1、Designation: E 2208 02Standard Guide forEvaluating Non-Contacting Optical Strain MeasurementSystems1This standard is issued under the fixed designation E 2208; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision, the year of last revisi
2、on. A number in parentheses indicates the year of last reapproval. Asuperscript epsilon (e) indicates an editorial change since the last revision or reapproval.1. Scope1.1 The purpose of this document is to assist potential usersin understanding the issues related to the accuracy of non-contacting s
3、train measurement systems and to provide acommon framework for quantitative comparison of opticalsystems. The output from a non-contacting optical strain anddeformation measurement system is generally divided intooptical data and image analysis data. Optical data containsinformation related to speci
4、men strains and the image analysisprocess converts the encoded optical information into straindata. The enclosed document describes potential sources oferror in the strain data and describes general methods forquantifying the error and estimating the accuracy of themeasurements when applying non-con
5、tacting methods to thestudy of events for which the optical integration time is muchsmaller than the inverse of the maximum temporal frequency inthe encoded data (that is, events that can be regarded as staticduring the integration time). A brief application of the ap-proach, along with specific exa
6、mples defining the variousterms, is given in the Appendix.2. Referenced Documents2.1 ASTM Standards:E 8 Test Methods for Tension Testing of Metallic Materials2E 83 Practice for Verification and Classification of Exten-someters2E 251 Test Methods for Performance Characteristic ofBonded Resistance Str
7、ain Gages2E 399 Test Method for Plane Strain Fracture Toughness ofMetals2E 1823 Terminology Relating to Fatigue and Fracture Test-ing23. Terminology3.1 Definitions of Terms Specific to This Standard:3.1.1 accuracyquantitative relationship of the measure-ments to the value obtained by standard measur
8、ement tech-niques3.1.2 basic datadata obtained directly by the measure-ment system. For optical, non-contacting methods, a two-dimensional array of image intensity data is generally the basicdata.3.1.3 coherent illuminationlight source where the differ-ence in phase is solely a function of optical p
9、ath differences;interference is a direct consequence.3.1.4 decoded datameasurement information related tothe displacement or displacement gradient field.3.1.5 decoded data bandwidthspatial frequency range ofthe information after decoding of the optical data.3.1.6 derived datadata obtained through pr
10、ocessing of thebasic data. Typically, this is displacement field data.3.1.7 dynamic rangethe range of physical parameter val-ues for which measurements can be acquired with the mea-surement system.3.1.8 illumination wavelengthwavelength of illumination,z.3.1.9 incoherent illuminationlight source wit
11、h randomvariations in optical path differences; constructive or destruc-tive interference of waves is not possible.3.1.10 maximum temporal frequency of encoded datareciprocal of the shortest event time contained in the encodeddata (for example, time variations in displacement field).3.1.11 measureme
12、nt noisevariations in the measurementsthat are not related to actual changes in the physical propertybeing measured. May be quantified by statistical propertiessuch as standard deviation.3.1.12 measurement resolutionsmallest change in thephysical property that can be reliably measured.3.1.13 numeric
13、al aperture, (N.A.)non-dimensional mea-sure of diffraction-limitation for imaging system; N.A. = D/ffor a simple lens system, where D is lens diameter and f is lensfocal length.3.1.14 optical datarecorded images of specimen, contain-ing encoded information related to the displacement or dis-placemen
14、t gradient field, or both.1This guide is under the jurisdiction of ASTM Committee E08 on Fatigue andFracture and is the direct responsibility of Subcommittee E08.03 on AdvanceApparatus and Techniques.Current edition approved May 10, 2002. Published August 2002.2Annual Book of ASTM Standards, Vol 03.
