EVAUAIO OF

DULTXUEGAIN ATR

'4"f ~ UNTE STATDULEEXUR

___cGRADIENTlPATTERN

App-"robl c. x le s l U 1 7BY THE ASSOCIATE

OFFICE O~ MLITRYLEAESI

6,'''T~~ MIITR ACADEMYttL

CDYER:-I- -~

-~ Front, top to bottom: M113A1, US Army pattern; XMI Dual-Tex pattern;PbV 302, Swedish Army pattern.

Rear, top to bottom: M577A1, USAREUR pattern; KPzLeopard II, Bundeswehrpattern; FV.432, UK pattern.

S~~NOT___E: Further information and materials are available concerning the Dual-Tex

pattern. Requests should be forwarded to:

Office of Military LeadershipUS Military Academy

West Point, New York 10996Telephone points of contact: MAJ T.R. OtNeill or CPT William L. Johnsmeyer,

S~US Military Academy; AV 688-2515/4127.

4'7

-iMAJ Timthy RO'Neil

Offie ofMiliary eadeshi

US Mlitay Acdem

'SECURITY CLASSIFICATION OF THIS i9AGE (Wheun Dat. Ante,.41

REPOR T DOUETTO AEN BEFORE COMPLETING FORMj 2 . J O V A C E S S O N E C IP I EN T 'S C A

T A L O G N U M B E P .

Ev )_ (utl S. TYPE OF REPORT A PERIOD COVERED

EDUAL-TEX: Evaluation of Dual-Texture Technical beprt,

Goncalment PatternficandInblorb. experi-e-ta -aou-a

Mand w Ain:ter(sow /o'nditin. proiaey20SMltr

Affadem cfMltayLadeteeerployp as subjcts Color slid sUMERiSwer usied tto simuliate fiaeldcndton.Th xermntlpatr

wasectoun tob/infcnl oeefcietan thr aterOfic of IntttonlRsac71.NMERO3AE

MIT

DUAL- TEX:

EVALUATION OF

DUAL - TEXTURE GRADIENT PATTERN

BY

THE ASSOCIATES,

OFFICE OF MILITARY LEADERSHIP

UNITED STATES MILITARY ACADEMY

WEST POINT, NEW YORK

MAJ Timothy R. O'Neill, ARCPT William L. Johnsmeyer, IN

CPT James M. Brusitus, ARMAJ Bruce T. Caine, INCPT Mark L. Frey, ENLTC Thomas G. Johnson, INCPT Harry N. Lumpkin, QMCPT Thomas N. Meriwether, AR sor.MAJ Kim H. Olmstead, AR

-

CPT David L. Taylor, AR Bufi Selen tC fluff Sectilo 0-

............... ... ... ... ... ... .........ISTMIMiOR/AAILANUITY CODES

* .----_ _C_'_

Thi. study compared the effectiveness of an experimental camouflage conceal-ment pattern (Dual-Texture Gradient Pattern) with a selection of existing pa.tternsand unpatterned targets in summer and winter (snow) conditions. Approximately 260US Military Academy cadets were employed as subjects. Color slide series were usedto sim-tilate field conditions. The experimental pattern was found to be sigaificant-ly more effective than other pattern measures tested. Field validation was recom-mended.

NOTE: Any conclusions in this report are not to be construed as .

official US Military Academy or Department of the Army Positions !S~unless so designated by other authorized documents.

,DIS, is um is prepar for off l p es on~~~~~lY./Its~~a cte s may nt be r~d~ -q~s:uei

!whl or in p without he s ific rmsion the er-,"i natUS ltrAce i echne

' ACKN(OWLEDGEN ENT SThe authors wish to acknowledge the assistance of Ym. Lawrence

J. Gale, US Army Mobility Equipment -Research and Development Command, andLTC George D. Waters, Director, Office of Institutional Research, Us mili-Utary Academy; and particularly the assistance of numerous cadets who acted

as controllers during the course of the experiment. Finally, special acknow-i ledgement is due to Professor E. R. Long of the Department of Psychology, ~Universityr of Worth Carolina at Chapel Hill, whose guidance at the incep-d";''' fination ofm hdooythis project several years ago contributed significantly to the i

f l h l

' V I I I

TABLE 17 CONTENTS

PARA TITLE PAGE

1 BACKGROUND .. .. ........ ....................... . .......... .. . ...

2 OBJECTIVE. .. ............ ..................... . .................. 1-4

3 SCOPE . .. ....... . ...... ..... . ...... ..... . ....... .... . .......4 SUMMARY OF RESULTS. .. ........ ................................... 1-55 CONIUnSIOiS. . .......... .. ... . ... . ... . ..... 1-56 RECONMEN-DATIONS. 6.............. ................... .......... 1-5

7 CONDUCT OF TESTING. .. ....... . ...................... .............. 1-6

APPENDICES

I PATTERNS

II PIGMENT SAMPLES

III PHOTOGRAPHS

SIV SUBJECT SELECTION DATA

V DATA ENUMERATION

VI REFERENCES

VII DISTRIBUTION

44

.3i

1. BACKGROUND

1.1

Recent trends in tactical doctrine for United States ground forcessuggest the inevitability of "fighting outnumbered" in the initial stages of thenext war - a position of numerical inferiority which onlr increases the need forthe most effective training, doctrine and equipment. Most strategies for fight-ing outnumbered have relied upon increasing defensive effectiveness with the goalof producing extremely unbalanced favorable exchange rfitios - through improvedhardware, sophisticated doetrine and perfection of di.ect-fire gunnery tech-niques. There is, however, a coiilementary passive Approach to the problem ofimproving exchange ratios: perfection of countersu:veillance techniques whichwill reduce the vulnerability of critical weapons srstems in the Active Defenseposture.1.2 ,

The increased emphasis on passive countersurveillance technology has

been impressive in the last 4-5 years, particularly in view of its effective ne- 2glect since the end of World War II. The new interest includes development andtesting of pattern-painting designs and techniques, artificial garnish, reduc-tion of thermal signatures and other efforts. The improvements in reduction ofequipment signature has been validated in field tests (see references 1-3, ap-pendix VI), and the general concepts have gained wide acceptance. The existingmeasures, however, offer anything but a panacea for the problems of battlefieldcountersurveillance.

