UNIVERSITY OF CALIFORNIA, SAN DIEGO May 1991 UC SAN DIEGO LIBRARY 3 1822 04429 7430 OPTICAT CVETEMS GROUP TECHN Offsite TE NO. 229 (Annex-JO rnals) QC 974.5 • T43 no. 229 Horizon Scanning Imager Project Review Viewgraphs J. E. Shields M. E. Karr UNIVERSITY OF CALIFORNIA SAN DIEGO The material contained in this note is to be considered proprietary in nature and is not authorized for distribution without the prior consent of the Marine Physical Laboratory and the Air Force Geophysics Laboratory TTYO .OF. Contract Monitor, Dr. H. A. Brown Atmospheric Sciences Division ili.it FORA LIGHT ORNIA HE To A. 1868 W. SI Prepared for The Geophysics Laboratory, Air Force Systems Command United States Air Force, Hanscom AFB, Massachusetts 01731 under contract No. F19628-88-C-0154 SCRIPPS INSTITUTION OF OCEANOGRAPHY MARINE PHYSICAL LAB San Diego, CA 92152-6400 UNIVERSITY OF CALIFORNIA, SAN DIEGO II III TIL TETTIIN INILI III UUDI I II DI III 1 UNIT TH DU 1 . DINI 1 III 11 LII 1 IIIIIII III UNI 11 III 3 1822 04429 7430 This technical note documents the viewgraphs used during the May 91 HSI project review. A summary of each set of viewgraphs is included. The presentation consisted of the following areas: HSI Software Progress Report 2. A Sensitivity Study of Daytime Visibility Determination with HSI 3. Determination of Visibility at Night Using the HSI Horizon Scanning Imager Software Progress Report Status of Programs - All software that requires the VS100 board have been converted to use the VS100 board. These programs are listed below. Primary Programs Setupvis - Creates azimuth selection file, allows user to change input parameters (range, threshold and inherent contrast). Vistex 15 - Automated visibility determination program Wsiten - Cloud cover determination program Utility Programs Fixhome - Fine tunes home position. AV5 - Program is now called Snglvis. Computes visibility for a range of inherent contrasts. Seevis - Displays image disk file to RGB screen. Savimg - Allows user to digitize an image, then save it to disk. Savrgb - Allows user to save image currently on RGB screen to disk. Viewvis - Reads 8mm data tape created by Vistex 15. Programs such as Move which move the rotary table to a user selected azimuth and Printvis which prints and translates visibility data files into a spreadsheet format are not affected by the VS100 board. Several suggestions were made for updates to the HSI software. These updates have been completed. 1) The cloud cover information created by the Wsiten program is now being saved to disk. The file consists of the total and opaque cloud cover amounts and a time stamp indicating when the cloud cover measurements were made. 2) The user may now define up to 16 targets in Setupvis and in the new program Snglvis (which is an upgraded version of AV5). 3) The target location information may now be found in two places. The target locations have always been available in the target files created by the Setupvis program. In addition, the locations are now embedded as header data in each RGB image. 4) A copy of the image used to compute visibility is now saved to quadrant 3 of the VS100 board image plane. The image saved does not include the target boxes or identifiers. The user may access this image using the program Viewvis and then analyze it with another program such as Snglvis. 5) The user can define a horizon ROI up to 256x256 pixels in Setupvis and Snglvis. The above changes have been added to the VS100 version of the HSI software only. Several new programs have been created for use with the VS100 board. 1) Snglvis is an upgraded version of the old AV5 program. This program computes visibility for a range of inherent contrasts. Several features have been added to Snglvis: A) All target files created with the old AV5 program and Setupvis can be used as input files. B) The range and threshold contrast can be changed C) Targets can be re-registered, added or deleted. D) Up to 16 targets can be defined. E) The horizon region of interest may be up to 256x256 pixels F) The RGB image can be archived. 