Ground Penetrating Radar
GRORADAR™ by Gary R. Olhoeft, PhD
[ Home ] [ Up ] [ Tutorial ] [ Links ] [ Reading ] [ Bibliography ] [ Educational Software ] [ GRORADAR Ground Penetrating Radar Gary R. Olhoeft ]

Download Adobe PDF format manual

GRORADARÔ

Version 03.99

Preliminary Manual

Copyright Ó 1997-1999 by Gary R. Olhoeft.

Golden, CO 80402-1520 USA

GRORADARÔ is a trademark of Gary R. Olhoeft registered in the State of Colorado.

SINGLE END USER LICENSE AGREEMENT

Do not load or use GRORADAR(tm) until you have read the following terms and conditions. By loading or using GRORADAR(tm), you agree to the terms of this software license agreement.

You have a license to copy this software, GRORADAR(tm) onto a single computer, and you may make one backup copy subject to these conditions: 1) you may not copy, modify, rent, sell, distribute or transfer any part of GRORADAR(tm) and agree to prevent unauthorized copying (you may copy and distribute the accompanying data and the version of GRORADAR(tm) named GRO_DEMO.EXE that is for demonstration only). 2) you may not sublicense nor permit simultaneous use by more than one user. 3) if the main copy resides on a desktop or tower computer, the backup copy may reside in a notebook or laptop computer. 4) you may not reverse engineer, decompile nor disassemble GRORADAR(tm).

Title and ownership of all copies of GRORADAR(tm) remain with Gary R. Olhoeft. GRORADAR(tm) is trademarked and copyrighted, and is protected by the laws of the United States and other countries through international treaties. You may not remove copyright nor trademark notices from GRORADAR(tm). Gary R. Olhoeft may make changes to GRORADAR(tm) at any time and is not obligated to support or update all copies of GRORADAR(tm) beyond the first year after purchase.

This software is provided "as is" without any express or implied warranty of any kind including warranties of merchantability, noninfringement, fitness for a particular purpose, or accuracy, except for a 30-day money back guarantee if GRORADAR(tm) is incompatible with and will not execute on your computer. In no event shall Gary R. Olhoeft be liable for any damages whatsoever (including without limitation, lost profits, business interruption, or lost information), arising out of the use or inability to use GRORADAR(tm). Some jurisdictions prohibit exclusion or limitation of liability for implied warranties or consequential or incidental damages, so this may not apply to you. Defective media should be returned for replacement.

Gary R. Olhoeft may terminate this agreement at any time if you violate its terms. Upon termination, you will immediately return or destroy all copies of GRORADAR(tm). Claims under this agreement shall be governed by the laws of the State of Colorado and remedy is limited to the refund of the purchase price less shipping and handling costs.

RESTRICTED RIGHTS NOTICE

Use, duplication, or disclosure by the Government is subject to
restrictions as set forth in FAR52.227-14 and DFAR252.227-7013 et seq.
or their successors. Use of GRORADAR(tm) by the government constitutes
acknowledgement of Gary R. Olhoeft's proprietary rights. The contractor
manufacturer is Gary R. Olhoeft, P.O.Box 1520, Golden, CO 80402-1520
Copyright does not extend to include fraction of code derived
from USGS reports 92-532, and 95-58 (see GRORADAR.NOT).

 FAST START

To begin using the licensed, registered version of this program, type:

GRORADAR

and press Enter (or Return on some keyboards). Either the distribution CD-ROM must be in the CD drive or GRORADARÔ must be registered to a specific hard drive (as on a laptop computer without a CD drive). Without one of these requirements being met, GRORADARÔ will not run.

For help, type:

GRORADAR -?

and press Enter. At most points in the program, press ? for a suggestion of what to do next, or F1 for more general help.

For the demonstration program, type:

GRO_DEMO

and press Enter. GRO_DEMO has limited RAM availability and cannot produce any outputs, but it may be freely copied and distributed. GRO_DEMO does not require the presence of the CD-ROM nor any specific hard drive registration.

TECHNICAL SUPPORT

Licensed registered users of GRORADARÔ may report bugs or anomalies in the program to the contact point below. Bugs will be fixed. Please report bugs using the bug report form in the appendix. Requests for new features will be seriously considered. Licensed registered users will be given free upgrades for one year after purchase.

Contacts:

Gary R. Olhoeft, P.O. Box 1520, Golden, CO 80402-1520 USA,
golhoeft@g-p-r.com, (303) 273-9202 fax, (303) 279-7932 voice.

INTRODUCTION

This program will run on IntelÒ 80386 machines and higher. It is optimized for IntelÒ PentiumÔ and PentiumÒ II processors with 16 to 64 MB RAM (or more) and VESA VBE 2.0 or higher compatible graphics video cards. If the graphics card is not VESA compatible, use UNIVBE as a driver (available from SciTech Software, 916-894-8400, www.scitechsoft.com).

GRORADARÔ must be run under MS-DOSÔ , and works best from the DOS under MicrosoftÒ Windows 95Ô , found when F8 is used to boot to MS-DOSÒ during Win98 startup, or under IBMÒ PC-DOS 7. It has been tested and found to run under MS-DOSÔ 3.3x under 80386 CPU's, but runs very slowly, an 80486 is better, but a PentiumÔ or PentiumÒ II is recommended. It will run in the MS-DOSÔ box under Windows 95Ô (using up to 4 GB RAM but slower than under DOS), but it will not run under Windows NTÔ . It has its own built in 32-bit DOS-Extender to access up to 4 GB RAM, perform 32-bit processing under DOS and uses the DPMI extender interface in a DOS box under Windows 95Ô . NOTE that the demonstration version of GRORADARÔ (called GRO_DEMO) can not produce output and is limited to 4 MB RAM.. GRORADARÔ is not guaranteed to run on all possible combinations of computer hardware, operating systems, and software.