One striking shortcoming in the present system is inherent in thenature of pattern-painting: that this measure alone is not sufficient to reduce

: target signature to favorable levels. The present US Arnrj pattern has been pro-ven effective in comparison to unpatterned vehicleR, but it is stipulated in allsuch reports that the pattern itself must be enhanced with astute employment ofnatural and artificial garnish (nets, disruptors, natural shrubbery, carefulsiting and other techniques). This assumption is acceptable in most cases, andcertainly for relatively static targets: headquarters complexes, firing bat-1 tery positions, and for most combat systems. However, the developing ActiveDefense posture poses special problems:

" 1.3.1 There is no reduction in the need for effective camouflage for criti-* cal systems - particularly for direct-fire anti-armor systems (MBT, MICV and ITV)

and their early and potentially prolonged committment to close combat argues,I in fact, for extraordinary countersurveillance measures. Such systems are thebasic instruments of favorable exchange rates, and their vulnerability must asa consequence be reduced as lar as possible without degrading the effectiveness oftheir fighting capabilities.

1.3.2 The satisfaction of the requirement for concealment may not be real-izable in the present context because of the explicit requirement for extensivegarnish. These critical systems must be capable of displacing more frequentlyin the Active Defense than has been the case in the past. The use of extensivecamouflage netting and related kits - to say nothing of frequent gathering ofnatural garnish - may prove so time-consuming that tactical displacement wouldbe slowed below levels acceptable in the current doctrine; or, even more likely,the use of such measures would soon be abandoned in combat as a necessary trade-t off to mobility. In dither case, overall system and tactical unit effective-

Im

ness would be degraded in a situation which begs all the enhancement possible.

1.41

ato The need for extensive use of garnish is due largely to practical limi-tations inherent in the idea of pattern-painting. The specific identifying charac-teristics (signatures) of most targets are so overwhelming that simple applicationof paints is, taken alone, relatively ineffectual. The obvious question, then, is:to what extent can the effectiveness of patterns be improved to reduce the needfor bulky garnish which restricts tactical mobility?

1.5 The Problem

1.5.1 Camouflage as a military technique consists of the purposeful degrada-tion of a target signature with the objective of reducing an eneqm's ability todetect the presence of the target and identify its type. The term signature is amilitary term which will be defined for the purposes of this study as: a schemaor organized aggregate of distinctive features unique to a specific stimulus cate-gory." The various schemata which are characteristic of imilitary targets includevisual (w:aite light), visual (infrared), thermal, auditory, radar, and olfactory;this study is concerned principally with visual signature in the visible lightspectrum.

1.5.2 Two variables affect the power of a concealment pattern: observercharacteristics and target/pattern characteristics. It is generally hypothesized(though a lack of empirical demonstration has historically made these presumptionstentative) that observer properties include:

1.5.2.1 Innate perceptual organizing properties (gestalt principles): closure, 4.good figure, continuation, similarity and related individual characteristics.

1.5.2.2 Individual cue-search habits: learned or innate visual search patterns.

1.5.2.3 Prior learning and perceptual set: familiarity with target stimuluscategories and their attehdant schemata. These might include prior experience inthe field with vehicles; familiarity with shadow patterns (see figure 1), shape,color, and related signatures which prime the observer to perceive such expectedstimuli.

._ . .r .

1.5.3 Pattern properties contribute to concealment in a variety of ways, butmost properties can only be evaluated subjectively. Observation of patteron perfor-mance in a variety of situations suggests to the authors that the variables most fre-quently involved are, in approximate order of importance:

1.5.3.1 Value contrast of pattern and ground: if the overall value (relativebrightness) of the target is significantly different from the background, detectionis usually quite easy.

1.5t3.2 Color contrast of psntteri and ground: this may seem intuitively ob-vious, yet most pattern applications in use in US Army units violate this principle.

1.5.3.3 Ifitra-pattern Value differentiation: if sufficient value contrast isnot present in the patterni, the pattern will merge at soie range into a monocolor, andthe disruptive effect will be lost.

1.5.3.4 Texture-gradient contrast with the ground: most patterns match thetexture of the environment only at relatively long ranges; this ignores the problemof observation under optical magnification (such as gunner's sights, rangefinders andother common observation aids). At closer examination, broad patterns will "standout" against the background.

1.5.4 Given these characteristics and constraints, the problem is one of de-

veloping a concealment pattern which offers maximum practical concealment value underall common threats without requiring the use of extensive garnish which restrictstactical mobility.

1.6 The Dual-Texture Gradient Pattern

1.6.1 The problem cited above hs generally been regarded by camoufluers asvirtually insoluble. However, recent examination of alternative pattern schemes sug-gests that there may be measures which will meet the concealment/mobility requirementsof the Active Defense. The Dual-Texture Gradient Pattern developed within the Psy-chology Committee, Office of Military Leadership, US Military AcadenW, is intended tofill this specific need.

1.6.2 The pattern (see figure 2 and appendix. I) is derived from the US Araypattern system developed by the US Army Mobility Equipment Research and Development Com-mand, (MERADCOM), at Fort Belvoir, VA, a measure which has been in widespread usefor several years. At longer ranges and without optical enhancement it is not read-ily distinguishable from the standard measure. However, the broad color patterns ofthe US Pray Pattern have been randomly disrupted to produce a higher texture gradientunder closer observation.

!!1.6.3 At longer ranges, the Dual-Texture Gradient (DTG) pattern merges into_a mroattern of broad light and dark are.s' which matches the texture of the ground

that distance. At closer range (under optical magnification) a micropattern re-solves which provides a continuing match with the background texture. This micropat-tern is conceived as based on a square grid: not because squares are particularlyeffective shapes for concealment, but because pattern design is simplified by devel-oping the micropattern from the US DAlT pattern (as a macropattern) by computer,as a relatively simple linear progranming technique.

1.6.4 The obviously more complex pattern may be criticised for difficulty inapplication at unit/organizational level. However, the pattern is not nearly as pre-cise in practice as the drawings suggest. In fact, the "squares" provide only a guide-line for the drafter and painter and the pattern would be in application much lessformal.