2) Intgr8 allows the user to operate the Frame Integration Control Box via the computer keyboard. Images can be saved to the hard disk. 3) Getnite will allow the user to move to pre-programmed azimuths and apply a range of frame integration levels to each azimuth. All images are saved to Exabyte. Detailed descriptions and instructions for these programs will be delivered with the updated Night Visibility Hardware/Software package. Hardware Necessary to Run New Software ITI VS100 image processing board Imaging Technology's VS100 board replaces the FG100 board in the HSI system in order to take advantage of the external snap trigger feature of the VS100 board. The cue to grab an image is provided by the Frame Integration Control Box after the requested number of frames have been integrated. 80386 CPU We have found that with the presence of the VS100 board, we could no longer write to the Exabyte in our present computer setup. We were able to trace the problem to the 80286 CPU. By replacing the 80286 CPU with an 80386 CPU all Exabyte problems went away. Modified Frame Integration Control Box The upgraded Frame Integration Control Box has three new features: 1) Thumbwheel switch for local integration control 2) Local (manual) or Remote (computer) control of frame integration level. 3) Autolris cutoff controls have been added so that when frame integration is invoked the autoiris is set to full open. Autolris cutoff switch - An autoiris cutoff switch must be added to the HSI 2710) camera in order to completely open the autoiris when frame integration is being used. The V$100 board: Software Related Specifics There are a few things to know about the VS100 board if one is going to write software for it. First, some of the reserved bits in each register may be handled differently in the VS100 board than in the FG100 board. There are gain and offset controls in the VS100 board that are not present in the FG100 board. The above mentioned features are documented in the VS100 users manual. We use Werner Frei's VS100 version of the ImageTool software library for our image processing applications. The difference in register settings is transparent to the programmer when the VS100 version of Image Tool is used. A couple of VS100 characteristics that I could not find documented are 1) it is necessary to set the offset voltage before an image can be digitized and 2) if the zoom register is set before writing to Exabyte, the image will appear at 2X normal. Horizon Scanning Imager Software Progress Report Status of Programs • Hardware necessary to run new programs • The VS100 board: Things you need to know Conversion of Horizon Scanning Imager Software for the VS100 Primary Programs Utility Programs SETUPVIS VISTEX 15 WSITEN FIXHOME AV5 SEEVIS SAVIMG SAVRGB VIEWVIS ...... . Horizon Scanning Imager Software Updates • Cloud cover information written to separate file • Number of targets available: 16 • Target information is embedded in image • Image without target boxes saved to Exabyte • User may choose horizon ROI upto 256x256 pixels New Programs • SNGLVIS • INTGR8 · (GETNITE) Hardware Necessary to Run New Software • 80386 CPU • ITI VS100 image processing board • Modified Frame Integration Control Box The VS100 board: Software Related Specifics • Different register settings • Write to Offset register before digitize Zoom register set before write to Exabyte --> zoom - -- -• •-. -- .. ..... Annotations of Viewgraphs for Presentation on A Sensitivity Study of Daytime Visibility Determination with the llorizon Scanning Imager Further details on all of the material presented herein is contained in Technical Note No. 227. 1. Title and Outline We will show plots of the sensitivity of visibility, as determined by the FISI, to various measured and input parameters. This is followed by a discussion of potential system improvements implied by these results. Sensitivity to: Parameters which will be discussed in the presentation. 2 Determination of Visibility from Measured Contrast The form of the visibility equation which has been used for daytime visibility determinations. It is based on the Koschmieder equation, but determines visibility from measured values of apparent contrast. 