15、01.1Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959, United States.3.1.15 optical data bandwidthspatial frequency range ofthe optical pattern (for example, fringes, speckle pattern, etc.)that can be recorded in the images without aliasing or loss of
16、information.3.1.16 optical integration timetime over which digitalimage data is averaged to obtain a discretely sampled repre-sentation of the object.3.1.17 optical resolution, (OR)distance, d = z / (2 N.A.),between a pair of lines that can be quantatively determined.3.1.18 quantization levelnumber
17、of bits used in the digitalrecording of optical data by each sensor for image analysis.The quantization level is one of the parameters determining thefidelity of the recorded optical images. It is determined by thecamera selected for imaging and typically is 8 bits for mostcameras.3.1.19 recording r
18、esolution (pixels/length), knumber ofoptical sensor elements (pixels) used to record an image of aregion of length L on object.3.1.20 spatial resolution for encoded dataone-half of theperiod of the highest frequency component contained in thefrequency band of the encoded data.3.1.21 spatial resoluti
19、on for optical dataone-half of theperiod of the highest frequency component contained in thefrequency band of the optical data. Note that decoded data mayhave a lower spatial resolution due to the decoding process.3.1.22 systematic errorsbiased variations in the measure-ments due to the effects of t
20、est environment, hardware and/orsoftware. Test environment effects include changes in tempera-ture, humidity, lighting, out-of-plane displacements (for 2-Dsystems) etc. Hardware effects include lens aberrations, ther-mal drift in recording media, variations in sensing elements,interlacing of lines,
21、phase lag due to refresh rates, depth of fieldfor recording system, etc. Software effects include interpola-tion errors, search algorithm processes, image boundary ef-fects, etc.4. Description of General Optical Non-Contacting StrainMeasurement Systems4.1 Figs. 1 and 2 show schematics of typical moi
22、r anddigital image correlation setups used to make displacementfield measurements. In its most basic form, an optical non-contacting strain measurement system such as shown in Figs. 1and 2, consists of five components. The five components are(a) an illumination source, (b) a test specimen, (c) a met
23、hod toapply forces to the specimen, (d) a recording media to obtainimages of the object at each load level of interest and (e)animage analysis procedure to convert the encoded deformationinformation into strain data. Since the encoded information inthe optical images may be related either to displac
24、ement fieldcomponents or to the displacement gradient field components,image analysis procedures will be somewhat different for eachcase. However, regardless of which form is encoded in theimages, the images are the Basic Data and the displacementfields and the strain fields will be part of the Deri
25、ved Data. Thisguide is primarily concerned with general features of (a) theillumination source, (d) image recording components, and (e)image analysis procedures. ASTM standards for specimendesign and loading, such as Test Methods E 8 for tensile testingof metals or Test Method E 399 for plane strain
26、 fracturetoughness provide the basis for (b, c).5. Error Sources5.1 At each stage of the flow of data in the measurementsystem, errors can be introduced. These are considered in thesequence in which they occur in this guide.FIG. 1 Typical Optical Moir Systems for In-Plane Displacement MeasurementE22
27、080225.2 Errors Introduced in Recording ProcessSince themedia used to record Basic Data can introduce additional errorsin the Derived Data, each set of experimental data must includea detailed description of the recording media used. If a digitalcamera system is used to record images, data to be inc
28、ludedshould be the camera manufacturer, camera output form (forexample, analog or digital), camera spatial resolution, dataacquisition board type, pixel quantization level (for example, 8bits), ratio of pixel dimensions, lens type and manufacturer.When photographic film is used to record images, the
29、 filmcharacteristics and method of processing, as well as lens typeand manufacturer used in imaging should be documented.5.3 Errors Due to Extraneous VibrationsDepending uponthe measurement resolution, system vibrations can increaseerrors in the encoded information which may result in addi-tional ex
30、traction errors. Provided that the period of vibration issufficiently small relative to the integration time, and theamplitude of the disturbance is small relative to the quantitybeing measured, sensor averaging may reduce the effect ofvibrations on the displacement fields and the strain fields.5.4
31、Errors Due to Lighting VariationsSince the BasicData is image data, lighting variations during the experimentmay affect (1) the actual encoded information (for example,phase shift in coherent methods) and (2) extraction of theencoded information. For incoherent methods, light variationsof several qu
32、antization levels may degrade the Derived Dataextracted from the images. Similar effects are possible forcoherent methods if there are, for instance, slight changes inthe wavelength of the illumination. In both cases, use of imageprocessing methods that are insensitive to lighting variations(for exa
