ModelVision software, Research and Services
Tensor Research Newsletter
ModelVision 18 - December 2024
​A lot has been happening at Tensor Research since our last newsletter, so we want to tell you about what will be in ModelVision 18.0 and wish you all the best for the festive season and 2025.
Contents
Highlights from ModelVision 18.0
Beta program and incentives for you to join
Highlights from ModelVision 18.0
Visualisation enhancements
have improved our imaging of new structural symbols for points, geophysical grids, continuous/discrete colour legends with range and increment controls, colour contouring and real-time updates with many Apply buttons added to property dialogs. The common multi-map crosshair is a great help when working with different image and body maps. Use it easily with the linked pan and zoom feature.
Tiled and scale-linked windows showing the new muti-map crosshair cursor, legends and colour and size modulated symbols.
More on Point Visualisation, AutoMag and QuickDepth
Here we show some examples using flexible data point annotation, AutoMag solution visualisation and QuickDepth symbols.
In the first example of point annotation, the symbols sizes and colour selections would not normally be used in this way, but they do demonstrate the level of control. Importing geochemistry data or rock property measurements can also be used in this way.​
Point data are displayed as colour and size modulated symbols with colour = depth below ground, diameter = depth quality estimate. The title box in the left corner is designed to show the names of the channels and their relative locations. Annotations clockwise from left side are mauve for magnetic susceptibility (cgs), black for formation width (m) and white for depth below ground (m). Font size, decimal points, position and orientation are controlled for each annotation. The top image is RTP with RTP contours. The bottom image is a geo-registered JPG bitmap overlain by RTP contours with the point symbols as the top layer.
​​In the left-hand AutoMag example below, the conventional Smart points representing strike and dip directions are colour-coded by the Level number as the width of the sample data expands by a factor of 2. The right-hand image is using points generated via the Standard Points button on the AutoMag toolbar.
Left - AutoMag solutions as smart points with symbols showing tabular body strike and dip directions.
Right – Smart points converted to standard points using the trend direction as an arrow scaled by trend confidence.
QuickDepth-specific symbols have been included in this update along with other minor improvements. A further update for QuickDepth users is expected after release 18.0 is finalised. The example below shows how specific symbols have been used to have some resemblance to the body type (sheet, pipe, ellipsoid, edge) selected by the interpreter.
QuickDepth results from the tutorial training set based on a complete set of supported body types.
(a) 3D perspective view of the model and survey lines. (b) QuickDepth symbols with depth annotation.
(c) QuickDepth symbols and depths over the top of the original model. (d) QuickDepth symbols and depth annotations over the normalised source strength (NSS) image.
At last, a full text search Help System
ModelVision’s help system has been modernised to take advantage of recent advances in search engine technology. Our help system, user guide, supporting documentation and tutorial are moving to the web for quick searching for solutions to your specific problem. We have even included an experimental AI Assistant that uses our documentation with a ChatGPT style large language model (LLM). Don’t worry, if the web is not accessible, you will still have access to new versions of the local help system and PDF formatted guides. We have many researchers around the world who undertake fundamental research using the tools provided in ModelVision so it is important that they can reference the User Guide.
Example full text search to find a list of FFT filters where references to line and grid filters were returned.
Genius Search AI Assistant!
The AI Assistant is a welcome addition to the learning process because you can ask it plain English questions and refine them as you go. In the previous example of a full text search, your question is more focussed on a specific outcome.
Here, I asked AI Assistant the question “Can I use FFT to transform total magnetic intensity grids to the magnetic tensor?”. In response, it gives you a summary of what it has found and a link to the most useful page in the documentation. There may be more than one link depending on the nature of the search.
Example of the AI Assistant being used to find references to the transformation of TMI to the magnetic gradient tensor.
New computational engine for the Calculator
Our calculator operates on lines, grids, points and drillholes and now has a completely new and more powerful computational engine. Extended function list, logic branching and scripting allow you to apply reusable code each time you receive a new survey dataset.
The calculator has a new computation engine, extended function set and multi-line scripting.
Coding with the Calculator
Extended function list, logic branching and scripting allow you to build reusable code each time you receive a new survey dataset. This script is used to convert magnetic tensor data to commonly used parameters such as the analytic signal of the Bz component (ASz), invariants I1 and I2, Tilt angle, dimensionality (DI) and SI stands for shape index (Cevallos, 2014).