., , _ - __ ~

Figure 2. Dial-Texture Gradient Pattern (xMI)

1.6.5 The DTG pattern is not designed for use without garnish of any kind;in all probablility, this goal is unattainable. However, it seems reasonable to pre-siune that an extremely effective pattern could at least dispense with the bulkieritems, particularly nets and disruptor kits which must be anchored to the ground anddetached and stowed when the vehicle must move. This is the goal of the Dual-TextureGradient Pattern.

1.6.6 Before continuing development of the pattern or moving to broader fieldtests, it appeared necessary to demonstrate a significant advantage in performance for'.he DTG pattern over other patterns in use.

vi 2. OBJECTIVE

The purpose of the research described in this report was the compara-tive evaluation of the Dual-Texture Gradient Pattern against selected existing measures.To be considered for even limited application in the Active Defense, the DTG mustdemonstrate, a significant advantage in terms of range to detection and range to identi-fication over at least the present (MERADCOM) measure and unpatterned targets.

3. SCOPE

3.1

Because of practical physical restraints and because of the need toisolate pattern configuration as the independent variable, the test was conducted an,a laboratory simulation of field environment. This naturally introduced some dis-tortion of the actual processes involved in field detection and identification; how-ever, these distortions were uniform over all patterns compared.

3.2

-1 The test was conducted from August 1976 - February 1977 at Wes; Point,

1 -4_ -

i Bew York, by selected members of the psychology and Leadership Committees, Office oft Military Leadership, United States Military Academy. The research is still in pro-

gress at this time.

3.3

Ddring the experiment,,hptrokimately 260 US Military Academy cadetswere employed as subjects; the-expdrimefit was used to supplemeint instihction in*visual perception and experimental method as part of a Third Class (sophomo)re) 'coursein introductory psychology. Det-ails of subject selection and characteristics axe atAppendix IV.3.4

Funding of the research was at the expense of the officers concerned,

with the exception of 20 gallons of paint provided as a courtesy by the US Army Mo-bility Equipment Resharch aid Development Command.

4. SUMoMA oF REsuLTS

4.1

The Dual-Texture Gradient Pattern, tested under laboratory simulationof field environmeyda subje the experiment wanly more effective than the US Arc npattern and a solid forest green control target in summer temperate ensironment underthe specific conditions tested.

4.2

The DTG Pattern proved significantly more effectine othan the US Arneand an adaptation of the Swedish Aro patternsovduring winter (snow) conditions, bMtproved under the specific conditions tested not significantly more effective than a.solid white control panel.

5. CONCSUSIULS

5.1That the Dual-Texture Gradient Pattern appears to offer the potsnrial

for significant improvcmens over present measures in a variety of environments.

5.2"That the Dual-Texture Gradient Pattern py reqfes field vaiidatiohe under

tactical conditionhs.

6. RECOmmENDATIuS

6.1That field validatien tests of the DTG pattern be undertaken to deter-

mine the actual degree of improvnement over present measures, and that the ivpict ofthe reduction in system vulnerability produced by application of the uD pattern beevaluated, along with fractical measurement of time and cost constraints to determine

the total value of the measure ini terms of systems effectiveness.

4, ....."''....!.....I .. I. .. ! F ! 1 I" ! I

C I I I I 1

7. CONDUCT OF TESTING

7.1 ObjectiveDetermine the relative effectiveness of the Dual-Texture Gradient

Pattern in terms of range to detection and range to identification against existingmeasures.

7.2 Method

7.2.1 General.The basis of the laboratory simulation was the use of several series

of 35MM dolor transparencies taken at various ranges from most distant to least dis-tant from target panels painted with the camouflage patterns being compared. Subjectsviewed the slide series and detected the targets and identified their form at vary-ing ranges due to pattern characteri.atics and individual abilities. Mean detectionand identification ranges foz each pattern were recorded and compared.

7.2.2 Panel Preparation: In both summer and winter phases, the targets were4 x 8 foot panels of insulation board with a 2-foot section of corner removed at anangle of 45 degrees (see appendi-: 1). The panels were painted with a mixture of stan-dard pigment provided by MERADCOM (Forest Green and Field Drab) and matte-finish latexpaints locally purchased and matched to the US Army standard camouflage pallette.

7.2.3 Two separate test sites were used for preparation of the slide series:

7.2.3.1 Summer phase: open field bordered by trees and shrubs at Stewart ArmySubpost, Newburgh, NY. The field was surveyed and marked to allow a straight line ofapproach to the targets. Targets were emplaced at a point in the tree line and a pathmeasured in 25 foot intervals to a distance of 675 feet. This path served as theguideline for positioning the camera. A series of 22 35mm slides was taken of eachof three targets: DTG, US Arqr Pattern, control Danel (Forest Green). The photoseries was taken between 1330 and 1530; there was ..o cloud cover; since the distarcewas relatively short, coefficient of transmission was considered a negligible factor.In order to maintain uniformity between target-series, the following procedure wasfollowed:

a. Photographs were taken at 25-foot surveyed intervals, from 675 to75 feet.

b. Target panels were shifted at each position without moving thecamera; in this way, the elapsed time for all three photographs at each position wasextremely short (less than two minutes in most cases), and illumination conditionswere thus kept extremely uniform among target-series.

7.2.3.2 Winter phase: Upper Reservoir for the Village of Cornwall-on-Hudson.This is a small (less than 1O00-meter diameter) lake b 'ordered by high ground and hard-wood trees and shrubs. At the time of the photo-series execution, the lake was frozen

and snow-covered, providing an ext:.,emely level open art a with favorable background.Procedures were the same as those described in paragraph 7.2.3.1, except that a fourthtarget panel had been added, using the Swedish Armi pattern. All panels had beenmodified for winter (snow) environment (see photographs, appendix III).

7.2.4 Cadet subjects viewed the various slide series for summer and winterphases. The subjects were not familiar with target characteristics and had not beengiven the specific objectives of the experiment. The following proceduces were used:

-64

7.2.4.1 Groups of 4-5 subjects were given the following instructions beforemo;'ing to the test position:

"You will be shown a seris of 35MM color slides of an open field and

Sa tree line. Each picture will be on the screen for five* seconds. Each pic-ture will be closer to the tree line than the one before it -- in effect, youwill seem to be moving toward the tree line.