If in a 1024x768x256-color graphics mode, press P at any time to make a file for hardcopy. In image modes, the file will be encapsulated Adobe PostScriptÔ (.EPS), and in vector line graphs, the file will be HP-GLÔ (.PLT). Options will allow changing the size of the paper, modifying margins, adding labels, rotating the plot, making the file directly LaserJetÔ compatible (just copy the file to the printer port connected to a PostScript LaserJet), and more. For non-PostScript™ printers, use  Alladin Enterprises' Ghostscript, which can put PostScriptÔ files out onto other devices (http://www.cs.wisc.edu/~ghost). Other programs such as Birmy Graphics Corporation (407-768-6766, www.birmy.com) PowerRIP PostScript interprerters may also be used to drive other printers and non-PostScript devices. NOTE that the demonstration version of GRORADARÔ can not produce output.

GRORADARÔ is extensively corrected, modified, extended (1996-1998), and partially derived from code in (no copyright claimed for fraction of code from the following):

Powers, M.H. and Olhoeft, G.R., 1995, GPRMODV2: One-Dimensional Full Waveform Forward Modeling of Dispersive Ground Penetrating Radar Data Version 2.0: USGS Open File Report 95-58, 41p. + floppy disk.

Powers, M.H., Duke, S.K., Huffman III, A.C. and Olhoeft, G.R., 1992, GPRMODEL: one-dimensional full waveform forward modeling of ground penetrating radar data: USGS Open File Report 92-532, 22p. + floppy.

Duke, S.K., 1990, Calibration of ground penetrating radar and calculation of attenuation and dielectric permittivity versus depth: MSc Thesis, Dept. of Geophysics, Colorado School of Mines, Golden, 236 p.

Additional information may be found in the program documentation, GRORADAR.NOT, and in the README.TXT file.

Example: GSSI SIR-10 Data of a metallic pipe

After starting the program by typing GRORADAR and pressing Enter, there will be several introductory screens, ultimately arriving at the MAIN MENU. During the introductory screens, there will be an opportunity to change the graphics mode and hardcopy parameters. This may be done now or later. There will also be an opportunity to see an explanation of how the Debye-Pellat single relaxation equation describes frequency dependent dispersion caused by finite velocity charge motion (and ultimately, through the Cole-Cole equation, relates to water content).

From the MAIN MENU, press F for Field data display and import. A box will appear with *.dzt, *.dt1 and *.raw. Using the arrow and delete keys, delete all but *.dzt and press Enter. Another box appears containing a directory of example data files with the extension .dzt. Use the down arrow key to move the highlight to the file PIPE.DZT and press Enter.

A box appears giving the header information from the field data file. Press Escape to return to the MAIN MENU or Enter to continue and read the file into memory. Then, another box appears, asking "If markers are found in the data, are they to be used in rubber sheeting?" Markers may appear in GPR data for many reasons. One of the most common uses for markers is to locate the data to real physical locations marked by flags and location surveyed against a benchmark. Answer by pressing Y.

Another question appears in the box, "What is the marker spacing in meters?" If the actual (x,y,z) locations of the markers are known or measured, they may be read into this program from an ASCII file by entering -1. Or if the simple traverse distance between flags is known, enter that. In this case, the ground is flat and the flags are 1 meter apart, so type 1 (the number one) and press Enter.

Another box appears containing the information from the GPR data header file. Press Enter and another screen appears. Across the bottom of the screen are instructions, with further help available by pressing F1 or ? to get a suggestion of what to do next. The screen also shows a grey scale image of the radar cross section and vector line drawing of the wavelet under the cursor.

Using the down arrow key, move the cursor down 36 points until it hits the first energy being recorded. At the lower right of the graph, there are three labels showing the vertical cursor location in scan sample points, two-way travel time in nanoseconds, and depth in meters using the relative dielectric permittivity shown just above (in this case, it shows k=4.0). Press the 0 (zero) key to reset time zero to this position. Time zero is a function of cable lengths, setup parameters, antenna configurations, and not known to the radar system directly. Press F1 for a help screen showing all the options available at this point. Press any key until the radar data image reappears.

If in a 1024x768x256-color graphics mode, press P at any time to make a file for hardcopy. In image modes, the file will be encapsulated Adobe PostScriptÔ (.EPS), and in vector line graphs, the file will be HP-GLÔ (.PLT). Options will allow changing the size of the paper, rotating the plot, making the file LaserJetÔ compatible, and more.  Pressing P now in this example produces the left side of the the front cover of this manual labeled BEFORE(copying the file GPR_BEF.EPS to a LaserJet or other PostScript compatible device will produce hardcopy of this image). NOTE that the demonstration version of GRORADARÔ , GRO_DEMO can not produce output.

Note the banding in the radar data image caused by mismatch in coupling to the soil. Press B to remove the average background and this banding. Note this also removes the first arriving energy that tells where zero is located.

Note the vertical bands in the data caused by radio transmitter noise at horizontal traverse positions of 5, 81, 132, 152, and 183 scans. The horizontal scan count is shown just below the image. Press Z and type 20, Enter, to use a median cross spot filter to remove this RF noise. Now press S to see the histogram of pixel values, and type 5, Enter to perform an image processing histogram equalization contrast stretch. The histogram of the data before and after stretching will appear in place of the image. Press Enter a second time to see the stretched image data.

Now press G, and the wavelet plot will be replaced by the range gain applied during data acquisition. Press Enter or the right arrow key to move the red circle cursor to the second range gain point (and simultaneous move to the second value across the bottom of the screen). Type 25 and press Enter to change this range gain value. In a similar manner, change the next to values to 50, the next two to 60, the next to 70, and the last to 80. Across the bottom of the screen should now read 10 (hilighted red), 25, 50, 50, 60, 60, 70, 80. Now press Shift-Enter to re-gain the image data. Now press Escape once.

Press R to rubber sheet the horizontal scan traverse locations to the known flag marker locations. In the process of the spline rubber sheeting, notice several black vertical bars appear in the image. This is missing data that can now be filled in by interpolation of nearest data by pressing I. Notice also, the labeling of scan location across the bottom of the screen is now in meters horizontal traverse, scans across the screen, and scans in the data file..

