396Page10of29
SymbolsExplanation&definitions
ijd
Euclideandistance
(0)D
Distancematrix
KForthecorrugatedbrightnesscontrast
ACorrugatedvisiblesize
DViewingdistance
3.2.2AdditionalAssumptions
Thismodelassumesthatthesamebasicmodelofthemodelassumptions.
3.2.3TheFoundationofModel
Theshortestdistanceanalysisconductedthatlookatcomparativeanalysisbetween
thevariousoperationalbasecolorvaluesbasedonthesizeoftheclusterenvironment.
Whenclusteranalysis,eachofthecombatzoneasacolorsampletopography,with
i=1,2,3,4,5,6,respectively,marine(navyblue),grassland(grassgreen),forest(ink
green),desert(khaki),snow(snowwhite),speciallandforms(multicolor),theworld''s
landscapeswithj,calculatedseparatelyforeachregionandthegloballandscape
geomorphologyEuclideandistanceijdinMATLAB
Inthechoiceofcolor,shouldmeettheCMYKvalues(CMYKvaluesmeanscyan,
magenta,yellow,blackandfourcolors,itsprimarycolors(primaryhues)ismagenta
(C),cyan(M),yellow(Y),ThereforereferredCMY,asshowninclaim2)ofthethree
primarycolorsasshowninprinciple,becauseeachofthedifferentcolorscanbe
accuratelyrepresentedbyCMYKvalues.Thisarticleprovidesseablue(Ba102),grass
green(Gg327),darkgreen(Gb350),khaki(Yk58),SnowWhite(WS25),special
landformscolor(CMYK10),accordingtotheprinciplesofthethreeprimarycolors,
getsixkindsofcombatzonelandscapecolorvalues(table3.4).
Team#1396Page11of29
Figure3.4Schematictrichromatic
Table3.4colorvalues
ColorEnglishNumberColorvalue
NavyblueBa102
GrassgreenGg327
DarkgreenGb350
KhakiYk58
SnowWhiteWs25
SpecialcolorCMYK10
3.2.4SolutionandResult
1)Theshortestdistanceanalysis
UsingSPSSfordataprocessing,clusteringobtained(Table3.5)withapedigree
chart(Figure3.5).
Table3.5ProximityMatrix
CaseSquaredEuclideanDistance1:Ba2:Gg3:Gb4:Yk5:Ws6:CMYK
1:Ba.00050625.00061504.0001936.0005929.0008464.000
2:Gg50625.000.000529.00072361.00091204.000100489.000
3:Gb61504.000529.000.00085264.000105625.000115600.000
4:Yk1936.00072361.00085264.000.0001089.0002304.000
5:Ws5929.00091204.000105625.0001089.000.000225.000
6:CMYK8464.000100489.000115600.0002304.000225.000.000
Team#1396Page12of29
Figure3.5pedigreechart
Clustertable(Table3.5),togetherwiththeleveloftheclusteringprocessvarious
typesofconsolidationthatoutclearlyfromthepedigreechart(Figure3.5),wecan
clearlyseethewholeprocessofclusteringandrankingthecorrespondingcolor.
2)Euclideandistanceanalysis
Makingeachcolorvaluevariation(Figure3.6)usingMATLAB,FIG1,2,3,4,5,6
representBa,Gg,Gb,Yk,Ws,CMYK.
11.522.533.544.555.56
0
50
100
150
200
250
300
350
Figure3.6Changesinthevalueofeachcolorchart
Euclideandistanceofthesixkindsoflocalcolorintothreecategories
????
??
--25,58
-and-10,102
--327,350
WsandYk
CMYKBa
GgandGb
??
??
?
(3.8)
Eachcolorasasample,andi=1,2,3,4,5,6representsBa,Gg,Gb,Yk,Ws,CMYK
colorrespectivelysix,usingequation(1)betweenthetwogroupscalculated
Euclideandistance.
12
2
1
d(,)()pijninj
ij
xxxx
?
?????
???
(3.9)
Team#1396Page13of29
Suchas:1221221[(102327)]225dd????,
1222332[(327350)]23dd????
Resultingdistancematrix(0)Dasfollows(duetosymmetry,onlywritethe
diagonalandlowertriangularsectionandrowandcolumnpositionsmarked
correspondingcategory):
(0)
0
2250
248230
442692920
77302325330
9231734048150
D
??
??
???
??
??