44j "A target has been placed in the tree line. It is visible, but par-Lially concealed. (The target is a panel cut in one of the four shapes shown

~on this card)**"Your task is to search the tree line until you have detected the tar-

get. DO NOT GUESS: be fairly certain you see the target before you signal detec-- -tion. When you do see the target, signal 'stop' and the projectionist will

stop the series on that slide. At that time you will be asked by the controlleroto point out the target.

"Do you have any questions?"

7.2.4.2 At this point, subjects were moved to the test position, where the"slide series was administered individually to each subject.7.2.4.3 Each subject was seated approximately one meter from a translucent rear-projection screen. The controller was seated to the rear of the subject and out ofthe subject's view to avoid inadvertant nonverbal =es. The slide series was projectedfrom the rear of the screen; stimulus size was 1.4 meters horizontally by 1.0 metersvertically (see photograph, appendix III). Subject was provided a pointer and, in thewinter series, a card showing four alternative target shapes (see figure 3).***

M

-- '-

Figure 3. Sample stimulus card.

7.2.4.4 When subjects made a detection, the controller required verificationby asking the cadet to point to the target with the pointer; in the cases of incor-rect detection, the controller replied: "that is not correct, continue the series."

* Summer only; in the winter phase, exposure time was 15 seconds per slide.** Parenthetical instruction in winter phase only.** In the summer series, the subject was given the sample card only after a correctdetection.

15

Ip

This procedure was repeated until a correct detection was made; incorrect detectionsand correct detections were recorded by slide number.

7.2.4.5- When a correct detection occurred, the subject was read the secondpart of the instruetionic:

"Yo have correctly detected the target. ""Your next task is to identity the geometric shap6 of the target. The

slide series will continue from the slide at Which you detected the target. Thetarget is a panel cut in one of the shapes shown on the stimulus card. When youare fairly certain you can identify the shape as A. B, C or Di ss 'stop' and '

M give the identifying letter of the shape You think is correct. khether youridentification is correct or not, the series will continue.* -f, after the firstidentification, you wish to change your answef, the procedure is the same as be-fore: sey 'stop' stud-give the new identifying letter.

"Do you have any questionsV"

The slide series was continued from this point until a correct identification had beenmade and recorded.

7.2.4.6 At the end of the series, each scurject was ubetwnished not to discuss thenature of the experiment until his individual instructor indicated that the experimentwas over, in order to avoid inadvertntly disclosing tetarget location and shape tofuture subjects.

7.3 Results and Analysis

7.3.1 Summer Phase7.3.2.1 Target detection and identification means for experimental groups

id the summer phase are shown below:

DEV2CTION IDENTIFICATIONGROUP A (US ARM PATTERN) 11.97 14.22

GROUP B (DTG) 15-33 17.72GROUP- C (CONTROL) 12.68 13.97-

7.3.if2 e: an scores are exprensed as slide number from 1-22, since the slidesonly approximate the conditions at the actual ground ranges. Analysis indicatedythatthe means for group A anCon id rnot iffer significantly; Group B mean differed fromGro.24. A and C in the predicted direction and beyond the .01 level of significance.tResults and analytsis of the summer phase test are exumtned in the intermhrepoer on

that hasve (inoerednce 4, appendix Vrtt).

7.3.2 Winter Phase7.3.2.1 Target detection and identification mean scores are shown below:

This instruction was added to prevent identification by elimination.

DETO DETFCTO

iDTCTO IDNIICTO N

GRU 1 (U ;RN PATEN 10;54 44

GROUP 2 (DTG PATTERN) 18.1471 19.3529 34

GROUP 4 (CONTROL PANEL) 18.9643 20.2857 P.8

7.3.2.1 Analysis

The population parameters for eight target series patterns were esti-mated. The mean score parameters estimated are listed below:

-4

GR1D 1 US Arf Pattern detection mean score.CR 2D - (G Pattern 1t 9" "i 2

7 2 3D - Swedish Pattern " Ith e4D p Control Pattern o e t s

V l, - US Army Pattern identification mean score.-2I DTG Pattern "It I

~21 T atr It itS31 " Swedish Pattern " "1 Control Panel II t

S1 Analyses of sample distributions indicated that only the US Army Pat-I ttrn target detection and identification scores provided a normal distribution. For

the other target series, photographs were not taken sufficiently close to the targetto allow all subjects to detect the panel. Consequently it was assumed that theywould have detected and identified on the next slide (slide 23), and the score of 23was assigned to those who fell into this category. This was a conservative measuremaking significance of t-tests more difficult to obtain as it forced the estimatedmeans closer together. Since the t-test is robust and sample size was sufficientlylarge to allow departure from normality, independent t-tests for differences be-tween means could still be conducted. The hypotheses and results are listed at Table

A post-hoc comparison of US Army and control panels detection and identi-fication scores was conducted using the Scheff9 post-hoc comparison test. This

* test requires an overall significant F from an analysis of variance. Consequently,one-way analysis of variance tests were conducted for both detection and identificationscores comparing all four pattern series. The analysis of variance tests are detailedat Table 2. The results of the Scheff4 post-hoc comparisons are listed below:

H 0 : t'lD - 4D 0H 1 : u - F4D 0

-11.85

F-R msm

HO - H' 0

H1 : 11I - '4I 0

HO : 12 41 = 0 -. 78 60 Accept H0Hi : L21 41 < 0

m l. Test was significant beyond the .001 level.Im - Test was significant beyond the .05 level.!1 I

TABLEA. 1 Xk 4 ANALYSIS, OF VARIANCE FOR-TARGET-DETECTION

Source df Sum of Squares Mean Squares FBetween G-oups 3 1657.83 552.61 17.64*

MWithin Groups 132 4136.3ii 31.33Total 135 5t93.93

b. 1 x 4 ANALYSIS OF VARIANCE FOR TARGET IDENTIFICATION

Source df Sin of Squares Mean Squares FBetween Groups 3 174o.97 580.32 26.42 *

Within Groupr 132 2898.92 21.96Total 135 4639.88

Significant beyond the .001 level.

7.4 Discussion

Evaluation of the results in a test of this nature must go beyondsimple statis -tical analysis; a wealth of subjective experience emerged from the

~relatively straightforward methodology which is of interest to the Ar and isin some cases a starting point for fu~rther hy-potheses. The following observationsseem worthy of interest:

7.4.1 The nature of camouflage as a perceptual phenomenon has been somethingof a mystery. The general ambivalence towards research in camouflage is sumarized

by John R. Bloomfield in Visual Search:

Although of major importance, problems of camouflage have received,little systematic investigation. This is due, in part, to the great difficul-ties in defining camouflage situations. .. While it is possible to make quiteprecise theoretical statements about the artificial stimuli used in competitionsearch tasks, we are still a very long way from being able to deal with thekind of complexities involved in camouflage situations. (Reference 5, appen-dix VI.)

