1、The Institute of Electrical and Electronics Engineers, Inc.3 Park Avenue, New York, NY 10016-5997, USACopyright 2005 by the Institute of Electrical and Electronics Engineers, Inc.All rights reserved. Published 1995. Printed in the United States of America.IEEE is a registered trademark in the U.S. P
2、atent however, copies can be obtained from the IEEE, 445 Hoes Lane, P.O. Box 1331, Piscataway,NJ 08855-1331, USA.IEEERESOLUTION OF CAMERA SYSTEMS, 1993 TECHNIQUES Std 208-1995 34. Test and measuring equipment requirements4.1 Test chartThe essential tool for measurement of resolution is a suitable te
3、st chart. The IEEE Resolution Chart, 1993(gure B.1 in annex B) is recommended for this purpose. This chart may be used directly for measurementof television cameras.NOTEIf the measured amplitudes at numerous TV line numbers are plotted, the resulting graph is dened as the hori-zontal resolution resp
4、onse characteristic. Horizontal resolution response is referenced to TV lines (square wave pat-terns), and is dened as a percentage of the referenced value, 100%.In addition to being used for the measurement of resolution, the IEEE Resolution Chart, 1993may beemployed as an aid in checking scanning
5、linearity, aspect ratio, interlacing, shading, streaking, and ringing.4.2 Picture monitorWhen a picture monitor is used to measure the resolution of a camera system, a monochrome monitor isrecommended, and it should have a resolution capability exceeding that of the camera system to avoid mea-sureme
6、nt errors.4.3 Oscilloscope/wave form monitorFor the measurement of the resolution response of a camera system, an oscilloscope or wave form monitorof sufciently wide bandwidth is required to avoid measurement errors. Frequency response should beuniform and equal to or greater than that required to a
7、ccurately measure the system/circuit under test. Thehighest fundamental electronic signal frequency resulting from the IEEE Resolution Chart, 1993(see gureB.1) when used with a television system conforming to system M (525 line, 59.94 eldsee table A.1), is12.7 MHz.The electronic signal frequency may
8、 be calculated for any value of TV lines on the resolution chart by thefollowing formula:Frequency = (1/2) (TVL) (Width Height) (1 Horizontal Active Line Time)The oscilloscope/wave form monitor should be equipped with a line selector that will permit the selection,display, and identication of any de
9、sired horizontal scan interval once each frame time.IEEEStd 208-1995 IEEE STANDARD ON VIDEO TECHNIQUES: MEASUREMENT OF45. Measurement procedure5.1 ConditionsBefore a signicant measurement of resolution is made, it is essential that the television camera systemunder test be properly adjusted, and tha
10、t the IEEE Resolution Chart, 1993(see gure B.1) be oriented andaligned perpendicular to the optical axis to permit the camera to accurately reproduce its image. Care shouldbe applied to assuring that the base of the camera and the edge of the chart rest on, or be referenced to, thesame level surface
11、. Further, the geometric center of the chart must be aligned, and perpendicular with theoptical axis (center line) of the camera lens and optical system. Finally, after adjusting lens focal length (ifequipped) and other lens adjustments to their optimum settings (see condition 5.1.7), the distance b
12、etweenthe chart and the camera should be adjusted until the entire resolution chart is just completely reproduced bythe camera. Two arrowheads located at each edge of the resolution chart are provided to aid in the correctorientation and the complete composition of the chart within the picture raste
13、r observed on the monitor. Thetest pattern on the chart is correctly oriented and composed when the camera is adjusted such that the pointof every arrow just touches its corresponding edge of the picture monitor raster.After the test pattern has been correctly oriented with respect to the camera, th
14、e following items are amongthose that shall be given attention:5.1.1 Scanning and interlace For cameras that utilize imaging tubes, care should be taken to check and adjust for proper scanning size,scanning linearity, aspect ratio, and interlace. For measurement and adjustment of scanning size and l
15、inearityand aspect ratio, it is suggested that reference be made to IEEE Std 202-1954.2Interlace will affect vertical resolution; hence, any adjustments in the system that inuence interlace shouldbe optimized. 5.1.2 ShadingIf the camera equipment employs signals for camera-shading correction, two me
16、thods for proper adjustmentare suggested:a) Visual inspection of the picture monitor to determine if the background is an even grayb) Use of the wave form monitor to determine whether the average picture signal axis is parallel to theblack level line, both at line and eld frequencies5.1.3 Nonlinear
17、video signal processingThe inuence of nonlinear video signal processing such as gamma correction, white compression (knee),black compression, and expansion (stretch), should be avoided to assure accurate and uniform resolutionmeasurement.5.1.4 Image enhancementThe inuence of image enhancing video si