(3.10)
Thesizeofeachelementin(0)Dvalue,reflectingthedegreeproximitybetween
thesixcolors.Forexample,theEuclideandistancebetweenWsandCMYKminimum
(15),reflectingboththeclosestcolors,camouflagetechniquesalsoveryclosetothe
environment.(0)D,Itcanbeseenthatthesmallestelement65({6},{5})15Dd??.5G
And6Gwillbeatthelevelof15andmergedintoanewclass7{5,6}G?,andthen
usetheshortestdistancerecurrenceformula:
(,)min{|,}rkijrkDGGdiGjG???
min{min{|,},min{|,}}min{(,),(,)}ijpkijqk
pkqk
diGjGdiGjGDGGDGG??????(3.11)
Calculate7GVersus1234GGGG、、、.Obtainthefollowingmatrix:
(1)
{5,6}0
1770
23022250
3352248230
433442692920
D
??
??
?
??
??
(3.12)
Similarlyget(2)(3)(4)DDD、、thedistancematrix.
(2)
10
4440
{2,3}2252690
{56}77333020
D
??
??
?
????,
Team#1396Page14of29
(3)
10
{2,3}2250
{4,5,6}442690
D
????
?
??
(3.13)
(4){1,4,5,6}0{2,3}2250D???????
AccordingtothedistancematrixgeneratedpedigreechartshowninFigure3.7in
matlab.
Figure3.7Distancematrixgenerationpedigreechart
Insummary,weareinthehorizontaldistanceofthefirst15pooledsamples5and6
toobtainthenewclass7G.So,finallytheformationofacategory,theresultsismore
suitableenvironmentundertheguiseofordinaryandspecialenvironment.
3)Calculatetheripplevalue
Principleofthecorrugationshapeandconfigurationare:
(1)Shapeofcorrugationdeformationcamouflageconsistsofirregularcurvilinear
profile.
(2)Thesamecolorcorrugatedshouldadoptdifferentshapes,sizesripple.
(3)Intermediatecolorsandcontrastsrippleripplesintheequipmentshouldbe
staggeredconfiguration.
(4)Modificationcamouflagebellowsshouldnotsymmetricconfiguration.
(5)Corrugatedcontouredgeequipmentshouldnotinterrupt,shouldbeextendedto
theothersurfacetogolongwhenextendedridgewithcorrugatedequipmentshould
intersectatanacuteangle.
(6)Atthetopoftheequipmentshouldbemoredarkcorrugatedconfiguration,the
darksideoftheconfigurationshouldbemorebrightripplesandrippleswillextend
dark.
(7)Portsiteequipmentshouldbeconfigureddarkripples,butnotduplicatethe
outlineemptysite.
(8)Equippedwithprojectionspreferablydarkbellowsconfiguration,therecess
shouldbebrightcorrugatedconfiguration,andthecentercorrugationshouldcoincide
withtheprojectingorrecessedportionwithavertex.
Inordertopreventthegenerationofspaceblendingvariouscolorsripple
Team#1396Page15of29
phenomenon,camouflagerippledimensionsmustbeseeninapredeterminedviewing
distance.Accordingcorrugateddifferentbrightnesscontrast,GJB453givenripple
visiblesizeformula:
Atthetime0.4K?,
(2.5~3)3400DA?(3.14)
Atthetime0.20.4K??,
(3~4)3400DA?(3.15)
Where:Kcorrugatedbrightnessandcontrast;Acorrugatedvisiblesize;Dforthe
viewingdistance.
Theyear2000GJB4004provisions,
0.0009AD?(3.16)
Equation(3)shouldbeunderthepremiseof(1)takesalargervalueaccordingto
theformula.Grounddeformationcamouflageequipmentdesignedviewingdistanceof
800~3000m,correspondingcamouflageripplesizeshouldbebetween0.72~2.70m.
3.2.5AnalysisoftheResult
BecauseclusteranalysisisSPSSsoftware(CMYKandBa)intoonegroup,rather
thanthe(CMYKandWs)clusteredtogether,soclusteranalysismodelestablisheda
systemerroroccurs.Themodelisnotthebestmodel,itisalsoappliedthismodelto
improvetheEuclideandistance,allowinghigherreliabilitymodel,bettercamouflage
effect.
3.2.6StrengthandWeakness
Benefits:
Clusteranalysisclearlydescribethesizeofthecolorvaluesofeachoperational
environment,ameasureofthesimilaritybetweendifferentdatasources;higher
dimensionalEuclideandistanceanalysisunaffected,sothatthecredibilityofthe
model,camouflageeffectbetter.
Cons:
Dimensionalobservationofthemodelisnotexactlythesame,sotheexistenceof
objectiverealityduringtheclusteringfeaturecertainaspectsoftheroleofdifferences
havebeenexaggeratedorreducedpossible.
