TS
Tectonics and Structural Geology

Teaching

Minds over Methods: Virtual Microscopy for Geosciences

Minds over Methods: Virtual Microscopy for Geosciences
The next “Minds over Methods” blogpost is a group effort of Liene Spruženiece (left) – postdoctoral researcher at RWTH Aachen and her colleagues Joyce Schmatz, Simon Virgo and Janos L. Urai.

Credit: Liene Spruženiece

The Virtual Microscope is a collaborative project between RWTH Aachen University and Fraunhofer Institute for Applied Information Technology (Schmatz et al., 2010; Virgo et al., 2016).

In the multitude of tools to analyze rocks, the optical microscope is still one of the first go-to methods for rock characterization. A look on a petrographic thin section under an optical microscope gives a quick overview of the mineralogy and fabric of a rock and helps to determine the areas of interest for more detailed analysis with other methods. It is also a crucial part for teaching geology classes.

Most geoscience institutions already have a petrographic microscope that is equipped with a camera and maybe an automated stage for capturing image mosaics. However, these images make only limited use of the vast amount of information available, such as the overview of the sample fabric at different magnifications or the change of optical mineral properties (pleochroism and extinction behavior) with the rotation of polarizers. Although there have been recent developments in the field of virtual microscopy (NASA, 2007; The Open University; 2010; Tetley and Daczko, 2014), to our knowledge, none of the existing systems has been able to fully emulate a user experience in a virtual environment that is equal to using a traditional analogue microscope.

 

Figure 1. The setup of the PetroScan microscope. Credit: Fraunhofer FIT

How does it work?

The idea behind the virtual microscope is to capture the full information of entire rock thin-sections in a digital format, including the possibility to switch between plain light and crossed polarizers, rotate the polarizers and zoom everywhere in the sample at high magnifications.

The hardware part of our system consists of a fully automated petrographic microscope, connected to a computer (Figure 1). The software contains three modules; a PetroScan launcher for setting up the image acquisition, optimizer for combining and interpolating the scanned images and a viewer program (Figure 3).

 

Figure 2. Automated sample stage allows capturing high resolution images of entire thin sections by combining thousands of images taken sequentially along a predefined grid. Credit: Jochen Hürtgen

High-resolution mosaics (Figure 2) containing up to a million images can be acquired in a few hours. With crossed polarizers, the scans are performed for several rotation angles. The movement accuracy of the sample stage is about 10 nm. This ensures precise overlap of the scanned mosaics at different rotation angles, where each corresponding pixel has the same x-y coordinates. Thus, the images of each rotation angle are precisely overlapping and allow to interpolate the change in brightness values for every individual pixel in the mosaics, producing smooth curves (Figure 3) that reflect extinction behavior in minerals (Heilbronner and Pauli, 1993).  Each attribute of these interpolated curves (phase, amplitude) of the sample can be viewed at any rotation angle.

 

Quantification of optical images

Figure 3. Screenshot of the PetroScan TileViewer window showing a thin section that is scanned under crossed polarizers. On the right side the scanned image is overlain by the phase map, where each pixel is color-coded according to the mineral extinction angles. Credit: Liene Spruženiece, Joyce Schmatz, Simon Virgo and Janos L. Urai

Virtual microscopy offers powerful means for quantitative analysis of giga pixel images to extract multi-scale information. It has a resolution of up to few micrometers for areas in sizes of 10 square centimeters. In its current form the viewer software that we developed contains a built-in toolbox for displaying and tresholding the intensity, saturation and hue values of the scanned images. This can be used for a quick estimation of sample porosity or proportions of different minerals. In addition, the scans with interpolated crossed-polarizer rotation also contain information of the mineral extinction behavior that is used to produce a “phase map” (Figure 3). The “phase map” displays the variations in the mineral optical axis orientations. Although it cannot provide absolute values for the crystallographic orientations, it allows a clear distinction between differently oriented mineral grains and allows easy visualization of qualitative lattice misorientations inside individual mineral grains. Furthermore, import and export functions are incorporated in the image viewing software. Thus more advanced image analysis can be carried out in specifically designated external softwares (e.g. Fiji/ImageJ, GiS, Matlab, Python), then inserted back in the PetroScan viewer software.