A multi-line script example applied to tensor data (BFnn) resampled from grids derived from a total magnetic intensity grid using ModelVision’s powerful suite of FFT tensor filters.
When finished, the calculator shows the last line in the script file. If bad syntax is encountered, it will stop at that line and provide an error message. Just update your MVP script text file, reload and Compute. Results of the calculation from one line over two magnetic pipes are shown in a Multi-Track view of 7 computed channels plus the original magnetic data. Note the symmetry of the ASz, Invariant 1 and Invariant 2 channels.
Multi-Track plot of Mag plus 7 computed data channels.
The IF statement allows you to include another cascading IF statement in place of a simple parameter or calculation. For example, say you have three parameters A, B, C, you can craft a multi-tiered expression where either output parameter can be replaced by a logical expression.
​
-
D = IF( A > B, C, A+B)
-
D = IF(A > B, (IF(A = NULL, C, A+B), A+B))
-
D = IF(A > B, (IF(A = NULL, C, A+B), (IF(A=B, C, A+B)))
​
While this is very powerful and efficient to use, it is difficult to spot logic problems and if you find you are having problems, break it out and create intermediate values that are used on the next line.
​
Reference: Cevallos, C. 2014 Automatic generation of 3D geophysical models using curvatures derived from airborne gravity gradient data. Geophysics 79 (5) p. G49-G58.
Create models from data points
For researchers interested in studying large scale models, the point to body conversion tool has been expanded to include tabular, spheres, ellipsoids, circular and elliptical pipes. Use ModelVision survey simulations to test your survey specifications before you fly. It is now easy to create a model of near surface maghaemite nodules using randomly distributed ellipsoids or spheres along with your target model. You can build the full model point data in Excel or other applications such as Python or MATLAB which is much faster than creating the multi-body model manually with the Create Body tool.
Example of a large point to body conversion from the Cloncurry region where the Tabular body type is applied to all points in the set.
Turn data points into models
Test your survey specifications before you fly by simulation. It is now easy to create a model of near surface maghaemite nodules using randomly distributed ellipsoids or spheres along with your target model. You can build the full model point data in Excel or other applications such as Python or MATLAB which is much faster than creating the model manually using the Create Body tool. This example used a random noise generator to create the easting, northing, and elevation. Although not done here, you can also simulate random magnetisation directions and amplitudes. These models are used as a basis for testing different processing techniques and survey specifications for optimum data collection.
Lab measurements of density or magnetic properties can be converted to a located point set and turned into simple spheres that can be overlain on geological and image maps. At any time, you can convert the body to another body type and start modelling. If you have many samples, use the bodies (samples) in the RockMod module to assess their possible lithology classifications by overlaying the lab measurements on one or more of the charts.
A map view and zoomed 3D perspective view of a near-surface noise simulation using randomly distributed ellipsoids to emulate near-surface noise sources with random magnetisation directions immediately above manually positioned circular pipes.
And that’s just some highlights. To find out more about the many other new features and improvements, read more ...
Beta program and incentives for you to join
We are in the final stages of Beta testing so I would like to ask you whether you would like advanced access to the pre-release version of ModelVision 18.0 to do some Beta testing. We value external input from testers because you will use different datasets and solve different problems. We do use ModelVision on many live projects, especially when using RPD Mapping results, or generating research material for model surveys.
Our offer to you is 50% off your annual Support and Updates renewal cost from the time it is next renewed. The offer is available to the first five users who sign up and complete a component of testing. If you work for a company that has multiple licences, the work you do on your licence will be credited to the company at the time the renewal is due.
Timing – We need the results of your tests before the end of January so that we can include as many fixes as possible prior to the release at the end of February.
​
Let us know if you can help (support@tensor-research.com.au) and we will send you details of the beta program along with documentation on the new features, fixes and known bugs.
Release schedule
If we get the results back from our beta testers in time we will include as many recommendations as possible for release on 28 February 2025.
Interesting developments with the magnetic tensor
Our interest in the magnetic gradient tensor started well before the end of the millennium when we were developing tools for the Falcon Gravity gradiometer system. While most of this focus was on building modelling and interpretation tools in ModelVision, De Beers was actively pursuing the development of an acquisition system for the magnetic tensor. Our research had shown us that the vector magnetic tensor contained enough information in just a few readings to identify the direction and location of the magnetic source. We worked with De Beers to develop methods to extract more geological detail, depth of burial, magnetic susceptibility and remanence properties.
