The G.E.E.P. development project at Geomatrix Earth Science

Geomatrix is partnering The University of Leicester in a KTP project, the aim of which is to further develop the MSP, or newly re-named GEEP system (Geophysical Exploration Equipment Platform) into a fully commercial product which can be operated by a third party of either a sale or rental basis, the KTP project comes to a conclusion in December 2006, with the GEEP available for commercial surveys by 3rd parties by mid 2006

GEEP Data sheet (360 kb pdf)

Geomatrix' partners in the FIESTA project were Glebe Mines, The University of Leicester, British Geological Survey, Geosoft who were all under the project management of MIRO and which came to a conclusion in 2003.

Geomatrix main involvement in this project was supplier of field geophysical instrumentation and data acquisition software, and to aid University of Leicester in the concept and building of the Multi-Sensor Platform or GEEP. The underlying principals of the GEEP are similar to that or an airborne platform, which allow multiple data sets to be acquired simultaneously and rapidly with as little degradation of signal quality as is possible. This allows large areas of land to be surveyed at a resolution that would not be otherwise possible and in a time that is far less than when acquiring the data by conventional techniques. Applications include Archaeological prospection, industrial mineral exploration and large area site investigation and environmental surveys.

A resume of the GEEP its applications and limitations associated with it are to be found below.

Survey techniques

The GEEP allows surveys to be conducted quickly and efficiently. By using a DGPS for location there is no need to set up a local coordinate survey grid, as all locations can be tied back to national grid coordinates. The platform can be towed across survey areas in an almost random pattern, although thought must be given to the position of survey lines to avoid special aliasing of data.
The survey technique that was found to be the most efficient in the field was to drive around the perimeter of the survey areas in ‘ever decreasing circles’. This allowed the platform to be towed continuously without worrying about tight corners etc. Then as the loops became tighter switch the survey pattern to parallel lines across the survey area. This allowed a complete, dense data coverage with multiple tie lines.

Instrumentation

Geophysical instruments that have been mounted onto the GEEP during the FIESTA project include Geometrics G858 Caesium magnetometers, Geonics EM38, Geonics EM31, Exploranium GR256 and an OMNI VLF system.

GEEP Data flow

Combining multiple instrument data files

Although the GEEP is collecting several different geophysical data sets, due to the nature of the data storage these data sets are not completely processed separately. As some processes would have to be performed repeatedly on each individual data set if they were processed on their own, such as a positional correction, it is more efficient to combine all of the data sets into one large data file.
As the data sets are being collected simultaneously, the common factor between each data set is the time. Thus the way to produce a single data file is to combine the data records that were collected at the same time. This is not a trivial procedure.
How does the user know the time that a reading has been taken?
There are few instruments that fix a time stamp to the serial output string. Even when this is the case, the time stamp will only refer to the individual instrument’s internal clock, and have no relation to data collected on another instrument at the same time.
In addition to this, instruments produce an output signal at different rates, for example some at ten times a second – others only once a second. Other instruments only output a signal at approximate rates. In a 1 second period an instrument may produce ten readings, in another 1 second period there may only be 8 readings output.
More complication is added with the fact that the instruments are not ‘synchronised’ – two instruments may output data once a second; the first instrument may output a signal at the second, but the second instrument may be producing a signal half a second behind the first instrument.
The problem is how to combine signals such as these that may be out of phase or irregularly timed and have no common time identifier.

The key to combining the different data sets with the GEEP system lies with the MagLog logging software. MagLog collects a transmitted data string from the WLAN signal and adds a computer time stamp to the end of the string at the time the data was received. Each data type is then saved in a separate file. The time stamp is derived from the time of the logging computer, which may not be correct relative to GMT. In this manner the time stamp is independent of the recording instruments, but it is common to all of them.

MagLog removes the problem of knowing when an item of data was recorded – but it does not remove the other problems mentioned concerning sampling rates. One way of doing this is to determine the sample rate of the output file, and combine the input data files to match that rate from a set starting time.
The instrument records are unlikely to lie exactly on the sample interval for the reasons mentioned above, so a search window is employed to take the nearest recorded value for the sample time to ensure that the records that are combined are as close as possible.
If the sample rate of the combined file is set to that of the most frequently sampled instrument, then the less frequent sampled instruments will be missing data in the combined file records. Whereas if the sample rate is at the same rate as the slowest sampled instrument then every record in the combined file will be filled.

Although the computer time field is the key field with regard to combining the data sets together, the other common factor that the individual data files share is location. All of the data points recorded need a location on the ground, and this is derived from the DGPS signal. As there must be a location for each data point the DGPS data can be interpolated across the missing records if the sample rate of the combined file is greater than that of the DGPS input file. If this is the case then the DGPS data is linearly interpolated across missing values. The slowest rate of data output from the GBXPro instrument is one reading every second, so assuming good DGPS fixes, the worst interpolation errors are going to occur over a one second period. This interpolation is carried out for the DGPS values and not the other data sets because in any one-second period the DGPS generally moves in a straight line between two points. The same prediction cannot be confidently or accurately made for the data from other geophysical instruments.