IV.Conclusions
Camouflageassesstheeffectofgreatmilitaryvalue,thebasemodelandthemodel
proposedtoenhancethetexture,colorassessmentmethods,andgivesitappliescamo
uflageeffectevaluation;indexbyassigningweightscaneffectivelydeterminethediff
erentindicatorsinthesamecontextforthecamouflageeffectofcontribution;textentr
opymethodusedtoobjectivelydeterminetheweightindicators.Insummary,thebasi
cmodelcamouflagemulti-indexmodelsandenhancethecomprehensiveevaluationm
ethodtoeffectivelyandobjectivelyevaluatethefullcamouflagepatternandcolorwit
Team#1396Page16of29
hrespecttothemeritsofthedifferentlevelsofbackgroundlandscapes,resultinginth
ebestenvironmentcamouflagepattern.
V.FutureWork
5.1Pairsofelementsapplycamouflagefactoranalysis
Inthepresentwar,inordertoachievethedesiredeffectofaparticularenvironment
camouflage,camouflageisthebestchoice.Theaimistochoosethefactoranalysis
factorwithafewmorevariablesdescribingtheeffectoftheimpactonthecamouflage
case,hasreachedthebestcamouflagequantitativeresults.
Applicationoffactoranalysistoclusteranalysisresultinginthreecamouflage
patternsaccordingtotheirdegreeoftoningandsurroundingspixelfitting
comprehensiveselectionofregional,pigment,temperature,light,corrugatedthese
fivefactorsastheevaluationindex,whichisacomprehensiveasawaytoadapttothe
realitiesofwarneededtocamouflagepatterns;
Saaty,whoproposedtheuseofa1-9scale--aijvalues1,2,...,9anditsreciprocal
number1,1/2,...,1/9qualitativetoquantitativeconversionissues:
Table5.1Qualitativeandquantitativefactorsproblemintosizetable
Size123456789
Ci:Cj
importanceSame—
Slightly
stronger—Strong—
Significantly
stronger—
Definitely
stronger
Assumingthatregion,camouflage,temperature,light,rippleswereD1、D2、D3、
D4、D5Therelationshipbetweentherelativeimportanceofeachfactorbetweentwo
relativelyquantitativeresultsareasfollows:
Table5.2Factorsaffectingtherelationshipbetweentables
QualitativeresultsQuantitativeresults
DiVersusDjaffectIdenticalDi:Dj=1:1
DiVersusDjaffectSlightlystrongerDi:Dj=3:1
DiVersusDjaffectStrongDi:Dj=5:1
DiVersusDjaffectSignificantlystrongerDi:Dj=7:1
DiVersusDjaffectDefinitelystrongerDi:Dj=9:1
DiVersusDjaffectBetweensaidtwolevelsDi:Dj=2,4,6,8:1
DiVersusDjaffectandContrarytotheabovecaseDi:Dj=1:1,2,…,9
Accordingtothecharacteristicsofwarcombinedwiththerelativerelationship
betweensubjectiveawarenessofthefactorsconsideredareasfollows:
Team#1396Page17of29
1112131415
2122232425
3132333435
4142434445
5152535455
/////
/////
/////
/////
/////
DDDDDDDDDD
DDDDDDDDDD
DDDDDDDDDDD
DDDDDDDDDD
DDDDDDDDDD
?
(5.1)
Theuseoftheelementsontheotherfourelementsastheratiooftheelementinthe
camouflageeffectofinfluencingfactors,Obtainthefollowingtableforeach
weightingfactor
Table5.3Factorweightingcoefficienttable
D1D2D3D4D5
1.000.500.201.000.33
2.001.000.500.200.25
5.002.001.000.332.00
1.005.000.331.000.33
3.004.000.503.001.00
Togivethecorrespondingelementcorrespondingweightsworthtogotothenext
relationship:
D1/D1=1、D1/D2=2、D1/D3=5、D1/D4=1、D1/D5=3
D2/D1=1/2、D2/D2=1、D2/D3=2、D2/D4=5、D2/D5=4
D3/D1=1/5、D3/D2=1/2、D3/D3=1、D3/D4=1/3、D3/D5=1/2
D4/D1=1、D4/D2=1/5、D4/D3=3、D4/D4=1、D4/D5=3
D5/D1=1/3、D5/D2=1/4、D5/D3=2、D5/D4=1/3、D5/D5=1
ExtractedbytheabovefactorsaffectthedataobtainedCamouflagefactormatrixas
follows:
1
12513
1/21254
1/51/211/31/2
11/5313
1/31/421/31
D?