However, the authors are not uncomfortable with the summary of indi-vidual observer properties at paragraph 1.5.2. This is due to an experiment con-ducted at the same time as the effort described in this report which is reportedunder separate cover since it'is of interest to different agencies than the presentreport (Reference 6, apperldix VI). In this experiment, the results of individual

" performance on two perceptual psychometric tests designed for the purpose werecorrelated with performance on the slide series described in this report. When

,44 &b

ri A

certain physic!ogical characteristic were controlled, the multiple R was shown bo behigher than expected (0.49); while this is not powerful enough to be predictive,it does cast some light on the nature of camouflage, giving empirical support to thehypothesis that detection oP camouflage is a combination of visual search habits andfairly specific and stablie perceptual organizing properties. This is of some impor-tance, since the DTG pattern was designed based on the presumption that these fac-tors were of some strength in camouflage detection. The reader is urged to comparethe specific findings of the referenced report with the results of the present ex-perimental effort.

7.4.2 The performance of the control pattern in summer and winter phasesis somewhat confusing on the surface. The green panel performed as predicted in thesummer phase, significantly less effective than the DTG Pattern, but not signifi-cantly different from the US Army Pattern.* In the winter, however, the controlpanel was not significantly different from performance of the DTG, and significantlymore effective than either conventional pattern (US Army or Swedish). This appearsdue to two special factors:

S7.4.2.1 Many subjects failed to detect either DTG or control panel bythe endof the slide series. Such observers received the score 23 (there were 22 slides),since the next logical slide would have ',een at a distance of approximately 50 feet.The authors presumed that most subjects could not have avoided seeing any of thepanels at so close a range. However, this produced an artificially skewed distri-

Sbution of scores for DTG and Control panels; hence, for comparison between thesepanels, the means are somewhat bogus, and the true means might have yielded a moreI, definitive comparison.7.4.2.2 The site chosen appears to have favored the control panel (see photo-

rgraph, appendix III). There was very little clutter in the area of target position,and the background in the immediate area consisted of deep snow and very light brush.In addition, for most of the slide series a level road embankment coincided with thetop edge of the panel, making shape cues virtually nonexistent. The problem was

__ I recognized by the camera team during the last stages of the slide series execution,but time and funds were not available for a replacement series.

7.4.3 Subjects made a number of comments during the slide series test. Inmany cases, subjects were unable to recognize the DTG as a pattern even when the tar-get outline was traced on the screen by the controller. The conventional patternswere much more obvious, since their low texture (in comparison with the background)simply replaced a target shape-schema with a clearly visible pattern schema.i '-7.4.4 A shortcoming of unstructured camouflage demonstrations (the sort whichhave traditionally given pattern-mainting a bad name) was evident in the test, Therole of perceptual set is extremely powerful. The panels - all panels - in the win-ter series were clearly visible to tVe controllers from the first slide, and the con-trollers expressed occasional asto:. Ahment at the subjects' inability to see thetarget. This illustrates what is probably the most important single factor in camou-flage detection: knowing the nature and location of the target will defeat any meas-ure known. If you know what the target look& like and where it is, its signature willusually be overwhelming; but this does not mean that it will be detected easily -bya nalve observer. This general point is illustrated by a popular classroom demonstra-tion of perceptual set: a degraded gestalt-completion slide with visual "noise" de-picting the face of Jesus; most students fail to see the face until it is explicit-ly delineated, but once perceived, the image is there to stay (see figure 4).

This is not surprising, not an indication of any ineffectiveness of the presentUS Army (MERADCOM) pattern; the panels were not garnished (in order to isolate thepattern as independent variable), and the US Army pattern is specifically designedfor use with garnish. The DTG was predicted as the high performer in this scenario.

1-1

7.4.5 The slide series seems to offer potential as a training aid at unitlevel. It removes unwanted variables and suggests any number of refinementsi and,importantly, is capable of production at unit level with a minimum of expense.The authors caution, however, that certain relatively stable and innate skills arepowerful contributors to detection ability with ambiguous stimuli, and there isconsequently a probable indivividual ceiling on ability, at least as far as purelynative perceptual properties are concerned. Hence, great weight should be givento familiarizing soldiers with important target signatures in ambiguous or degradedsituations, then validating their familiarity with these cues in the slide seriesformat.

7.5 Conclusions

7.5.1 The Dual-Texture Gradient Pattern proved significantly more effectivethan the US Army Pattern in Summer and Winter conditions in terms of range to detec-tion and identification.

7.5.2 The Dual-Texture Gradient Pattern proved significantly more effectivethan the Swedish Ar.'W Pattern in Winter conditions; the comparison with this pat-tern was made only in Winter.

7.5.3 The Dual-Texture Gradient Pattern proved more effective than an un-patterned control panel in Summer; in Winter, apparently due to special environ-mental factors and problems with data distribution, there was no significant dif-ference between DTG and the white control target.

"IM.

Z' l'

Uki

APPENDIXC I. PATTERNS

Patterns shown in this appendix are those used on the target pane].s.Patterns were identical for both summer and winter phases; however, the light greenused in summer was painted over with matte white in winter. Colors shown abbrevia-ted in the drawings are:

FG =Forest Green; 10 Light Green; FD =Field Drab; B =Black

1 A. DUJAL TEXT1URE GRADIENT PATTERN

01 E3E3 0

CMf

B. US ARSW (MERADCOM) PATTERN

IT

All-

4 4 4'*,-kzA -,tY -. '.-,>.,)O"Z,'--

4

I - - -

-Ip 1,

C * SWEDISH ARMY PATTERN

sJ

LG

'C

I C

ClA

C4

I ,-. 'C

SA

APPENDIX II. PIGMENT SAMPLES

The following color samples areprvddfrcmaio;F esGreen and Field Drab colors were provided by pSMrADovie oIMsn.ra

*1Forest Green Field Drab

Light Green White

Black

APEDX I.POTGAH

A. SME ETST STWR IYSBOTB. WNE ETST UPRIEVICRWL-l-USN

4

-

;r * -,4KcZ>

-t

2,;

-F

9';;4

C>$4 *9S

'F""'4

"4

St

4

r d4Ws-$tiL.