Due to the complexities of the transfer of the serial data from the instrument to the logging machine, the output serial string from some instruments can become distorted. This can involve omitting characters from the string, or adding extra characters in the middle of the string. These records can be easily removed from the instrument data files, as they do not follow the set format of the instrument record.

I
University of Leicester Interpol

Combining each of the single geophysical data files with the DGPS file so that the instrument data can be associated with a position is a lengthy and time-consuming procedure. For this reason the Interpol software was written.
Interpol allows the serial data files to be combined at a set sampling interval. The user, depending on the frequency of data output by the instruments, can determine this interval. For example if all the instruments sample at ten times a second, then the Interpol output file can be sampled at anything from ten times a second to once a second. If the instruments only output data every second, nothing extra will be gained from producing an output file at a greater sampling interval than this.
If an instrument samples at a rate less than that set for the output file, then the records in the output file for that instrument will appear as the ‘null data’ identifier, an ‘*’.
The output file produced by the Interpol software is a set format, so that the instrument data will appear in the same column regardless of what other instruments are being used.
This allows import routines to other software packages, such as Oasis Montaj to be easily created.

Geosoft Oasis Montaj

Oasis Montaj is being used as the main piece of software for data processing the GEEP data. Data is imported from the Interpol program into a large database for each field area. As there are many Interpol files to import – generally one for each field, an import template was created. In addition to increasing the speed at which data was imported this allowed the database to be set up the same for each field area.

Some of the processes carried out within Oasis Montaj include:

Performing the projection transformation from WGS84 to British National Grid coordinates
Removing drop outs (zero values) and spikes from the magnetic data
Subtracting the base station magnetometer data from the field magnetometer values (diurnal correction)
Applying a height correction for the DGPS data
Applying positional corrections for the port and starboard magnetic sensors
Applying heading corrections to the magnetic data
Leveling magnetic data with regard different base station positions
Applying the positional corrections for the EM38.
Filtering and smoothing of data

DGPS data

The DGPS antenna is situated at the front end of the GEEP, 1.40m above the ground. The console unit is located on the towing vehicle in an attempt to remove all unnecessary magnetic material from the platform. The signal is transmitted from the antenna to the console via coaxial cable within the tow cable.
From early work on the accuracy and repeatability of the GBXPro DGPS data an accuracy of  0.4m has been recorded. With the introduction of another differential beacon station at Wormleighton since these findings, the signal to noise ration and signal strength of the differential signal has been increased. The positional accuracy investigations have not been repeated in depth since the introduction of the new beacon, but the data quality, and reliability appear to have increased with its introduction.

The DGPS height data is not so well constrained as the XY positional data. This is due to the position of the satellites in three-dimensional space around the sensor.
A position in the XY plane can be derived well from 4 or more satellite signals to the object as the satellites are likely to be well spaced around the object in the XY planes of 3D space. To accurately derive a Z position of an object on the surface of the Earth from orbiting satellites the position is not going to be so well constrained. This is because the satellites are all essentially the same large distance away from the object, with the XY distance differences becoming small in comparison. This problem is compounded by the fact that the satellites that the positional fix is derived from keep changing. As the number of satellites changes in the data often a ‘jump’ in the height data can be clearly seen. This affect is reduced as the number of satellite signals received increases. Data repeatability is thus dependant on the number of satellites.

Processing of the DGPS data involved the removal of data points that were derived from using less that 4 satellites and a differential signal latency of greater than 12 seconds. In addition the data had to be corrected for the height of the antenna above the ground.

Compass data

The fluxgate compass is mounted at the front of the platform along with the majority of the other electronic components of the system. The compass is mounted in a foam block as a means of damping out vibration noise.
The data recorded by the fluxgate compass is a very important component of the GEEP system. The data is used not only to derive the position of the instruments in relation to the DGPS antenna, but also gives an additional method of checking the quality of the DGPS positions. Initial trials with the compass found that the instrument was very sensitive and large, unstable variations in the heading, pitch and roll of the instrument were recorded at the slightest movement of the platform. Vibrations were also being transmitted from the towing vehicle along the tow cable – causing the deck to vibrate. Many attempts were made to reduce the noise that was being recorded, including increasing the rigidity of the platform ‘deck’ and cushioning the instrument with foam. The method that was found to be the most successful was the internal digital damping process as the instrument recorded data. The damping was set so that the instrument internally used a rolling average over 20 records (2 seconds), this was found to leave the sensitivity high enough to detect the motion of the sledge without being affected by the vibrations.

Little other processing was needed on the compass data, except the removal of bad data strings, which was done automatically using the Interpol program.

Since the conclusion of the FIESTA project, Geomatrix has been involved in further developing the GEEP as part of a MIST award (Minerals Industry Sustainable Technology) along with the University of Leicester, and is actively looking to further develop the system in applications such as Archaeological prospection where large tracts of land need to be rapidly surveyed.

GEEP acquired magnetometer data at Wroxeter Roman City

The main applications for the GEEP include mineral exploration, environmental and brown field investigations and archaeological investigations. For further information on the use of geophysics in archaeology, the Journal Archaeological Prospection may be considered, details can be found at www.interscience.wiley.com/journal/arp

page modified 25 September 2006