(5.2)
UsingSPSSsoftwaretoanalyzeallfactorsaffectingcamouflagefractionalanalysis
gravelsoilmap.
Figure
5.1Factoringravelsoil
Team#1396Page18of29
Ascanbeseenfromthefigure,thefirsttwofactorssteepdiscount(factorand
factorsteep),andbacktoflattenreflectedfromthesideofthesetwofactorsasthe
maincomponentfactor,andthereforetakeintoaccountthecamouflagedesignisthat
weshouldgiveprioritytothesetwofactorswereincamouflageclothing,vehicles,
weaponsandothercamouflage.Followsasacomponentmatrix:
Table5.4BasedonPrincipalComponentAnalysisMatrixFactor
ComponentMatrix
Component
123
D10.9880.0771.07
D2-0.0530.8630.81
D30.983-0.0510.93
D4-0.1510.8690.75
D50.9650.1561.12
Abovetablecanvisuallysee,factorsaffectingthecamouflageeffectofD1factor
(geographical)andD5factor(ripple)decision.Sowecanseemoreofcamouflage
designshouldbecombinedwithgeographicandrippledesign,designpatterns,such
aspostersattachedpicturelater.
5.2Predictionandsimulationofneuralnetworkmodel
Thefollowingtableisanintegratedvalueofeachgeographicalareaandavariety
oflocalspecies;colorvaluesofthevariousregionsofcolors.Eachcorresponding
geographicalcamouflageclothing,vehiclesandweapons,thevariancebetweenthe
originaldatainthetargetareaofthevariousgeographicalpicturesofthelargest
classesandcalculatedbyMATLAB.(Asdefinedherein:ageographicalpictureas
backgroundtothepicturecamouflageclothingandequipmentasthetargetfigure,and
thusthroughthetargetfigureinthebackgrounddeterminestheproportionofthesize
oftheeffectofcamouflagetechniques).
Table5.5Camouflagetechniquesmainingredientlist
ConditionColorArea/onehundredmillionhectaresColorvalueOriginalOtsu
OceanBa264.0810264843.5252412
ForestGg45.0032770188.5225070.2
DesertGb59.04350108544.8209705.9
SnowYk31.458124479.7241926.5
PrairieWs22.0025233696.4227397.6
CityGb1.537038383182816
Special
landformsCMYK29.0310106689.3223221.4
OnthetablecombinedwithneuralnetworkmodeldatausingSPSSsoftwarefor
Team#1396Page19of29
datareliablypredictedbythesimulationfunctionMATLABsimulationreliability
throughaneuralnetworkmodelforlocalcolorvalueandcolorvaluescamouflage
fatiguesandregionalrippleandripplecarryon,eventuallycometoaparticulararea
canbesolvedcombatcamouflagecolorsandpatterns,sothatthecamouflageeffect
achievethebestresultsinaspecificgeographical.
Fromthisfiguremayreflectthearea/millionhectaresaffectcolorvalue,the
originaltargetareaistheidealenvironmenttocamouflagetheproportionofdata
values,itiseasytoseethatthebiggestfactorswere,colorvalueandtheoriginalvalue
ofthetargetarea.
Figure5.2Camouflageeffectonthreefactorscontributedmap
Thefollowingdataontheuseoftablemadeoutofmulti-storeyMATALB
inductionprinciplemadethreeinductionrelationshipdiagramshownbelow:
Figure5.3Manyperceptualmapneuralnetwork
Thefigureisthroughtheestablishmentofneuralnetworkmodel,theoriginaltarget
data,colorvaluedata,andgeographicareadataasafactoroftheinputlayer,and
finallythroughtheintermediatestepfunctionhiddenlayerneuronsandlinear
mappingandlinearsensorusedinoperationalenvironmentmappingrelations
camouflagepatterndatavaluesideallytheoutputlayer,andfinallythroughMATLAB
idealenvironmentforcamouflagedatavaluesofmodelingandsimulationweremade
underdifferentenvironmentalsimulationoutthebestcamouflageclothing,
camouflagedisguisedweaponsdesignandcolorvaluesripplepattern.