Now move the cursors to the top of the hyperbola at horizontal scan 106 (file scan 122) 2.78 meters traverse, and vertical sample 130 points, 4.6 ns, and 0.34 m depth, by using the arrow keys (alt-arrow to move faster). Press V and a yellow hyperbola will be overlain on the image. Across the bottom of the screen, the relative dielectric permittivity and the size of the target scatterer may now be changed, causing the hyperbola shape to change, and to try achieving a better match to the hyperbola in the image. Use the up and down arrow keys to change the permittivity and watch the hyperbola change. Put the permittivity value back to 4.0. Press the right arrow key to move the red hilight to the target radius of 0.01 meters. Using the up arrow and page up keys, increase this value from 0.01 to 0.45 and watch the hyperbola change shape and size. A red ellipse will indicate the size of the target scatterer. The scatterer is actually a circle, but plots as an ellipse due to vertical exaggeration in the image. This is calibrating the two-way travel time to depth to the top of the hyperbola. The green line at the bottom of the image is the amplitude of the radar data along the hyperbola (similar to seismic AVO), press A to toggle this amplitude plot on and off. At the top of the image, the region above the horizontal green line is in the near field of the antenna. As this program uses far field assumptions, the region between the green line and the surface can not be processed nor modeled accurately. Press Escape once. The right half of the front cover of this manual, labelled AFTER (copying the file GPR_AFT.EPS to a LaserJet or other PostScript compatible device will produce hardcopy of this image), is what is recorded in a *.EPS file now by pressing P if in a 1024x768x256-color graphics mode.

Press H, and watch the progress indicator run across the bottom of the screen. When it reaches the right side, an image processing mask approximating a migration will have been completed. Press S, type 5, press Enter, press Enter again to improve the contrast, and the results will show the collapse or focusing of the hyperbola to its vertex (migration). The file GPR_MIGR.EPS copied to a LaserJet or other PostScript compatible device shows the hardcopy of this migrated image.

Press Enter a third time, and return to the main menu. Press X, and the original, unprocessed raw data scan under the cursor location in the image will be imported into the 1D full waveform modeling program, including the hyperbola fitted estimation of the relative dielectric permittivity and depth. The red curve is the field scan, and the yellow curve is a full waveform model computed from parameters that are listed on the right side of the screen. The pulse polarity looks reversed, so press R (not all antennas and cables are connected for the same polarity). Amplitudes are too high, so press C to enter the electrical conductivity parameter menu (on the right side of the screen) and press or hold down the Home key until the conductivity reaches 40.00 mS/m. Note the increasing decay of the yellow curve from left to right with increasing conduction losses. In these parameter menus, use the right-left arrows to move between parameters, and the up-down arrow, home-end, and page-up-down keys to change values. Press F1 for additional help. Press S to turn multiples off and watch how the model changes, and press S again to turn them back on.

Press E to return to the dielectric parameter menu (pressing M goes to the magnetic permeability menu). Press the right arrow key four times until the hilight moves to the 0.34m depth. Press the page up key 3 times to increase 0.34 to 0.37 and watch the yellow model curve move to the right. Use the left arrow key to return to the first dielectric parameter of 4.0. The four parameters across here are the Cole-Cole equation parameters describing frequency dependence of the complex dielectric permittivity (if you return to the Main Menu and press I to see the introductory screens again, you will see these Cole-Cole equations and the assumptions of the model). Note that the field data (red curve) near 5 and 6 nanoseconds has two negative going peaks of different amplitude. This is an indication of dispersion, or frequency dependence in electromagnetic properties.

Use the arrow keys to make these parameter values read 4.2, 3.6, 0.1, and 1.0. While changing the values, note the impact on the position, amplitude and shape of the yellow curve, and on its fit to the red curve field data. It will now be neccesary to return and lower the conductivity to match the overall amplitudes better. This has now improved the depth interpretation and derived material properties information for the soil between the surface and the buried object (in this case a 90 cm diameter metal pipe buried 0.37 m deep to the top of the pipe). This is not a unique answer, as other combinations of parameters may produce an equivalent fit. Press P now to make a hardcopy of the screen through an HP-GL file. The figure below was produced by importing this file (GPR_FIG2.PLT) into Microsoft WordÔ version 7.0 for Windows 95Ô :

Press F to see the frequency power spectra of the wavelet (on the bottom third of the screen) and the frequency dependence of the material properties. Press F2 and enter a soil mixing model. Press the capital letter to increase a value, and a lower case letter to decrease a value using the first letter of each component (sand, clay, water, iron). It will create the Cole-Cole parameters for that soil mixture. In this case, attempting to fit the pipe data shows only a fraction of a percent of water is required to give this small amount of dispersion. Press Escape and then shift-Enter to return to the model. Press Z to save the model wiggle trace (it will appear as a white wiggle trace). Now vary some of the parameters to see how the white model compares with the yellow model generated by parameter variation. Return the model parameters to those producing the best fit to the data.

Press shift-Enter to return to the Main Menu. Press S to save the model in a file, type a filename like "pipe" and press Enter. If a box appears asking Overwrite? that means the filename already exists. Press N and Enter to enter a new filename or Y and Enter to overwrite the existing one. When such a model file exists, on entering the program pressing R will read the model and data parameters and go directly into the full waveform modeling part of the program.

Press F to reimport the field data and continue as above to the data image (Yes to markers, 1 m spacing). The yellow curve in the wavelet window is the 1D full waveform model just created. Move the horizontal cursor to the scan 122 that was modeled, then move it from side to side to see how features in the data (white wavelet) compare and change against the model (yellow wavelet).

Press Q to quit, and you're ready to start over and process another file. (It is also possible to press F and import another field file at this point but doing this repeatedly will cause the program to crash on some machines as allocable memory is limited.)

After starting the program by typing GRORADAR and pressing Enter, there will be several introductory screens, ultimately arriving at the MAIN MENU.

Press F for Field data display and import. A box will appear with *.dzt, *.dt1 and *.raw. Using the arrow and delete keys, delete all but *.dt1 and press Enter. Another box appears containing a directory of files with the extension .dt1. Use the down arrow key to move the highlight to the file KEVIN003.DT1 and press Enter.