'S 4

544,>1 t4'

9

*4

I4

I"C, 3

4

'F

-I'5'-I

S

-El'4

I

4'Ii .7;fr55542FA4 777iF44

it-I 4 4 . 4-

>,:4,S4 '

95)

0t

MAW

Eli

04.

0

0r

01,

I U

A, 4

'5?

4'

-A

A

'A

A

I!.

4

Al

4

'A tnd,'.AAAA, -AS-.- AA 4 ,I'd'

APPENDIX IV. SIECT SELECTION DATA

A. Subjects were male cadets from the Class of 1979 (sophomores) assigned toclasses in General Psychology (PL 202). Approximately 260 subjects were employedduring the course of the experiment.

B. Subjects varied from 18 to 21 years of age.

C. Following visual requirements are uniform for the United States Corps ofCadets:

Vision: (Disqualification criterion) Distant visual acuity not correctibleto at least- 20720 in each eye with spectacle lenses.

Miscle Balance: (Disqualification, criterion)

Esophoria aver 15 prism diopters.Exophoria aver 10 prism diopters.Byperphoria over 2 prism diopters..Stabismus (tropia) of any degree."!

SColor Vision: Mst be able to distinguish vivid red end vivid green.

Refractive Error: (Disqualification criterior ) Myopia over 5.50 diopters in anymeridian.

Hyperopia over 5.50 diopters in any meridian.Astigmatism all type over 3 diopters.Anisometropia over MO5 diopters.

7577777' qvt

APPEIDIX V. DATA

The individual subject data listed is reproducei from computer output.Key to listed variablea-

VAR 1 Subject Identification NumberVAR 2 Section Number (Identifies instructor/controller)VAR 3 Wears spectacles? (1 = yes; 0 = no)VAR 4 Wears contacts? (1 = yes; 0 = no)VAR 5 Color blind? (1 = no)VAR 6 Degraded Letter Test Score (Not used in this report; see ref

4 and 5, appendix VI)VAR 7 Cue-Search Test Score (Not used in this report; see ref 4 and

5, appendix VI)VAR 8 Distance at Target Detection (expressed as slide number,

1-23)VAR 9 Distance at Target Identification (expressed. as slide number,

1-23)VAR 10 Number of incorrect detectionsVAR Ui Number of incorrect identificationsVAR 12 Group Number: 1 = US Arqj Pattern

2 = DTG Pattern3 = Swedish Arqy Pattern4 = Control Pattern

Li

*1l

Ii In km..

-0 -0 4c 0

44

4z40o 00 0 C 0 00 0 C00

00 c o00 0 0 0.c o 0 0 0

Nu *j *l fn I, v *n *u *- II N N

c. cc~ O m 00 cc~ w' 00 w c0z

49I

,so>

4j z 0 0 0 00 0 0 0 I 0 z 0 n~w A. * *L . . .A i . ) . . * * . "A

LA) u O rI) 0 W jol 0 u nf% rv 'O v) ) Lf rA' CO VA' rj0 & r

000 00 00 000 00

000 00 O 0-C 2 tA U) A * *A * A

0p -c- - > 0

4j Aj > 4- 4j

MM . 17 4

V00 0000 00 0000 0.000 00 oc

LLJ"I N 0%U0V

4cc .4

0) 0 000 00

13 n 0nCD n-0 fn 00 n 0

4iitNj

tvf vN f t Clr

0rC rc =Mn rf k

-C< 4c- - x-c-

00 z.0000 0 a 07 00a 0 C

LL- tn- -0 ( C %1 n - 'm0 W r.J' ) o-0 n a 0 a

a.- I v 4C w C T- - 14 'x 4 t

L'4

00u0 00 0 0 0 000 0 000 00 0 0 0

tn tn tn55*

U '0. 1.- 1.- 1.- 0 N- - 1.-- o " I.- - -U , -* NN0C- C 0 -44C -4C 04 4-4 44c 444

000 0 0 00 0 0

LA UN *n LA LM V% * *A*

4 CP. ~ 0 0t 40.0 Itde- C- -C -C- C-C- C-

0

In00 0l0 00 f 00 0 n1 nC)I 00

00o 0 0 00 00 0 0 C o

OC - 40 .l -0.

uj r4r

0c -0-C

Wi U% - gm-I.-I.

4 ~ - 44 4 4444 44

I.2 -C0

n -p

o 0 00 In 20w 0MaI0 00 0o-n

14.c

c00 000 000

000 000099 99 99 999.

tn -oL

M, ^ N F^~ em .f, .2 I N w-UN-.- In NN '0 NN, Na. 3

'I- -C c4 44C -C4 - 44 -C~44

co m

ci 4 0l0 z0 0 C3 2 0.3~~ w r 0 0

-C~4 CIO00 00 00 00 00O 00 00

w w 1c .U.U U. U. U.

wL 2 00000000 0 )C IA,

4~ ~ 2 In (1 n11 nLa 'o 0 1- -0I

cc M-c W at 0: W -ec 2 I W z MC-

4 . 0 44c 0 444- 0 4m44, -X 44c 0 4944 0 444'~~~~t Lb w. L ... L ... L ... L ... L .. I 000 I 000

,*A,., 00 S4CiC0. am~~

-0e

0 0 000000 0 0 I0 00 000 00 0

tn X) M 0 -n c- -nNJ11

0 00 0 0 m

Lu LfI1 II UL. f, U.. I, N ^ L- J r 74 2- 4xt

C C

CI 00 00 0 0I0

00~~ .- 0 0

ti0

M",.