 

Applications in teaching

Figure 4. Mineralogy and fabrics of the scanned samples can be segmented and quantified using built-in toolboxes. The recognition of features is greatly improved by the possibility to zoom in the samples, switch between the plain light and crossed polarizers and rotate the polarizers. Credit: Liene Spruženiece, Joyce Schmatz, Simon Virgo and Janos L. Urai

At RWTH Aachen University, virtual petrography has been used as a teaching tool for several years. For example, it is extensively used in the microtectonics course. Each class consists of a short introduction in some of the common rock microstructures, such as veins, cataclasites, mylonites, dissolution-precipitation features and others. This is followed by an hour-long presentation by a student group, where they describe and discuss the respective microstructure in a scanned thin-section, projected to a screen with a high-resolution beamer. The students use laser pointers for characterizing the features and origins of the microstructure, switch between views in plain-polarized or crossed-polarized light, zoom in and out across the sample, adjust illumination settings, rotate polarizers and do basic image analysis (Figure 4). This has proven to be a highly engaging teaching method. The students in the audience question the presenting group asking to provide a closer look or more details on any interesting microstructural feature. Often hypotheses are formed by the listeners and immediately tested by the presenters, allowing to perform an investigation and agree on a reasonable explanation at the end of the class. Virtual petrography can be especially important in teaching institutions that do not own well-equipped microscopy labs. It only requires a computer and a projector. The thin section scans and PetroScan viewing software will be available for download online. Thus, such a method will not incur additional costs to the institutions, at the same time will provide a large collection of a high variety of geological samples.

 

The vision

We imagine a world-wide community built around the platform of virtual microscopy, where the information from different analytical methods can be exchanged between users and stored in open databases, available for teaching or research purposes. Many advantages arise from such a transformation. Work can be carried out anywhere without a need to access microscopes, several users can simultaneously view the same samples, data can be easily quantified and integrated between different methods, and thin-section libraries can be shared and exchanged by the user community.

Further plans for developing this method include collaboration with the computer vision department at RWTH Aachen University in order to create deep learning algorithms that allows quick and precise segmentation of rock microfabrics, such as mineral content and distribution, grain boundaries, grain sizes, porosity, etc. This has been made possible by recently obtained funding from the RWTH Aachen University “Exploratory Research Space – ERS”. The final data set will be public and shared with communities.

 

Edited by Derya Gürer

 

References

Heilbronner, R.P., Pauli, C. (1993). Integrated spatial and orientation analysis of quartz c-axes by computer-aided microscopy. Journal of Structural Geology, 15, 369-382.

Tetley, M.G., Daczko, N.R. (2014). Virtual Petrographic Microscope: a multi-platform education and research software tool to analyse rock thin-sections. Australian Journal of Earth Sciences 61, 631-637.

NASA (2007). Virtual Microscope. Available at: http://virtual.itg.uiuc.edu/

Schmatz J., Urai J.L., Bublat, M., Berlage, T. (2010). PetroScan – Virtual microscopy. EGU General Assembly, EGU2010-10061.

The Open University (2010). The Virtual Microscope for Earth Sciences Project. Available at: http://www.virtualmicroscope.org/.

Virgo S., Heup, T., Urai J.L., Berlage, T. (2016). Virtual Petrography (ViP) – A virtual microscope for the geosciences. EGU General Assembly, EGU2016-14669.

Cargèse Earthquake Summer School 2017

Cargèse Earthquake Summer School 2017


Earthquakes: nucleation, triggering, rupture, and relationships to aseismic processes – 
2-6 October 2017, Cargèse (Corsica)

A good spot to ponder over earthquake physics… or life! Credits: Elenora van Rijsingen

A summer school in October, isn’t that a bit late? Well, not if it is held in Cargèse, a small town at the coast of Corsica! After a successful first edition in 2014, scientists from all over the world gathered again last week at the beautifully located ‘Institut d’Etudes Scientifique’ in Cargèse, to learn, share, discuss, agree and sometimes disagree about all facets of earthquakes.

The scientific program of the course was built around  several keynote lectures per day, given by well-known scientists in these disciplines like Satoshi Ide, Chris Marone, Bill Elsworth, Gregory Beroza, Shamita Das and many more. In order to give the participants of the course the opportunity to share their own work as well, the keynote lectures were alternated with short talks and poster sessions.