4. Sensitivity of Derived Visibility to Inherent Contrast Note that the error approaches zero as the apparent contrast approaches .05, which occurs when the target range approaches the visibility. Errors can be significant for high apparent contrasts, i. e. when the target range is much smaller than the visibility. Changes in inherent contrast are expected, using natural targets; they probably provide the basic limiting parameter in the IISI. It is important to use targets at ranges near the visibility to minimize this impact. III 5. Sensitivity of Derived Visibility to inherent Contrast, when the Inherent Contrast is Lower Errors become larger than in slide 4. This illustrates the importance of using dark targets. Sensitivity of Derived Visibility to Measured Target Radiance Uncertainty The error is reasonably low, however it is largest when the target range is close to the visibility. For this reason, it is important to keep measurement uncertainties as low as possible, so that we may safely use targets at range close to visibility, where the inherent contrast sensitivity is least. 5. Sensitivity of Derived Visibility to Measured Target Radiance Uncertainty when the llorizon Radiant Signal is 100 This is a similar plot to the above, however it shows the effect when the auto-iris is set to keep the horizon near 100 rather than 200. Errors are significantly higher, illustrating the advantage 10 setting horizon signals near 200. Sensitivity of Derived Visibility to Measured Ilorizon Radiance Uncertainty Like ploi 4, the errors are reasonably small, but become larger when the target range is close to the visibility. 7. Sensitivity of Derived Visibility to Measured Ilorizon Radiance Uncertainty when the llorizon Radiant Signal is 100 The error is significantly larger than when the horizon radiant signal is near 200, as in the above plot. Again, it is advantageous to keep the horizon signal near 200. 8. Sensitivity of Derived Visibility to Non-Linearities in Camera Response This shows the impact of the slight non-linearities normally encountered with CID sensors. Correction is straight-forward and computationally fast. Note that the error is roughly independent of apparent contrast; thus it would not appear as a discrepancy between different targets, but rather as an overall bias on the visibility determined by all of the targets. 9. Sensitivity of Derived Visibility to Non-Linearities in Camera Response for a llorizon Radiant Signal of 100 The error is somewhat changed, and the change is camera-dependent. 10. Sensitivity of Derived Visibility to Non-Linearities in Camera Response for a llorizon Radiant Signal of 220 Above a signal of 200, the CID cameras sometimes become increasingly non- uniform. In this plot, one of the four cameras shows this effect, with a significantly larger error when the horizon radiance climbs above 200. 11. Sensitivity of Derived Visibility to Precision and Stability of the Non- linearity Curve I shows the error associated with our ability to precisely measure the non- linearity. Curve 2 shows the impact of a change in the linear response of a camera over a year interval. Curve 3 shows the impact of the non-linearity after a year if no correction for the initial (pre-year) response is made. The change is large, because the camera had an abnormally high full dark level, as well as a truncated high end signal. Whereas correction of the non-linearity is important, this slide illustrates that monitoring of the camera response, through occasional full dark checks, is also important. 12. Summary Plot Showing the Co impact, the impact of target radiance change, horizon radiance change, and non-linearity. As noted earlier, we wish to minimize the instrumental sources of uncertainty, so that targets near the left end of the scale, i. e. with ranges close to the visibility, may be used. In this way the impact of changing Co may be minimized. IU 13. Improvements Relating to Measurement Accuracy a) The impact of some sources of error is minimized by using horizon radiant signals near 200. b) Applying the lincarily can largely remove one source of CITOL. c) Monitoring the full dark signal becomes important if the linearily correction is used. d) We need to evaluate the various sources of measurement uncertainty, to determine whether random noise and non-uniformity need to be improved or corrected. e) Having targets at a greater number of ranges allows us to stay within the optimal Cr range. f) Verifying target range