Team#1396Page20of29
Simulationresultsareasfollows:
Figure5.4JungleCamouflagesimulationrenderings
CamouflageEffectWebsiteat:
http://www.szj2014com.icoc.cc/pod.jsp?id=23&_php=2_313_1
5.3Urbancamouflageoutlook
ApplicationcamouflagetechnologyismainlyusedinWorldWarII,theKorean
War.TheVietnamWar,whichischaracterizedbythemainbattlefieldmainlyin
jungles,grasslands,deserts,snowandothernaturalenvironments,socamouflage
technologyduringWorldWarIIanimportantmeansofcamouflageshadow
possession.Whichappearedinavarietyofcamouflageclothing,camouflagetank
weaponsandcamouflage,camouflagetechnologyinventionandapplicationwecan
seenowthatthewarisnowavailableforthefirstleap?
Butthefuturebattlefield,forjunglecamouflageclasswilltendtoapplymore
urbanizedcamouflage,whichisrootedinthetransformationofcontemporaryand
futurebattlefield.Duetothebattlefieldbythetransformationofurbanjungle,sothe
citywillbecomeapopularcamouflagetechnologyarea,thecitywillbeurban
camouflagecolortoneconsistent,texturestendtobemoreurbanlinesofhorizontal
andverticallinestexture.
UrbanCamouflagetechnologyisbasedonaprogramundertheguiseofurban
background,urbancamouflagefatiguesthebestresults,bothurbanwearcamouflage
clothingandequipmentisnotbeingputpasspasserby.Thisiswhatweneedforawar
basedonfutureprospectsforthefuturecamouflageeffect,asshownbelow:
Team#1396Page21of29
Figure5.5Simulationrenderingsofurbancamouflage
ApplicationMATLABsoftwaresimulationoftheurbancamouflagefatigues,the
soldiersinthecitytomoveseemingly"Lingboweibu"lurking,reflectsthestealth
technology.
Team#1396Page22of29
VI.References
[1]FordeBergOlsen.MethodsforEvaluatingThermalCamouflage[R].ADA456649,
Decemder.2005.
[2]MIL-C-53004B.CamouflageSystems.Modular.Lightweight,XSynthetic.Woodland
Desertandsow[s].AliphaticspecificationofU.S.A,1989.
[3]ZhangJiangchong.Camotechnology[M].Beijing:ChainTextilePress.March5,2002.
[4]ChenjiveMatlabatreasuryofknowledge[M],Beijing:electronicindustrypress,
September.2013.
[5]WangZirong,YuDabingSunXiao,Coatingandstealthtechnology[J].Journalof
CoatingsinChina,1999,(5):39-42
[6]MATLABTechnologyalliance,ZhangYanMATLABimageprocessingsuper
learningmanual.People’spostsandtelecommunicationspublishing
house,March,2014.
[7]ZhangDefeng,MATLABnumericalcomputationmethod[M],Mechanical
engineeringpress,Jan,2010.
[8]CaoYi,AnalysisontheTechnicalItemsofDistortionPatternpainting,Infrared
Technology,Fed.2008
Team#1396Page23of29
VII.Appendix
7.1Picturepixelareaandthelargestamongareasourcecategoryvariance
method
1=imread(''13.jpg'');
I=im2bw(I1);
L=bwlabel(I);
STATS=regionprops(L,''Area'');
w=[STATS.Area]
I=imread(''3.jpg'')
imshow(I)
inf=imfinfo(''3.jpg'');
X=rgb2gray(I);
imshow(X)
X2=grayslice(I,64);
imshow(X2,hot(64))
X3=im2bw(X);
imshow(X3)
7.2Changesinthevalueofeachcolorchart
x=[1,2,3,4,5,6];
y=[102,327,350,58,25,10];
plot(x,y)
7.3PicturesbyMATLABdigitalimag
processingprogram
g0=imread(''3.jpg'')
g1=imread(''11.jpg'')
g0=g0(:,:,2);
figure(1);imshow(g0)
I=imread(''3.jpg'');
J=imadjust(I,[0.2,0.4],[]);
subplot(221),imshow(I);
subplot(222),imshow(J)
subplot(223),imshow(g0)
subplot(224),imshow(g1)
title(''IterativeThresholdbinarizedimage'')
7.4.Iterativethresholdbinarization
f=imread(''1.jpg'');
subplot(121);
imshow(f);
title(''Theoriginalimage'')
Team#1396Page24of29
f=double(f);
T=(min(f(:))+max(f(:)))/2
done=false;
i=0;
while~done
r1=find(f<=T);
r2=find(f>T);
Tnew=(mean(f(r1))+mean(f(r2)))/2;
done=abs(Tnew-T)<1;
T=Tnew;
i=i+1;
end
f(r1)=0;
f(r2)=1;
subplot(122)
imshow(f)
title(''IterativeThresholdbinarizedimage'')
|
|