​
This research evolved into a service that we call RPD Mapping which stands for Rock Property and Depth mapping. Our data comes from full tensor surveys or well specified conventional total magnetic intensity surveys. Of course, given the opportunity, I would prefer to work with new full tensor survey data, but that alone would never have funded our R&D costs. We had developed a method for transforming a conventional TMI survey into a full magnetic gradient tensor line survey. This meant that we could look at historical TMI surveys anywhere in the world with fresh eyes. What might have been missed in past interpretations?
​
Transformation of the TMI data to the magnetic tensor is well documented and has been available in ModelVision for over 15 years. Our Depth Module uses the magnetic tensor in QuickDepth and the magnetic tensor and tensor grids are generated automatically. We have completed at least 250,000 kilometres of RPD Mapping from surveys in many continents. We know it works because we have had a chance to compare the transform generated tensor with full tensor survey data. The comparison is limited by differing line spacing, line directions and ground clearance but the anomalies exhibit similar amplitudes and shape characteristics. For a perfect match, the data should be collected on the same platform and this will happen soon with developments taking place in Africa.
Conferences and publications
ASEG Exploration Geophysics just published Blair’s third and final PhD paper on the properties of the magnetic tensor.
McKenzie, K.B., Hansen, S.M. and Morrissey, J. (2024) The magnetic gradient tensor of a right circular cylinder: theoretical considerations in the determination of magnetisation direction, Exploration Geophysics, 55:6, 809-841
​
The 2024 SAGA Conference was held in Namibia and three papers were delivered on the magnetic tensor by colleagues. I was there in spirit.
​
Kirkpatrick, L., Vorster, A. and Gibson, L. (2024) Exploration target ranking and environmental risk mitigation using magnetic gradient tensors, multibeam echosounder and optical remotely sensed data. Extended Abstracts - 18th SAGA Biennial Conference & Exhibition 2024, 7 p.
Ugalde, H., Morris, W. and Kamath, A. (2024) Full tensor magnetic gradiometry: an assessment of what is being measured and comparison with total magnetic intensity. Presentation 18th SAGA Biennial Conference & Exhibition 2024, 6 p. Also published as Full-tensor magnetic gradiometry: Comparison with scalar total magnetic intensity, processing and visualization guidelines. EAGE Geophysical Prospecting, 1-12.
Vorster, A., Polome, L. and Pratt, D.A. (2024) Advancing full tensor magnetic gradiometer SQUID and UAV magnetic gradiometer systems and surveys for diamond exploration. Extended Abstracts - 18th SAGA Biennial Conference & Exhibition 2024, 6 p.
​
Hernan Ugalde, DIP Geosciences is our agent in Eastern Canada and South America and provides native language support to South American geophysicists.
​
Professor Emeritus Bill Morris has been a colleague for many years as Bill has a long history in magnetic rock properties, especially magnetic remanence. Bill and Hernan have trained many a student in the art of ModelVision.
​
The Innovation Award was given to Anre Vorster for his work on the full tensor magnetometer over many years. His presentation at the SAGA conference also showcased some of our joint work on aliasing and the importance of full tensor gridding. From my perspective, his presentation on the wing-tip mounted full tensor sensor in a horizontally aligned cryostat represents a major step forward in this technology. A towed bird sensor using the horizontal cryostat has also been developed for a helicopter borne system.
The Air Tractor with the larger stinger housing the horizontal cryostat FTMG SQUID system and the two smaller stingers on the opposite wingtips the G-822A magnetometers. Image from the SAGA extended abstract
The two G-822A total field magnetometers will provide the opportunity for a direct comparison of the transformation of TMI to the magnetic tensor with the measured tensor and direct estimation of the aliasing midway between the flight lines.
​
Time for some end of year good cheer!
Kerri, Blair, Tony and I wish you a very special Christmas with time to unwind from a very busy year. We look forward to hearing from those interested in beta testing and any other questions you have about ModelVision 18.0. We will still be hard at it working on the final release.
David Pratt
Manager Research and Development David.Pratt@tensor-research.com.au
PO Box 5189, Greenwich NSW 2065 Australia