A box appears with the header information stored at the time of data acquisition. Press Enter and another box appears, asking "If markers are found in the data, are they to be used in rubber sheeting?" Markers may appear in GPR data for many reasons. One of the most common uses for markers is to locate the data to real physical locations marked by flags and location surveyed against a benchmark. Answer by pressing Y. This dataset was taken with a marker spacing of 1 meter, so press 1 and Enter in response to the next question.

Another box appears containing the information from the GPR data header file. Press Enter and another screen appears. Across the bottom of the screen are instructions, with further help about instructions available by pressing F1 or ? to get a suggestion of what to do next. There are also a large grey scale image and vector line drawing of the wavelet under the cursor.

Using the down arrow key, move the cursor down 51 points until it hits the first energy being recorded. At the lower right of the graph, there are three labels showing the vertical cursor location in scan sample points, two-way travel time in nanoseconds, and depth in meters using the relative dielectric permittivity shown just above (in this case, it shows k=4.0). Press the 0 (zero) key to reset time zero to this position.

Note the top to bottom gradient in the radar data image. Press B to remove the average background and this gradient. Press S to see the histogram of data values, and press 7 Enter and Enter again to see the contrast stretched image.

Now press G, and the wavelet plot will be replaced by the range gain applied during data acquisition. Press Enter or the right arrow key to move the red circle cursor to the third range gain point (and simultaneous move to highlightin red the third value across the bottom of the screen). Type 5 and press Enter to change this range gain value. Similar change the next to values to 10, 15, 20, 20, 20, and the last to 20. Across the bottom of the screen should now read 0 (hilighted red), 0, 5, 10, 15, 20, 20, 20. Now press Shift-Enter to re-gain the image data. Now press Escape once. Press W to remove the wiggle trace.

Now press R to rubbersheet the image and I to interpolate missing scans. Press PgDn once and then hold down Control-Left arrow until the cursor at the bottom of the image reads 6.00 meters. Move the cursors to the top of the hyperbola at horizontal scan 646 (5.31 meters) and vertical sample 106 points, 11.0 ns, and 0.83 m depth, by using the arrow keys (alt-arrow to move faster). Press V and a yellow hyperbola will be overlain on the image. Across the bottom of the screen, the relative dielectric permittivity and the size of the target scatterer may now be changed, causing the hyperbola shape to change, and to try achieving a better match to the hyperbola in the image. Press the right arrow key to move the red hilight to the target radius of 0.01 meters. Using the up arrow and page up keys, increase this value from 0.01 to 0.60 and watch the hyperbola change shape and size. A red ellipse will appear to indicate the size of the target scatterer. It is actually a circle, but distorted by vertical or horizontal exaggeration. Use the left arrow to move the red hilight back to permittivity. Use the up and down arrow keys and the Home and End keys to change the permittivity and watch the hyperbola change. Put the permittivity value to 8.0. This is calibrating the two-way travel time to depth to the top of the hyperbola. Press Escape once.

Press F10 to change from processing the whole dataset to just the part shown on the screen. Press H, and watch the progress indicator run across the bottom of the screen. When it reaches the right side, an image processing mask approximating a migration will have been completed. Press S, type 2, press Enter, press Enter again to improve the contrast, and the results will show the collapse or focusing of the hyperbola to its vertex.

The lower right side of the image is now has too much gain, so press F10 then W and G to get the gain window back. Then change the last three values from 20 to 15, press Shift-Enter to apply, Esc and W to return to the full screen image. Press V again and control left arrow until the cursor is on 637 scans, 5.24 meters. Press right arrow and lower the target radius to 0.5 meters by pressing down arrow once. Note the fit of the top of the red ellipse to the curvature of the pattern in the image after migration, confirming the size of the object. Note also the second reflection at the bottom of the red ellipse. This could be a multiple or a reflection from the bottom of a non-metallic object.

Press Escape once and Enter to return to the main menu. Press X, and the original, unprocessed raw data scan under the cursor location in the image will be imported into the 1D full waveform modeling program, including the hyperbola fitted estimation of relative dielectric permittivity and depth. The red curve is the field scan, and the yellow curve is a full waveform model computed from parameters on the right side of the screen. The model shown assumes a metallic reflector and is not a good fit to the data. Press shift-Enter to return to the main menu.

Press R to read in a modeling file. A box with *.gpm will appear. Press Enter, and in the next box select KEVIN003.GPM and press Enter. Press F1 for additional help. Press S to turn multiples off, and press S again to turn them back on. Note how this model of an air-filled plastic pipe fits the data better. This modeling program begins assuming the second layer or target is made from metal. But this is actually a double walled plastic culvert. Press shift-Enter to return to main menu. Press M to go to model build/edit. Change number of layers to 7. Return to the model and see how well you can fit the data.

Press Q to quit, and you're ready to start over and process another file.

 Example: Sensors & Software PulseEkko Data with survey wheel

After starting the program by typing GRORADAR and pressing Enter, there will be several introductory screens, ultimately arriving at the MAIN MENU.

Press F for Field data display and import. A box will appear with *.dzt, *.dt1 and *.raw. Using the arrow and delete keys, delete all but *.dt1 and press Enter. Another box appears containing a directory of files with the extension .dt1. Use the down arrow key to move the highlight to the file KEVIN006.DT1 and press Enter.

A box appears with the header information stored at the time of data acquisition. Press Enter and another box appears, asking "If markers are found in the data, are they to be used in rubber sheeting?" Markers may appear in GPR data for many reasons. One of the most common uses for markers is to locate the data to real physical locations marked by flags and location surveyed against a benchmark. Answer by pressing N. This dataset was taken with a survey wheel and requires no rubber sheeting.

Another box appears containing the information from the GPR data header file. Press Enter and another screen appears. Across the bottom of the screen are instructions, with further help about instructions available by pressing F1 or ? to get a suggestion of what to do next. There are also a large grey scale image and vector line drawing of the wavelet under the cursor.

Using the down arrow key, move the cursor down 102 points until it hits the first energy being recorded. At the lower right of the graph, there are three labels showing the vertical cursor location in scan sample points, two-way travel time in nanoseconds, and depth in meters using the relative dielectric permittivity shown just above (in this case, it shows k=4.0). Press the 0 (zero) key to reset time zero to this position.