18、gnal processing such as aperture correction, contour correction(detail), and others, should be avoided to assure accurate and uniform resolution measurement.2Information on references can be found in clause 2.IEEERESOLUTION OF CAMERA SYSTEMS, 1993 TECHNIQUES Std 208-1995 55.1.5 Streaking and ringing
19、Streaking following any one of the horizontal black or white bars is an indication of low-frequency distor-tion in the video processing circuits.Ringing, or multiple echoes following any single vertical line, is a function of the high-frequency responseof the system. These echoes may be confused wit
20、h the multiple lines of the resolution wedge and hence leadto an inaccurate determination of resolution.5.1.6 FocusAll optical and electrical focusing at the camera should be optimized.5.1.7 Light levelThe desired light level value is that which permits all of the elements of the camera system to be
21、 adjusted totheir optimum performance positions. For instance, lens settings for iris, focus, and focal length, are usuallyoptimum at about the center of their adjustment ranges.5.1.8 Signal component to be measuredFor color cameras, it is recommended that resolution be measured only for the luminan
22、ce video signal. Colorsignal components contained in the output signal should be disabled or eliminated. The preferred point formeasurement of resolution is the camera output connection that the camera manufacturer directs be con-nected to a principal program input of a video system.5.2 Measurement
23、technique5.2.1 Measurement of limiting resolutionAfter assuring that the conditions described in 5.1 are satised, limiting resolution may be determined byobserving the reproduced resolution chart on a picture monitor. The limiting horizontal and vertical resolu-tion of the television camera chain an
24、d picture display combination is determined by observing the point atwhich the individual lines of the graduated wedges are no longer discernible as separately dened images.The resolution readings of both horizontal and vertical wedges will indicate the system performance underthe conditions of the
25、test.Camera systems may have different values of limiting resolution in different areas of the image. Thiswould be observed by comparing the wedges in the center circle with those in the four corner circles of thetest chart (see gure B.1). Unless specied with the measurement, the resolution cited ap
26、plies to the centralportion of the picture.5.2.2 Measurement of horizontal resolution response of camera systemsThe oscilloscope/wave form monitor should be connected to the luminance video signal output of the cam-era system to be measured. A reference reading should be obtained by noting the ampli
27、tude of the video sig-nal for the transition between the long horizontal black bars and the white background. The long black bars,located at the bottom of the large center circle, can be used to determine the reference black to white videolevel before measuring horizontal resolution response. Horizo
28、ntal resolution response at any TV line numberis measured as a percentage of the reference black to white video levelthat is dened to be 100%. As the vertical wedges of the test chart, used to measure horizontal resolution, are selected and displayed,there will be ve or nine cycles in the video sign
29、al. The duration of the sweep of the oscilloscope/wave formmonitor should be adjusted to permit a clear display of the electronic signal wave form corresponding toIEEEStd 208-1995 IEEE STANDARD ON VIDEO TECHNIQUES: MEASUREMENT OF6these bursts. Line numbers should be selected corresponding to the num
30、ber of TV lines on the chart, fromthe minimum, to the maximum for which all of the cycles are still discernible. The maximum number of TVlines still discernible corresponds to the limiting resolution. The relative peak-to-peak amplitude of thecycles should be noted for the different TV lines to give
31、 a complete measurement of video signal amplitudeversus television line number on the chart.NOTEThis measurement is affected by the fact that the resolution wedges of the chart produce an optical square wave,rather than a sine wave. This means that the reading of horizontal resolution response at a
32、given line number will beinuenced by the response of the system being measured to the harmonics of the square wave.Sampling techniques that may be utilized within the camera system, including the pixel array of CCD image sensors andvertical segmentation by scan lines, and others, will produce aliasi
33、ng. When the pattern of the test chart exceeds one halfthe sampling frequency, the camera system may produce a lower frequency alias component that is superimposed on thecorrect reproduction. All of the measurements previously described can be distorted by alias components.a)Limiting resolution.If t