is important, because an x% error in range translates directly into an x% error in the visibility determination. WW 1 g) In the future, taking advantage of the continually growing sensor technology through the use of more stable, noise free cameras should allow us to further minimize measurement error, and allow us to utilize the measurement regimes more favorable to minimizing the Co impacts. uy. 1. Improvements Relating to Non-ideal Conditions a) Limiting targets to Co close to .8 minimized most of the errors evaluated in this study. b) When it is not possible to use optimally black targets, using more restrictive V/R limits could be helpful. c) An automated check for clouds on the horizon, based on comparing signal and STD within two horizon ROI's could help avoid or detect error in the horizon radiance. (In this case, the error is actually an error in the assumption that the horizon radiance equals equilibrium radiance.) d) We need to determine whether the lens allows sufficient change in elevation 10 introduce significant error in the horizon radiance if a higher ROI is used. e) We would like to study the potential for determining, from the imagery, more exact Co values. f) Characterization of the diurnal variations in Co may also be useful, if they are sufficiently well-behaved. 15. Conclusions A Sensitivity Study of Daytime Visibility Determination with the Horizon Scanning Imager • Sensitivity of Measured and Input Parameters • Potential System Improvements Sensitivity to: • Variations in Inherent Contrast • Measured Target Radiance • Measured Horizon Radiance • Camera Responsivity (Linearity) • Range and Threshold Determination of Visibility from Measured Contrast ε = Contrast threshold Co = inherent contrast Cr = apparent contrast 100.00 80.00 Test 10 Sensitivity of Derived Visibility to a variation in actual co when a fixed input Coof.8 is used 60.00 40.00 % Change in Derived Visibility .9 20.00 8 0.00 - boce 0 OC -20.00 - -40.00 + 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast 40.00 . . . . . . . 20.00 Test 10 Sensitivity of Derived Visibility to a variation in actual Co when a - fixed input coof.5 is used 0.00 - -20.00 - - - - - - - . . % Change in Derived Visibility . P R O -40.00 - - - - - 000.00 - - - - - - - -60.00 - .. -80.00 ... . ...... .... . . . . . . . . . . . ! -100.00 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast . . . - - . . -. - . . . . .. -. - _ - . - . - 100.00 80.00 Test 2b Sensitivity of Derived Visibility to Measured Target Radiance Uncertainty CO = .8 L o = 200 . - . - 60.00 - - • . . . - 40.00 - . • - - - -- % Change in Derived Visibility 20.00 .4 .2 -- 0.00 -O- * - - +2 ... .. . ... - -20.00 - . - - - - - - . . . . . . . . . -40.00 + 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast . 100.00 80.00 Test 2a Sensitivity of Derived Visibility to Measured Target Radiance Unceriainty CO = .8 L9 = 100 60.00 . ... . 11. . . . . . . . . . 40.00 - % Change in Derived Visibility 20.00 .2 0.00 o Reparer e pappagesepsitena -20.0 -40.00 T 0.00 R 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast 100.00 - - -- - - Test 36 Sensitivity of Derived Visibility to - Measured Horizon Radiance Uncertainty Co = .8 L T = 2 0 0 80.00 60.00 - . . . . . . . . . . - - - . 40.00 % Change in Derived Visibility 20.00 : 47:29:6 ingen semasa . . . . . . 0.00 NON + -20.00 -40.00 0.00 0.10 0.20 0.30 0.40 0.50 0.60 -- --- -- Measured Apparent Contrast ...-..- - -- - -- - ...- .-.-.--- - -. - .. -..... --.. - ---- -- - - - 100.00 ----- -- --.. 80.00 Test 3a Sensitivity of Derived Visibility to Measured Horizon Radiance Uncertainty CO = .8 L9 = 100 . -- -- -- .... -- -- ... - . - . 60.00 . ..... . ... . .. - - .. 40.00 - % Change in Derived Visibility 20.00 +2 :::::::::: 0.00 . . . . -20.00 .. . . - - - - - - - - -40.0 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast - - - 80.00 - ... .. .. .. 60.00 Test 4b Sensitivity of Derived Visibility to Non-Linearities in Camera Response Co = .8 L T = 200 . .. ... . .. .. . .. .. .. .. .. . 