In No 'n 00 00 01 0C

0 00 0 0: 0 0

n 0 - IVO In ~ In0

tv eyrv

-c

Zi, 0 0 In 00 00 00 00*C)000 0 00 00 0

*i * ) . . . . .U . . . *j . . .

z 0 0 0 -

m4L UwL uLtI 101 -- - 00 -1

z c-I e( :o Z 7 a r(l t( c z o ra :c

-c4 t4 w >

00 0o 0

4444444

0 0 0 0 0D 0 0 0JO 0 00 0 0n0PC 0 0n 0 0l

0

00 00 - 0-f 0 0% 0

0o 0 00 0 0 0 co 5o000a C~..N ) 0 0oc 00 00 N

(U MfuMeu Eu AU

I. I- r, Elf % f , '0 N . v? CD M .1)rF In

U)

U. LO 0UN ?1 , EUN CA( .- 0 in l. N0- t'13 w0 0U .inr%

0

3M uj 0 0 0occ cc cc 00 2 c 00 2 r 00 cc 00Cc :a 2 ! 07 00 CC 00ccI

0- 41C 00 00 00 00 - 00 002 ~ - -C -* 0 < -C -* -* -C -C oI 4c M III ~ II 3 I I0.J > 1f). > >. > > >- > t .-> 4j - 4j :p.

PIP

jC 0

00 000 0000c0 0

-N -C-0-

00 0 000 0 0 00

Ii Vi0 )0 0 C.0 0 00 a 00ooC~ 0

00

00- 0 00 00. 0 0x 00V% = l

-U -) tN NU Ni N

00 000 000 000 000 0c 0 00

-w 000 000 00 00000 0 0IL * 0

w LU Ni NU N N N N7

I. 0 E I >2 >A >2C >A >21 tj >2 > % >2 IA >*2

RtV '- II $ A

w'

.21 >4

4 - 00000 0 0 0 0

9- 0. 'T% It It'-

00

CD0 0 00 00 00 00 00

4C 44 -C -C C, 44 4 c

Ix 0 gN i 0

-0' C) 090 ONC 0W 0 0

LLI xoU000 0 0

:z~ ' 10 10 I 0 1'cC K'.C -CC 4cr -C-14 C-

44

000000000 00N 9.99 .9 . 9 .9 99

800 00 0ir 3p 0 00 0 00

- - ~-C -0 -C -0.-N V

a 0 03 0 000

Z1..

00 00 0 0 0 00oN0 0 0 c o0 0) 0 0 0

N t * ** **

c -0~~~ 0> 00

4 00 00 0 00D,0 00 000 00 000i - - 00 -C 000 - 000 000 00 000 000

us. N ONN.tN- N-

c uU

w 2 N0

w'

00 CIO , 0 00 0 0

CIOC 00 0 0 00 00

LU U%.-Aa :

41~ N C0 0 0 040 -40 I l ?o

-4 -C

0 C

.844

Pn 0 0n0 In 0 -n00- 00

~0

L4U ruevr r V k

N- Ai M* N In (% r- *U t- 1; 1 *-^ .

r- c J , :r- g v\J f\N fJr mc cI

I.- ~ U) vo0 LA t C LL Oor LA -A LU LA L 0-C

U t, ri) rLI-rrL

r; 0 CD 0 ) 0 CDC0a 00: 4j 2 LA LA LA LA uA II . .

LU~~~~~ ~ ~ ~ 44- 3D0 n00 w . l( I ,t N C n .

44

00r a k I 0 10 00c 00

n LD

-r-

0 0 0l 0 0 f I10 an

a-09 9 9 C0 0 0 0

0It00 -

- --11A r qr i l 1 l

=0 0: Ix n0 0j cc x 1 0 0 0Z0 04z

2-. C3 :Zz C00 . 0 M0 a) z o

*A . . . . u . . . .i . * . . SL * * *-iL 0f a t no LA~ f-.,. 0o 0 A r , 10 c 0 V) o0

c Nv r4C cN dc 0 In

w0 00000 0 0 0 0 0 0 0

2-

w. t. fl - 1- -,- P- - 0 - 0; t- -

S 0 4K0 M- 0 0c 9 0cu . 0 2 0 0 2 00 ~ .0 2 0 0

-~0000 a 00

0 00 0a 0 0 00

0. .3.3.

0 0:

0~ 0 9 9 9 9.

00 0 00

00000C 0 0

N * * *4* ** * *

00t

PsjKurjr m1-f0 m a: c92: ccoi cccc 4

ao mcc o c (zcc 4 -C-C C 44c 44

00 00 0 00>

w c0 00a:0

00 000 00 000 00 000 00

00 00000 000 00 00 0

N * * * * 4 * * * * 6 * **

.. .

4 N4 U% .4 *ON.' . .

o O

I ~ ~ Lu 4j . (14(. L 1 1

Lu Lu 4.. .j 2 o0 < 4c 0 c 0

4c1 4c ..t 4c 4t'0l 4c > N J 4

W >> LI L

00 00 0 00 0i 00

00 0~ 00 0

.. .9 C! SS

0o 00CDo 0 0v 0 0r 0n

cuI0.] -2C 0 0 0o0

0D0I0 0

0 ' *01 ONC 0P ~ 0- 0wS

n a. -n N0 -n 1onc-n -n

coo0 0 0D 0 0

-C a- wIa - w a: -l 0 .0

I InI~ % 000 00 0 00 00 0 0

t.*S5555 ** 5 *- 000 00 0 001 0 00 0 00

I. I N .. . .N U.; N.s .f N . UP0 N . N. Ili 0 aN...-

x u. ci. wi wJ U - US -L .4 .iO. 4 . 41

2m 0

w P- I-- 1- - 1.- 0 .; P- I.--I. cc 2:o ca 2 cc 2: cc ccac2 za z 0 tc rc4

Co 00 00 c

Lf% LLCc r cc C3 ac CC

n C0 0 0 v 'n c

-c 44 44 -C 44444

0 C 0 0 0 00 C0

COO.p% O - CC.4

u'A

4.0 -MU~*C~4

Lu ccguriCD O z'' 0c z2 C000 2' 0 20' C) 2

0 ~ w4 9- 44 . C! 94 4 - 44 4 4-u in f'-I 0' .20 (nV 0' 0 v v 0' N' 0 '4I- L .JL Lu Lu Uj L

0 UJ t2 I.- 'D-- -- - 4 .u 0' -C -C -C C CC CZ C C- C< C2J wO > 00 > 00 2 0 C C Z C 0 0 2 0

Laii kA cI

900 4 09c 0: 00 00000

AI

41 '0A1014 ,

0o 00 0 40 0 C0 co;.j 410 Oui 00r m r l

-C 41 - -Z - -C < gsC - CN 9

z ' 0 0 0 CD0 0 00 0Lai .9 ..9 .9.. U)u

k 4 44 44 4 4j 41 4

00 0 00000000000

--9

i I-If$.