34、he alias component becomes predominant, it can be mistaken as extending the limitingresolution, even though the full number of cycles are not reproduced. The correct quantity of alternating blackand white lines should be clearly discernible for valid limiting resolution measurement.b) Horizontal res
35、olution response.The alias component can cause the ve or nine cycles of alternating black andwhite lines to have various peak amplitudes. To minimize the inuence of aliasing, the averages of the positiveand negative peaks should be used for measurement of relative response (see gure 1).5.3 Presentat
36、ion of dataThe following data should be recorded.5.3.1 Quantitative data presentation5.3.1.1 Test chart illuminationFor reference purposes, light level and lens settings ( numbers) should be recorded.5.3.1.2 Using the picture monitorAt the points of interest, the limiting horizontal resolution and l
37、imiting vertical resolution should be mea-sured and recorded.5.3.1.3 Using the oscilloscope/wave form monitorThe peak-to-peak amplitude of each burst relative to the appropriate television line number for each burstthat is of interest should be measured and recorded. Measurements, as instructed in 5
38、.2.2, should be taken.Where aliasing is present, special note should be taken.5.3.2 Graphical data presentationThe measured values may be plotted to display the horizontal resolution response characteristic. Such acurve denes the horizontal resolution response of the camera system. From this charact
39、eristic, the videosignal amplitude response is measured as a percentage of the reference amplitude for any number of televi-sion lines that the camera system can reproduce.A reading may be taken of one or more of the following signicant points, as appropriate:a) The chart line number for which the h
40、orizontal resolution response is half its reference value (halfamplitude response)b) The chart line number at which the horizontal resolution-response characteristic approaches zero(corresponds roughly to the limiting resolution)IEEERESOLUTION OF CAMERA SYSTEMS, 1993 TECHNIQUES Std 208-1995 7c) The
41、value of the horizontal resolution response corresponding to a chart line number of 350 lines(corresponds to the approximate maximum video frequency that may be transmitted under UnitedStates broadcasting standards for monochrome television)NOTEAn extensive discussion of the subject of resolution, e
42、specially in terms of the oscilloscope method of measure-ment, is given by O. H. Shade, Image Gradation, graininess and sharpness in television and motion picture systems,J. SMPTE, vol. 56, pp. 131-171, February, 1952; vol. 58, pp. 181-222, March, 1952; vol. 61, pp. 97-164, August, 1953;vol. 64, pp.
43、 593-617, November, 1955.Normal Resolution Responseand AliasingResolution ResponseFigure 1Effect of aliasing on the horizontal resolution response wave formIEEEStd 208-1995 IEEE STANDARD ON VIDEO TECHNIQUES: MEASUREMENT OF8Annex AInstructions for the use of the IEEE Resolution Chart, 1993(informativ
44、e) A.1 PurposeThe IEEE Resolution Chart, 1993was designed to provide a standard reference for the measurement of theresolution response of television cameras and other video devices. In addition, the chart serves as an aid inthe observation of other performance characteristics such as streaking; rin
45、ging; aliasing; and geometry, lin-earity, and aspect ratio of the picture. The chart may be used to observe some optical system performancecharacteristics as well.The horizontal resolution response of a television camera is usually limited by the resolving capability of theimaging device, not by the
46、 bandwidth of the video ampliers. Useful information and measurements of thepercentage of resolution response at numerous TV line numbers, and the limiting resolution response of atelevision camera, may be obtained by utilizing the IEEE Resolution Chart, 1993. For this purpose, horizon-tal and verti
47、cal wedges of TV line patterns, and other groups of TV line patterns have been arranged on thechart to permit resolution response measurement extending from 100 to 1000 TV lines.A.2 DescriptionThe 100 to 600 TV line number horizontal and vertical line pattern wedges in the center of the chart arecom
48、posed of ve black lines separated by four white lines. The 500 to 1000 TV line number wedges arecomposed of nine black lines, separated by eight white lines. All lines are of equal width at any TV linenumber point on each wedge. The numbers printed alongside the wedges correspond to the total number
49、 oflines (black and white) of the indicated thickness that may be placed adjacent to one another in the height ofthe chart. For example, if black and white lines having the thickness of those indicated at the 300 positionwere placed adjacent to one another, a total of 300 (black and white) lines can be tted into the height of thechart. Since the aspect ratio of the chart is 4 to 3, a total of 400 lines of this same thickness can be tted intothe width of the chart. The fundamental frequency generated by scanning through any TV line number onthe ve