40.00 . . NMM 1 = LINO20 2 = LINO24 = LINO28 = LINO32 20.00 - % Change in Derived Visibility 0.00 - w : O . - -20.00 - . ... ... . - 40.00 - ... - - - - ... . .. ... . -60.00 7 0.00 - 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast - - 80.00 Test 4a Sensitivity of Derived Visibility to Non-Linearities in Camera Response CO = .8 L o = 100 60.00 40.00 i = LINO20 = LINC24 = LINO28 = LINC32 -- - Mt - - 20.00 % Change in Derived Visibility *** 0.00 com wworznoxo Nr.9%..?. we -20.00 was. -40.00 - - -- -- - -60.00 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast 80.00 60.00 Test 4C Sensitivity of Derived Visibility to Non-Linearities in Camera Response CO = .8 L o = 220 - - 40.00 1 = LINC20 = LINO24 = LINO28 = LINC32 - AWN- 20.00 - % Change in Derived Visibility 0.00 tannonoscow.com ogromn.w. M ---. -20.00 . . . . . . . . . -40.00 -. . -60.00 .- 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast 100.00 . . . . . . . . . Test 5b Sensitivity of Derived Visibility to Precision and Stability of Non-Linearity Co = .8 L Q = 200 80.00 60.00 1 = Precision; LINO245 vs LINO24 2 = Stability; LIN042 vs LINO 24 3 = Stability; LIN042 vs linear response -.-.-.- . 40.00 .-.- % Change in Derived Visibility -.-.- 20.00 0.00 - -- - -- - - - -- -20.00 + -40.00 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent Contrast .. Loc 100.00 - - - - - - - - Summary a for Co = .8 L T = 200 . . - . 80.00 . . . . . .. . .. . - - . - . - -- 0 = 0 error = co change = + .1 2 = Target radiance change = + 4 counts 3 = Horizon radiance change = + 4 counts 4 = Non-linearity change - 60.00 -- -.- . . - . - - - - . -. . - .. - .- - . . . - . - - - . - . . 40.00 . % Change in Derived Visibility 20.00 COLOSSUS 0.00 MOGO .. ... -20.00 .... ..... ..................... -40.00 0.00 0.10 0.20 0.30 0.40 0.50 0.60 Measured Apparent contrast Improvements Relating to Measurement Accuracy • Control Horizon Radiance Near 200 Apply linearity Monitor full dark signal Evaluate magnitude of noise, non-uniformity Increase Number of Targets • Verify Target Range • In future: more stable, noise free camera Improvements Relating to Non-ideal Conditions Limit targets to Co close to .8 Use different V/R limits for non-ideal targets • Automated Cloud-check in Horizon ROI • • • Determine acceptable Horizon ROI elevations Potentially Extract Co values Potentially characterize Co Diurnal Variations Conclusions Sensitivity study reveals potential sources of observed visibility inconsistencies Results in potential for significant system improvements Annotations of Viewgraphs, for presentation on Determination of Visibility at Night Using the Horizon Scanning Imager Note: The following outline is intended as summary only. For further details, see Technical memos AV91-012t and AV91-013t. Title and Outline We will derive the defining equations for visibility at night, then discuss measurements techniques based on relative measurements and absolute measurements. The resulting equations are derived. Engineering considerations involved in the measurements are then discussed. 2. Derivation of Transmissometer Equation for Night Visibility a) Equation for illuminance of a point source of intensity I at distance r, in absence of attenuation b) The above with attenuation, in terms of beam transmittance and transmissivity (which is transmittance over unit distance). c) Allard's law, which comes from defining night vis as the distance at which the above illuminance reaches a threshold. Relation determined by NBS, and used commonly with transmissometers. This is based on the experimental determination that in the above equation, the threshold is itself dependent on the visibility. O 3. Derivation of Koschmieder Equation for Daytime Visibility The derivation begins with defining equations for apparent radiance and for contrast. (A discussion of the physical basis for these equations may be appropriate.) An expression for apparent contrast is derived: From this form of the equation, one shows that if a black target is used, and the background is the horizon sky, and the target is large enough so that the visual contrast threshold is close to a constant, the human detection of threshold contrast for a target at range r becomes a measure of the state of the atmosphere. That is, Cr becomes a function of transmissivity in these limited conditions. 4. Day Visibility vs Night Visibility A summary of the resulting equations, and their usage Fig 1. Day Visibility and Night Visibility as a function of Extinction Coefficient 5. 