4 0i 1 2 00 22! 0 -,D- $. $.-- 2 2 - 2ot0 4c 4A x 4 4 4 4 -E 4c

0C 0C 0

0C 0D00 0 0 4000 00 000Q

r -C -C -C 4 .0-0

>. > >>

-10 10 0st OIn vON0I 00.7 -C

-C -C 4cN

co10 ) o a 0 Lbr- 10 N-O It0 .J 07. 70'

mt000 000 000 0 00 000 000rAC 000x000 000 400 000 000 4c0 00-0 C C4c> > > > > > > > > > > > > > > >

N r J'x ..? mU.. a: cc cc- I 'A . s r.L,1jLL

R uj u w 7I - t (

2 L

-hL 0 a LU 0j . 0.

4~~ U) ww.. 41l t .l? U) O.'0 ' '-* )U) x)U)U

1. 0 - #- .-z I- "-.cc. 2 r cc ccc::i 2 2 c 2t 2re cc0~ < ;K -C 4c w . > 222 2 2 2 2 2 2b2 2 2

0 0

A~ 0#

.Mi

CD 00 00 0

0 0

0M 0 -n 00000

N * *~ ~~ 9 * 4* *9

0 7n ' 1 1 N '? Nl N N^1-

ru CQ

cc 0: m 0- 00 a: 0e0 j c:of t r

0) .

N4 - z N N

000 4Ao V 00 000 00 00

ILN Nj N kAN

Oio I-4 4 .- f-4 o- $.4 - " 4 '4. t-4 4 -2rg r r 2- o $- i-.

C -C < 00000-~

-

00 00 00c 0 002 ~~~~ ~ ~ ~ ~ * .C -C * . L . * * L * * 4 U *LU ~ '0t 0 0 d~uo 0 ONO 0 O O 0 V~0 0 .4 0 0 0.>

z'('-r 1-,

0 00 00 00C 0 0 00

0 00 00 00 0 0 00

00 1. 1 04 0 .0

0.0 0 00 0 0 a) 9

0 e 0 00g l . r 0m10i

-C4c4 4c 4c

00 0000

cr a: cc .

0~ - 0

00 c o0 0 0N 0 00

IVr-

00 00 09 . . . . 99 .

0JP 0P

(ALnV'0.Li 41 >

APPENDIX VI. REFEENES

1. (U) Adolph H. Humphreys and Sharon V. Jarvis, Camouflage Pattern PaintingReport of USAMERDC's Camouflage Support Team to MASSTER: US Army Mobility Equip-ment Research and Development Center Report iNo. 2090, February 1974.

2. (U) Sharon V. Jarvis, TECHNICAL MEMORANDUM: Fort Knox Test of CamouflagePattern Effectiveness: US Army Mobility Equipment Research and Development Center,August 1974.

3. John A. Barnes and N. William Doss, Human Engineering Laboratory CamouflageApplication Test (HELCAT) Observer Performance, Technical Memorandum 32-76: USArmy Human Engineering Laboratory, November 1976.

4. T. R. O'Neill et al., Interim Report - Evaluation of Dual-Texture GradientPattern and Investigation of Psychometric Correlates of Camouflaged Target Acquisi-tion and Identification: US Military Academy, November 1976.

5. T. R. O'Neill et al., Investigation of Psychometric Correlates of CamouflagedTarget Acquisition and Identification: US Military Academy, April 1977.

6. John R. Bloomfield, "Experiments in Visual Search", Visual Search: NationalAcadeey of Sciences, Washington, DC, 1973.

VA

'44

AMW

APPENID VII. DISTRIBUTION

ADRESSEE COPIES

Director, Technical Information 1Advanced Research Projects Agency1400 Wilson Blvd.Arlington, VA 22209

Defense Documentation Center 4Cameron StationAlexandria, VA 22314

Chief, Research and Development 2Washington, DC 20310

Commander, USATARADCOM 4ATTN: DRDTA-RHRPWarren, MI 48090

Commander 2US Army Natick LaboratoriesNatick, Massachusetts 01760

Commander 5MASSTERFort Hood, TX 76544

Chief of Engineers 2Department of the ArmyWashington, DC 20314

Commandant 1US Army Artillery SchoolFort Sill, Oklahoma 73503

Commandant 1US Army War CollegeCarlisle Barracks, PA 17013

President 4US Army Infantry BoardA Fort Benning, GA 31905

CommanderUS Army Training and Doctrine CommandFort Monroe, VA 23351

Commander 2US Army Engineer Topographic LaboratoriesFort Belvoir, VA 22060

A President 2US Army Airborne, Electronics & Special Warfare BoardFort Bragg, NC 28307

ka

President 2US Armry Armor and Engineer BoardFort Knox, KY 40121

Commander 2I The Combined Arms Center

Fort Leavenworth, Kansas 66027

Commandant 2US Arnwr Command and General Staff CollegeFort Leavenwortly, Kansas 66027

President 2US Aranr Artillery Board

I Fort 513, OK 75303

Commander 21 ~III US Armrj' CorpsFort Hood, 7X 765144Headquarters 6USAREUJR and 7th ArnqrApo wev York 09403

Headquarters 2I Corps (Gp)APO San Francisco 96358

Director 4US Arnqr Materiel Systems Analysis AgencyAberdeen Proving Ground, ND 21005

I. CommanderUS Arnqy Mobility Equipment Research and Development Command 10ATTN: DRXFB-RS (L. Gale)Fort Belvoir, VA 22060

Commander 5US Armyr Armor Center

A ATTN: Secretary of ArmorFort Knox, KY 40121

Commander 5'I US Arzwj Infantry Center

Fort Benning, GA 31905

Superintendent 1US Military AcadenWIWest Point, NY 109964Director 5 tI Office of Institutional ResearchDirector 5Office of Military Instruction

Director 10Office of Military Leadership

IP15 WNW WIF -

-, . -

Prof'essor and Head 2i Department of~ Earth Space and Graphic Sciences

US Military Acadeny~ej West Pcint, N~Y 10996

VII-3

. 4044111i I,

UIMA

*,,N -

.*

All

--