33、mple, normalized cross correlation) will increase theaccuracy of the extracted data.5.5 Errors Due to Rigid Body MotionDepending upon themeasurement resolution, rigid body translation and/or rotationmay severely impact the ability to extract encoded informationfrom the image data. For example, if th
34、e translation is largecompared to the measurement resolution and the opticalresolution of the recording media is low, then the highfrequency encoded information may be lost.5.6 Errors in Extraction ProcessThe encoded informationextracted from the recorded images is degraded by errorsintroduced by th
35、e image processing method used. Errorsintroduced by the extraction process can be a combination ofrandom errors as well as systematic errors (for example,peak-estimator bias or drift in Fourier correlation methods).Improved methods for image processing may significantlyreduce extraction errors and s
36、pecial care should be taken toreduce systematic errors.5.6.1 For example, one can define an engineering measureof normal strain along the “n” direction as:enn5Lnfinal2 Lninitial!LninitialHere, ennis defined by Lnfinal=(Ln+ Dun)2+(Dut1)2+(Dut2)1/2and (Dun, Dut1, Dut2) are finite changes in displace-m
37、ent along the perpendicular directions n, t1and t2for pointsat either end of line Ln. Thus, errors in strain enncan be due to(1) errors in the initial length of the line element and (2) errorsin the displacement components (Dun, Dut1, Dut2). In bothcases, extraction of Derived Data from the Basic Da
38、ta is thesource of error.5.7 Errors in Processing Extracted DataErrors are intro-duced when the form of extracted Derived Data in 5.6 isprocessed to obtain strain data. This process can involve a widerange of mathematical operations including (1) numericaldifferentiation of derived displacement data
39、 and (2) smoothingof displacement or displacement gradient data. Errors intro-duced by the choice of post-processing method can include, butare not limited to, (1) reduction of spatial resolution, (2)systematic under-prediction of strain in areas of high straingradients, (3) phase errors in the sign
40、al due to non-symmetricoperators etc.6. System Evaluation Process6.1 Each non-contacting optical strain measurement systemmust be evaluated to determine reliable estimates for theaccuracy of the resulting Derived Data. Given the wide rangeFIG. 2 Typical Digital Image Correlation Setups For (a) In-Pl
41、ane Displacement Measurement and (b) Three-Dimensional DisplacementMeasurementE2208023of methods that have been developed, this guide will notaddress specific details involving the application of any tech-nique. Rather, the guidelines are provided as a general frame-work for evaluation of non-contac
42、ting optical strain measure-ment systems.6.2 In the following sections, a direct comparison betweenestablished measurement methods and non-contacting methodsis recommended. However, it must be noted that, even thoughthis approach does provide a direct, quantitative measure ofagreement between two, i
43、ndependent measurement data sets.Practice E 83 and Test Methods E 251 provide only averagevalues for strain over a specific area on the specimen. Thus,good agreement with the average value obtained from theDerived Data in the same area does not verify (1) the accuracyof local variations observed in
44、the Derived Data or (2) theaccuracy of the Derived Data in regions outside the area wherecomparisons were made.6.2.1 For example, if a finite difference in displacementcomponents is used to determine strain components such asenn, then errors in relative displacement components can bedirectly related
45、 to strain errors using equations such as:EennDLnfinalLninitialWhere Eennis an estimate for the error in the normal strainenn. Here, DLnfinalis the error in final length due to errors in themeasured displacement components. Through calibration, thecontribution of inaccuracies in the relative displac
46、ement com-ponents (Dun, Dut1, Dut2) to strain error can be determined.However, the accuracy of the Derived Data outside of thisregion, which may be a region of importance, cannot beverified without additional comparisons.6.3 Comparison to Standard Measurement Methods forSimilar Test ConditionsNon-co
47、ntacting measurements canbe made under diverse conditions (for example, high tempera-ture, in-situ structures, laboratory test frames, vacuum). Due tothe diversity of conditions, calibration tests are recommendedon similar components under similar conditions to assessaccuracy. In this approach, the
48、effects of those phenomenapresent in the test condition but not accounted for in laboratorytests or computer simulations are directly included in the errorassessment.6.3.1 For these tests, direct comparison of the non-contacting measurements to independent measurements byestablished methods using do
49、cumented ASTM procedures (forexample, extensometers (Practice E 83), strain gages (TestMethods E 251) whenever possible is the most reliable way toobtain quantitative estimates for the accuracy of average valuesobtained from the Derived Data. If this approach is used, alldata acquisition and analysis procedures must remain the sameas used in the actual tests, with clear documentation providedto demonstrate that the same procedure has been used for bothtests.6.4 Comparison to Standard Measurement Methods forSimulated Test Conditions in Laboratory Environ