Note the top to bottom gradient in the radar data image. Press B to remove the average background and this gradient.

Now press G, and the wavelet plot will be replaced by the range gain applied during data acquisition. Press Enter or the right arrow key to move the red circle cursor to the second range gain point (and simultaneous move to the second value across the bottom of the screen). Type 10 and press Enter to change this range gain value. Similar change the next to values to 20, 30, 40, 50, 60, and the last to 70. Across the bottom of the screen should now read 0 (hilighted red), 10, 20, 30, 40, 50, 60, 70. Now press Shift-Enter to re-gain the image data. Now press Escape once.

Now move the cursors to the top of the hyperbola at horizontal scan 98 and vertical sample 220 points, 11.8 ns, and 0.89 m depth, by using the arrow keys (alt-arrow to move faster). Press V and a blue hyperbola will be overlain on the image. Across the bottom of the screen, the relative dielectric permittivity and the size of the target scatterer may now be changed, causing the hyperbola shape to change, and to try achieving a better match to the hyperbola in the image. Use the up and down arrow keys to change the permittivity and watch the hyperbola change. Put the permittivity value to 7.0. Press the right arrow key to move the red hilight to the target radius of 0.01 meters. Using the up arrow and page up keys, increase this value from 0.01 to 0.50 and watch the hyperbola change shape and size. This is calibrating the two-way travel time to depth to the top of the hyperbola. Press Escape once.

Press H, and watch the progress indicator run across the bottom of the screen. When it reaches the right side, an image processing mask approximating a migration will have been completed. Press S, type 10, press Enter, press Enter again to improve the contrast, and the results will show the collapse or focusing of the hyperbola to its vertex.

Press Enter a third time, and return to the main menu. Press Enter or X, and the original, unprocessed raw data scan under the cursor location in the image will be imported into the 1D full waveform modeling program, including the hyperbola fitted estimated relative dielectric permittivity and depth. The red curve is the field scan, and the yellow curve is a full waveform model computed from parameters on the right side of the screen. Press shift-Enter to return to the main menu. The data header file said this was an antenna with 450-MHz center frequency, so press T, type 450, press Enter, and press Enter again to return to the model. The pulse polarity looks reversed, so press R. Amplitudes are too high, so press C to enter the electrical conductivity parameter menu (on the right side of the screen) and press or hold down the Home key until the conductivity reaches 2.00 mS/m. Note the increasing decay of the yellow curve from left to right with increasing conduction losses. In these parameter menus, use the right-left arrows to move between parameters, and the up-down arrow, home-end, and page-up-down keys to change values. Press F1 for additional help. Press S to turn multiples off, and press S again to turn them back on.

Press E to return to the dielectric parameter menu (pressing M goes to the magnetic permeability menu). Press the up arrow key until the first dielectric parameter of 4.0 become 7.9. Watch the yellow model curve move into alignment with the red field data curve. The four parameters across here are the Cole-Cole equation parameters describing frequency dependence of the complex dielectric permittivity. Note that the field data (red curve) near 12 to 15 nanoseconds has two negative going peaks of nearly equal amplitude. This is an indication that there is little or no dispersion, or frequency dependence present in electromagnetic properties.

This modeling program begins assuming the second layer or target is made from metal. But this is actually a double walled plastic culvert. Press shift-Enter to return to main menu. Press M to go to model build/edit. Change number of layers to 7.

Press Q to quit, and you're ready to start over and process another file.

After starting the program by typing GRORADAR and pressing Enter, there will be several introductory screens, ultimately arriving at the MAIN MENU. During the introductory screens, there will be an opportunity to change the graphics mode and hardcopy parameters. This may be done now or later.

From the MAIN MENU, press F for Field data display and import. A box will appear with *.dzt, *.dt1 and *.raw. Using the arrow and delete keys, delete all but *.raw and press Enter. Another box appears containing a directory of files with the extension .raw. Use the down arrow key to move the highlight to the file 54EA048.RAW and press Enter. A box will appear saying the data type is not SIR10 nor PulseEKKO; press Enter.

Another box will appear, giving the required information to read in an arbitrary data file format. For this example, use the defaults and just press Enter.

Another screen appears, labeled TIME CALIBRATION. Old GSSI SIR-3, SIR-7 and SIR-8 systems recorded a 10-ns time calibrator and used the record to calibrate the two-way travel time range. Depending upon the quality of the data, there are several possible methods to extract this time calibration. The current screen shows three: using zero crossings, local minima, and derivative minima. Press Enter, and a fourth appears. The top half of the screen shows the real and imaginary (red & blue curves) parts of the FFT of the incoming time calibration record (see bottom half below). The white curve is a cursor that may be moved with the right, left and alt arrow keys, but it is initially located at the maxima in the FFT. The numbers along the axis are the two-way range in ns. The yellow numbers at the top and bottom of this top half of the screen show the number of data points per 10 ns and its conversion to two-way travel time range for the maximum and minimum in the FFT. The numbers labeled cursor are those for the position of the cursor.

In the bottom half of this screen, the white curve is the raw time calibration data, showing a 10-ns periodic pulse train. The red curve is the reverse FFT of the top half, filtered narrowly around the cursor position. The up, down and alt arrow keys change the phase, allowing the red zero crossings to align with the white data zero crossings (zero is in the middle of each plot vertically). The control right and left arrow keys move the yellow vertical line cursors in the lower half of the screen until they align with the white data zero crossings. Hold down the down arrow key until the red zero crossings align with the white zero crossings in the lower half of the screen. Now hold down the control-right arrow until the yellow cursors roughly align with the white data. This is a reasonable fit, giving a two-way travel time range of 90.5 ns time calibration. Press Enter, and a summary text screen appears. You must select and type in the two-way travel time range. Simply pressing Enter will accept either of two defaults: the value recorded with the data, or the value determined by the position of the cursor in the previous FFT screen (the case in the example), or another value may be typed followed by Enter. For this example, the correct answer is 90.5 ns, so simply press Enter.