6. Fig 2. Night Visibility as a function of Day Visibility Defining Equations and Relation to Measured Values At this point, the left side of this display, the defining equations, have been derived. The next part of the discussion will derive the equations which may be used to determine the visibility from measurements of either absolute or relative radiance of sources at night. These are shown on the right side of the display, along with the equation which has been used for determination of visibility from measurements of apparent contrast in the daytime HSI. 8. IISI Visibility Equations Based on Absolute Measurement of Lights at Night These equations are for extended sources, i.e. sources larger than a pixel. Analogous point source equations are contained in Memo AV91-012t. a) The transmittance is determined directly from measured target radiance. b) The transmissivity derived from the transmittance. This may be used with Allard's law or the NBS relation, to derive night visibility (the technique is iterative, see slide 2; slide numbers are listed on annotation) c) The resulting equation if the measured transmittance is used with Koschmieder's Law. This is derived on the next display. 9. Derivation of the Koschmieder Visibility using Absolute Measurements a) The Koschmieder equation from slide 3, expressed in terms of visibility. b) The measured transmissivity, from slide 8 c) Substituting b into a yields the desired equation 10. IIŞI Visibility Equations Based on Relative Measurement of Lights at Night These equations are for extended sources, i.e. sources larger than a pixel. Analogous point source equations are contained in Memo AV91-0121. a) The measured relative radiance is proportional to the beam transmittance and the relative inherent radiance of the source. b) If two measurements are ratioed, they are then a function of the transmissivity, the relative range to the targets, and the relative inherent radiances (which must be known or determined). c) Transmissivity may then be determined directly, and used iteratively with Allard's law or the NBS relation. d) It may also be used with the Koschmieder Law, as derived on the next slide. Derivation of the Koschmieder Visibility using Relative Measurements a) As before, the Koschmieder equation from slide 3. b) The transmissivity determined from relative measurements, from slide 10. (Physically the same transmissivity as from slides 8 and 9, but determined in a different way.) c) Substituting b into a yields the desired equation. A review of slide 7 at this point may be appropriate. Engineering Considerations These are several considerations involved in accurate determinations of the absolute or relative target radiance. Normal considerations we have dealt with in the daytime HSI usage are not reiterated. a) Point Source Implications: In the absence of the point spread function of the lens, which represents defocussing by the lens, measurement may not be reliable due to variations in sensitivity over the pixel area. A slight defocussing of the lens may become desirable, if the point spread function is not already sufficiently large. By thus spreading the flux over more than one pixel, the central pixel measure should become reliable. We also need to verify that our time averaging is sufficient to average out any effects of turbulence. And, for sources which are sub-pixel, we may need to make a correction for forward scattering which can partially fill the remainder of the pixel. b) Flux control algorithms need to be designed, i.e. what is the best integration period to use. This may be a constant for a given source, once night lights come on. More than one integration period may be required for a given scene with lights of different illuminance. The auto iris is operable during the day. At night, we intend to automatically disable the auto iris, leaving it fully open, so that it will not close down in response to the increasing signal caused by increasing integration periods. 