Now the range GAIN FUNCTION estimation screens appear. These derive a formula from the digitized analog recording of the range gain function. In this plot, the cyan lines are the range gain recording and the purple lines are the fitted envelope function. Press any key, and the linear and log plots of time varying (range gain) appear versus sample number.

Press Enter, and a box appears giving the header information from the field data file. Press Escape to return to the MAIN MENU or Enter to continue and read the file into memory. At this point a box may appear saying a marker file was found. Press R to read in the marker locations in scan space. Then, another box appears, asking "If markers are found in the data, are they to be used in rubber sheeting?" Markers may appear in GPR data for many reasons. One of the most common uses for markers is to locate the data to real physical locations marked by flags and location surveyed against a benchmark. Answer by pressing Y.

Another question appears in the box, "What is the marker spacing in meters?" If the actual (x,y,z) locations of the markers are known or measured, they may be entered via an ASCII file by entering -1. Or if the simple traverse distance between flags is known, enter that. In this case, type 1 and press Enter.

Another box appears containing the information from the GPR data header file. Press Enter and another screen appears. Across the bottom of the screen are instructions, with further help available by pressing F1 or ? to get a suggestion of what to do next. The screen also shows a large grey scale image and vector line drawing of the wavelet under the cursor.

Using the down arrow key, move the cursor down 62 points until it hits the first energy being recorded. At the lower right of the graph, there are three labels showing the vertical cursor location in scan sample points, two-way travel time in nanoseconds, and depth in meters using the relative dielectric permittivity shown just above (in this case, it shows k=4.0). Press the 0 (zero) key to reset time zero to this position.

Note the slight banding in the radar data image. Press B to remove the average background and this banding. Now press S and type 5, Enter to perform an image processing histogram equalization contrast stretch. The histogram of the data before and after stretching will appear in place of the image. Press Enter a second time to see the stretched image data.

Press R to rubber sheet the horizontal scan traverse locations to the known flag marker locations. In the process of the spline rubber sheeting, notice several black vertical bars appear in the image. This is missing data that can now be filled in by interpolation of nearest data by pressing I. Notice also, the labeling of scan location across the bottom of the screen is now in meters horizontal traverse.

Press W to toggle on and off the appearance of the wavelet wiggle plot. Use the page up and page down keys to rapidly scroll through the data.

Now move the cursors to horizontal scan 555 (file scan 929) 2.49 meters traverse, and vertical sample 232 points, 30.1 ns, and 1.01 m depth, by using the arrow keys (alt-arrow to move faster). Press V and a blue hyperbola will be overlain on the image. Across the bottom of the screen, the relative dielectric permittivity and the size of the target scatterer may now be changed, causing the hyperbola shape to change, and to try achieving a better match to the hyperbola in the image. Use the up and down arrow keys to change the permittivity and watch the hyperbola change. Put the permittivity value to 20.0. Note the fit to the diffraction events off the ends of the layer reflectors. Press Escape once.

Note at the bottom of the screen, pressing the F10 key will toggle further processing to be over the entire radar dataset or just over the part that is visible on the screen. Use just the screen part to run through "what if?" scenarios of processing quickly, then process the entire dataset for the final product. Press H, and watch the progress indicator run across the bottom of the screen. When it reaches the right side, an image processing mask approximating a migration will have been completed. Press S, type 10, press Enter, press Enter again to improve the contrast, and the results will show the collapse or focusing of the hyperbola to its vertex.

Press Enter a third time, and return to the main menu. Press Enter or X, and the original, unprocessed raw data scan under the cursor location in the image will be imported into the 1D full waveform modeling program, including the hyperbola fitted estimated relative dielectric permittivity and depth. The red curve is the field scan, and the yellow curve is a full waveform model computed from parameters that are listed on the right side of the screen. Proceed as in the first example for GSSI SIR-2/10 data.

Press Q to quit, and you're ready to start over and process another file. (It is also possible to press F and import another field file at this point.)

Perform this same sequence on the file 06WSN300.DZT, reading in the EDM coordinate file, but after pressing V, press T for topographic correction and F for fit to screen to get the image shown on the next page (GRO_FIG3.EPS). With the long axis of the page oriented left-right in landscape mode, the upper right corner of the figure shows the file name and, below the filename, a series of letters that give the processing history. Try to duplicate it.

Commands, Operations, Options

At the DOS prompt, type GRORADAR to start.

Options are

GRORADAR -? to get graphics help.

GRORADAR -g257 will start in VESA 640x480x256-color mode

GRORADAR -g259 will start in VESA 800x600x256-color mode

GRORADAR -g261 will start in VESA 1024x768x256-color mode

(1600x1200x256-color mode coming)

A series of information screens follows startup. At the screen titled, Relaxation Models of Dispersion, pressing F1 takes you into an explanation of charge motion, frequency dependence, and dispersion (keep pressing space to continue). Pressing any other key takes you past this to a screen, Nomenclature, that allows changing the graphics mode (press 1, 2, 3, or 4), and toggling of the graphics hardcopy mode (press L) and rotation (press R). Plots that are LaserJet compatible will copy and print directly on HP PCL-compatible laser printers. Plots that are not rotated, print in portrait mode (like the plots in this manual), and plots that are rotated print in landcape mode (rotated 90^o counter clockwise). Pressing space twice takes you to the Main Menu.

At the Main Menu, pressing the following keys perform the indicated action:

F allows importation of a field data set, processing and display, asks for path of filename to data, and then shows menu from which to select the data file. See the next page.

T allows selection of the transmitter center frequency (default 500 MHz)

X computes the 1D full waveform forward model and overlays on a data scan using the full waveform modeling program developed in Duke (1990)

D allows the display parameters to be altered:

B     beginning time on model plot

E ending time

O zero offset time

S sample interval

N number of samples in the 1D full waveform model after making changes, press Esc to cancel the changes, or Shift-Enter to accept the change and return to the Main Menu.

C changes the coupling ratio between the antenna and the ground

G changes the range gain applied to the data and model

R reads in a pre-existing model file

S saves the model parameters to a file

M allows building and editing of a model parameter list

H changes the heading for hardcopy plots

I re-displays the introductory sccreens

A acquires data through a Computer Boards PCM-DAS16D/16 digitizer card

Q quits or exits the program

F Field Data Display and Import Menu

Pressing F at the Main Menu asks for the path to the field data file, brings up a directory of files at that location, and allows selection of the file. After a file is highlighted and Enter is pressed, a series of information screens appear which vary with the data type:

DZT GSSI data shows the header file, and then asks

If markers are found in the data, are they to be used in rubbersheeting?

If the answer is NO, the next question is

How many meters traverse per scan?

If this answer is zero, the markers are ignored and assumed to be for annotation only.

If the answer is YES, the next question is

If the answer is a negative number (less than zero), the program searches for an ASCII file containing survey location data in the format: x.xxxxx y.yyyyy z.zzzzz 20charstring with the filename prefix the same as the data plus a .EDM extension. If this file cannot be found, it again asks for the marker spacing in meters.

A summary header file is then displayed, and Enter will read in the data or Escape will return to the Main Menu.

DT1 Sensors & Software data <>

RAW digitized data <>

There should now be a screen displayed that looks like this:

<>

A variety of processing, modeling and display options are available at this point. Press F1 for help, ? for a suggestion, Escape to return to the Main Menu, Enter to return to the Main Menu while importing the scan under the cursor for 1D full waveform modeling.

Field Menu sorted by keys:

F1 displays help screens.

F5 grab calibration scan reflection from steel plate for pavement thickness calibration

F9 grab data scan and automatically model for pavement thickness after F5 calibrate

F10 toggles the processing of only what is visible on the screen display or the entire data set.

F12 restores the original raw data if there is enough computer memory.

Arrow keys move cursor up/down/left/right.

Alt + arrows move faster.

Control + arrows scroll data slowly under cursor.

Page up/page down keys scroll the data left and right rapidly under cursor.

Home goes to beginning of data.

End goes to end of data.

| or \ search and find markers.

[ tracks a wavelet layer horizontally from current cursor position to the right.

< and > change the size of the wavelet tracker window.

] automatically model pavement thickness for n scans and give statistics of accuracy after F5 cal

0 resets the position of time zero to the current cursor position vertically.

A adds a mark at the cursor position.

B removes the average background scan (B also toggles it back).

C scrolls through various color palette options.

D deletes the mark under the cursor.

E slects and applies a variety of edge detection codes (Sobel, Roberts, extended, etc.).

F fits all the data onto one screen display.

G displays and allows editing of the Range Gain if the wavelet display is present.

H does hyperbola mask migration.

I applies a range image texture enhancement.

K displays 8-channel GSSI multipolarization data.

L applies a Laplacian unsharp masking and edge crisper.

M displays marker location information.

N displays marker name at cursor location if cursor is on a mark.

O (oh) outputs the data scan under the cursor to a separate binary file.

P creates a publication quality Encapsulated PostScript EPS hardcopy file of screen.

R does lateral (horizontal) rubber sheeting to constant scan spacing.

S displays a histogram of the data and allows a contrast stretch to be performed.

T performs a topographic correction to the data if location information is available.

U applies a cross spot remover noise filter.

V starts interactive hyperbola velocity matching if lateral spacing between scans is known.

W toggles wavelet display on/off.

X applies an X-spot remover noise filter.

Y applies an isotime histogram equalization contrast stretch.

Z selects and applies a variety of filters, such as the median spot remover noise filter (good at removing cellular phone noise).

eXecute Model Menu

Press X in the Main Menu to enter the full waveform modeling menu. If modeling involves comparison with real field data, first go through the Field menu to select a scan to model. Inside the full waveform model, the following keys are active:

F1 help.

Shift-Enter returns to Main Menu.

Right and Left arrows move highlight to different parameters in properties versus depth table on right side of screen. On the numeric keypad, the highlighted variable value is changed by ± 1, ± 0.1 or ± 0.01 by using, respectively the Home/End, up arrow/down arrow, and the PgUp/PgDn keys, and ± 0.001 using the + and - keys.

C displays a conductivity versus depth profile, E displays dielectric permittivity, and M displays magnetic permeability. Both permittivity and permeability are presented normalized to free space and as complex quantities obeying the Cole-Cole equation for a distribution of relaxation times.

D saves the field scan and Z saves the model scan for later use or comparison as parameters change.

F switches to a display of the frequency dependence of the layer properties selected along with the power spectra for the field data. While in the frequency dependence display, F2 brings in a mixing formula to estimate electrical and magnetic properties from volume or weight (toggled by V) mixtures of Sand, Clay, Water, Iron, and Air. Pressing the upper case first letter increases a component and the lower case decreases it.

Pressing P in any screen display makes an HP-GL file for hardcopy output.

APPENDIX Files on distribution CD-ROM:

\GRORADAR directory:

EQNTRPLX.FON font file required by GRORADAR
GRORADAR.EXE program, start by GRORADAR ? to get help
GRORADAR.EXE only on licensed, registered CD-ROM
(uses 32-bit DOS-EXTENDER, requires 386/486/Pentium/Pro/II)
(during Win95 startup, use F8 to exit WINDOWS 95 to DOS before running)(will use up to 4 gigabytes of memory)
(will run under Win95 DOS box but may only use 16 megabytes if machine has more than 64 megabytes total RAM)
<< NOTE: theoretically can use 4 gigabytes of RAM, but limited by BIOS implementations >>
(will NOT run under WinNT DOS box: must dual boot to DOS)

GRO_DEMO.EXE (demo version cannot make outputs and limited to 4 megabytes)
GRORADAR.NOT program notes
GRORADAR.DOC Microsoft Word 97 format manual
GRORADAR.HTML manual
GRORADAR.RTL manual
GRORADAR.PDF manual
README.TXT (this file)

*****************************************************************************
NOTE: The hardcopy output from this program is produced as HP-GL or Adobe
PostScript files that are suitable for copying to LaserJet printers or import into various programs. To print PostScript to other printers, use a translator like PowerRIP from Birmy Graphics Corp. (www.birmy.com), or QMS's UltraScript, or the Alladin Enterprises' GhostScript to send the files to Epson, Canon, HP, Encad or other inkjet printers.
*****************************************************************************