13. Engineering Considerations, Continued a) It is necessary to calibrate the effect of using integration periods as opposed to the normal live video. This is discussed more in slides 14 and 15. Cooling may improve signal to noise, and therefore improve achieved sensitivity. Depending on whether thermal noise is indeed the major source of the system noise, this may be an option to keep in mind. c) Determining target range and intensity is not so much a theoretical issue, as a job to be done. At least a first estimate of target range may be obtained by taking measurements at dusk when both the lights and the normal daytime targets may be imaged. The intensity of the lights may be most easily determined from measurements on a very clear night. 14. Measured Signal vs Integration Period Calibration This plot shows the measured signal as a function of integration period or number of frames of injection inhibit used. The several curves were acquired at different lamp positions. [The numbers identifying the curves represent the relative lamp position in logs; 1.0 lamp positions has .1 log less light impinging on the sensor than the 1.1 log position.] Note that the response is highly non-linear. Due to the asymptotic behavior of the curves at lower flux levels, we do not achieve as much gain with injection inhibit as expected. We are working (as of May 91) to determine the cause of this behavior and whether it can be mitigated. 5. Measured Signal vs Flux Level in Logs at several integration periods The curve farthest to the right was acquired at integration 0, which is equivalent to the normal live video calibration curve. These curves, and those in slide 14, must be measured in order to determine the relative apparent target radiance from measured signals at known integration periods. Determination of Visibility at Night Using the Horizon Scanning Imager • Visibility Definitions Relative vs Absolute Techniques • Resulting Equations Engineering Considerations Derivation of Transmissometer Equation E = 1/2 No Attenuation Attenuating Atmosphere E = Tr 1/72 = T'I/2 E4 = ITV/V2 S = 1 T IV? Allard's Law NBS Derivation of Koschmieder Equation the = tho T, +N, * Apparent Radiance tho-blo Contrast blo Co = image T . 1. CA = Co The If pho = plr = Lq and Co = 1 C, = T, = T' At threshhold Cr = ε, range = visibility E = TV Day Visibility E = TV Koschmieder w E = .05 WMO, Scatter Meters E = .055 Transmissometers Night Visibility S=ITV/V Transmissometers E = TV European; scatters meters? 100.00 10.00 Visibility (mi) Day Night 1.007 0.107 0.01 0.10 1.00 10.00 100.00 Extinction coefficient (1/mi) Fig. 1 Day Visibility and Night Visibility as a Function of Extinction Coefficient -- -- Night Visibility (mi) 0.10 0.10 1.00 - 10.00 - 100.00 .... .. .. . . - - . - - . - -..--- - .... ...... 1.00 Fig. 2 Night Visibility as a Function of Day Visibility Day Visibility (mi) 10.00 100.00 ... . - . ----... --. Defining Equations Relation to Measured Values meas Cr Kosch ε= TV/ meas abs. -V=rin ε / In Kosch) mas absv=rine v V = (1-2) In 1 meas rel. Allard Et = ITV/V2 E4 = ITV/V2 BYT = (h/to)" (NBS meas rel. To IIN 2No]"C= meas abs. meas re L2N, NJ HSI Visibility Equations Based on Absolute Measurement of Lights at Night Tr = thr/tho Measurement T = (tLr/tLo) !! Allard > = In > Koschmieder Analogous Point Source Equations Absolute; Koschmieder E = TV Koschmieder :: V = In ε/InT T=(L/ L ) Absolute Measurement .:. INT="" ( Thus V=r In ε / In Determination of Visibility HSI Visibility Equations Based on Relative Measurement of Lights at Night Nr a T No Measurement Ratio Allard V = (4-2) ine/n ( Koschmieder Analogous Point Source Equations Relative; Koschmieder E = TV Koschmieder .:: V = In ε/InT Relative Measurement Thus V = (1-12)e In / in my mind of Visibility Determination Engineering Considerations • Point Source Implications • Point Spread Function • Turbulence • Forward Scattering • Flux Control Auto Iris • Integration Calibration • Cooling Improved Sensitivity Signal to Noise Impacts • Determining Targets Range Intensity -- - - -- - - - 1.5 1.4 1.3 240.00 - 1.2 Measured Signal vs Integration Period Calibration (at Several Relative flux levels) 180.00 1.1 Measured Signal (0-255 bytes) 120.00 60.00 - 0.00 + 0.00 20.00 40.00 60.00 80.00 100.00 Integration Period (# Frames) Signal vs Flux Level in Logs at several integration intervals - 260.00 220.00 - 100V 50/20/10 int=0 180.00 Signal 140.00 + 100.00 60.00 20.00 0.00 0.40 0.80 1.20 1.60 Relative Flux (log)