Example radar data files included on CD-ROM version of GRORADAR:
*.DT1 are Sensors & Software PulseEkko format GPR data
*.HD are headers for *.DT1
*.DZT are Geophysical Survey Systems Inc. SIR-2/10 format data
*.RAW are GSSI SIR-7/8/3 raw digitized GPR data files
*.TC are GSSI SIR-7/8/3 time calibration files
*.RG are GSSI SIR-7/8/3 range gain calibration files
*.EDM are laser theodolite EDM ASCII location files
*.MRK are mark location files generated by GRORADAR
*.GPM are model parameter files generated by GRORADAR
*.EPS are Encapsulated PostScript files generated by GRORADAR
*.PLT are HPGL plotter files generated by GRORADAR
*.GPR/.2DP are Powers (1995) PhD thesis synthetic GPR data/parameter files
 

Example on cover of manual:
Yuma Proving Ground buried steel pipe
Data acquired by U.S. Geological Survey and published in:
Powers, M.H. and Olhoeft, G.R., 1995, GPRMODV2: One-Dimensional Full Waveform Forward Modeling of  Dispersive Ground Penetrating Radar Data Version 2.0: USGS Open File Report 95-58, 41p. + floppy disk.
Olhoeft, G. R., Powers, M. H. and Capron, D. E., 1994, Buried object detection with ground penetrating radar: in Proc. of Unexploded Ordnance (UXO) detection and range remediation conference, Golden, CO, May 17-19, 1994, p. 207-233.
PIPE.DZT Yuma Proving Ground Steel pipe, GSSI SIR-10A+ data
PIPE.GPM marks 1-m apart
GPR_BEF.EPS cover figure before printable to HP LaserJet with PostScript
GPR_AFT.EPS cover figure after background, median filter, contrast stretch
GPR_MIGR.EPS migrated GPR_AFT
GPR_FIG2.PLT figure 2 printable to HP LaserJet (HP-GL)
J06F04.DZT near-field buried metal & plastic land mines, and empty holes

Example with survey wheel versus continuous time acquisition:
St. Kevin Gulch, Colorado acid mine drainage site plastic culvert
Data acquired by Colorado School of Mines Dept of Geophysics Field Camp, 1997
KEVIN003.DT1
KEVIN003.HD marks 1-m apart Sensors & Software PulseEkko 1000
KEVIN006.DT1
KEVIN006.HD survey wheel

Example of arbitrary raw data file:
Borden PCE DNAPL Spill Site
Data acquired by U. S. Geological Survey and published in:
Sander, K.A., Olhoeft, G.R., and Lucius, J.E., 1992, Surface and borehole radar monitoring of a DNAPL spill in 3D versus frequency, look angle and time, in Bell, R.S., ed, Oakbrook, Il: Society of Engineering and Mineral Exploration Geophysics, Golden, CO, p. 455-469.
Sander, K. A. and Olhoeft, G. R., 1994, 500-MHz ground penetrating radar data collected during an intentional spill of
tetrachloroethylene at Canadian Forces Base Borden in 1991: U.S. Geological Survey Digital Data Series DDS-25, CD-ROM.
Brewster, M. L., Annan, A. P., Greenhouse, J. P., Kueper, B. H., Olhoeft, G. R., Redman, J. D. and Sander, K. A., 1995, Observed migration of a controlled DNAPL release by geophysical methods: Ground Water, v. 33, p. 977-987.
54EA048.RAW GSSI SIR-7 data recorded on FM analog tape and later digitized
54EA048.MRK marks 1-m apart
54EA048.RG
54EA048.TC

Example of topographic correction:
Magic Mountain Archaeological Site
Data acquired by Colorado School of Mines Dept of Geophysics Field Camp 1997
06WSN225.DT1 Sensor & Software PulseEkko 1000
06WSN225.EDM
06WSN225.HD
06WSN300.DZT GSSI SIR-8 modified by Gary Olhoeft for digital data acquisition
06WSN300.EDM
06WSN300.MRK
06WSN300.RG
06WSN300.TC
GPR_FIG3.EPS topo corrected image printable to HP LaserJet with PostScript

Synthetic radar data using Powers (1995) PhD modeling program:
Powers, M. H., 1995, Dispersive ground penetrating radar modeling in 2D: PhD thesis T-4820, Department of Geophysics, Colorado School of Mines, Golden, CO, 198p.
(see Figures 43 & 44 and accompanying discussion in thesis)
DPIPES01.GPR three 1-m radius pipes buried in lossy & dispersive media
DPIPES01.2DP modeling data (markers at 10 m spacing in both files)
DPIPES02.GPR three 1-m radius pipes buried in lossless media
DPIPES02.2DP

\BORDEN5 directory raw and processed data from the 1991 9-m cell DNAPL spill
\BORDEN5
\RAW 500-MHz 8-bit data files use GRORADAR(tm) to access
\RANGEGAI range gain files
\TIME_CAL time calibration files
\BOREHOLE 200-MHz hole-to-hole radar data
\LAB_DATA lab measurements on core samples (use EM_MODEL to view)
\SURVEY EDM (laser theodolite) survey data for 9-m cell surface
\IMG processed data in 4D (3D vs time)(use BORDEN4D to view)
(BORDEN4D requires 80486, 64 MB RAM, 1024x768x256 VESA graphics to run)

\GRAPHICS directory useful information to improve graphics performance

\HERSHEY directory fonts used by software on this CD-ROM

\GPR98 directory Gary Olhoeft's papers from GPR98

This web site is supported by sales of GRORADAR™ data acquisition, processing, modeling and display software.
Send mail to golhoeft@g-p-r.com with questions or comments about this web site.
Copyright © 1998-2000 Gary R. Olhoeft. All rights reserved.  Last modified: June 01, 2000.
All trademarks are owned by their respective companies. Copyright & Fair Use  Electronic Frontier Foundation
Please don't copy and redistribute anything without requesting permission.
Do freely make hyperlinks to these web pages through http://www.g-p-r.com.