Some free time to discuss in small groups. Credits: Elenora van Rijsingen

Topics like earthquake nucleation, triggering, rupture propagation, rate and state friction laws, induced seismicity and the wide range of ‘slow earthquakes’ were discussed. Due to the various backgrounds of both the participants and the keynote speakers, many different scales and aspects of these processes were addressed: from seismological observations to laboratory earthquakes, and from microfractures to the subduction megathrusts. Bridging the gaps between these different disciplines and scaling from the laboratory scale to the natural cases is a big challenge. Therefore, frequent interaction between the communities helps us to move forward together and better understand the intriguing processes behind earthquakes.

“On Friday evening we had a final discussion session which I enjoyed. All of us participants agreed on several common points like the connection with geological observations, simplifying our earthquake jargon and stimulate diversity by including more disciplines for potential future workshops. Considering the partial disagreements during session discussions and different standpoints from various communities this final agreement was a nice outlook. I hope this was not only because it was Friday evening and everybody was tired from an intense but inspiring week.” – Simon Preuss, PhD student at ETH Zurich

Posters were displayed outside throughout the week. Credits: Elenora van Rijsingen

And what better way to have this interaction in a beautiful and inspiring place like the Corsican coast? Fortunately, many of the participants remembered to bring their swimming gear so that they could go swimming during the long and lazy lunch breaks. Others would continue discussing at the posters or join the optional early afternoon sessions, which varied from software tutorial sessions to informal discussions about earthquake early warning systems and how to implement them. The small scale of the course, combined with the relaxed and informal atmosphere throughout the whole week made it a very successful event, almost like a scientific retreat! And the good news for the people who missed it: word is getting around that there might be a third edition of the course within a few years!

Teaching in the 21st century – a PICO session

Teaching in the 21st century – a PICO session

With the progress in the digital world there are more and more e-tools available for research and teaching. What are smart ways to make use of new techniques in teaching? For inspiration and learning, Hans de Bresser, Janos Urai and Neil Mancktelow convened a PICO session at the EGU 2017 General Assembly to showcase present-day e-learning opportunities to improve the efficiency and quality of teaching structural geology and tectonics. Despite an 8h30 morning slot and a limited number of abstracts, all spots during the 2 min madness were taken and all authors were talking the full 90 minutes – and some even well into the coffee break – about what they are using and how. Inspirational and fun!

Hans, you’re somewhat of an expert on teaching in Utrecht as former head of teaching in geosciences, how would you say that teaching has changed from back when you were a student?

”In my time as a student, I spent hours learning to identify rocks and recognizing minerals in thin sections, browsing back and forward in text books packed with determination tables and graphs of which more than half was not relevant for me. I also invested a lot of time in making maps in the field, carefully adding measurements and colors, hoping that I did it right the first time (never) and that I wouldn’t spill coffee over my precious products (it happened). And I sat in classes in which professors talked for hours, repeating the content of books that I had in front of me, while my level of activity in class was so low that preventing to fall asleep was a serious challenge.  I really learned a lot, definitely had a lot of fun, but looking back I feel it could have been done more efficiently. New styles of teaching, such as blended learning and flipping the class room, and state-of-the-art e-tools for data collection, modelling and visualization now help us to be very efficient and improve the quality of teaching. And it can be fun”.

 

Broadly speaking the presented aids can be divided in the following 3 categories, for each category we give examples below.

1. How to bring the real and experimental world into the classroom?

2. How to make life easier for a teacher?

3. Can we add extra information using the virtual world?

 

Category 1:  How to bring the real and experimental world into the classroom?

Benjamin Craven explaining his PICO: Fieldwork Skills in Virtual Worlds. Credit: Anne Pluymakers 

Virtual landscape, presented by Benjamin Craven: a 3D model of a field area, to bring the real world in the classroom, and increasing the efficiency of real world field teaching. Different packages of open source software make it easy to design your own landscapes, though you can also make use of the already made world. One could also teach students photogrammetry, to enable them to make their own 3D models using photos made with a smart phone, of rocks or other items in the classroom. One idea is to ask students to create a geological map in the virtual “field”, but then to give each student only a limited amount of time to do so. This allows them to learn how to plan their time in mapping projects.

Drones and 3D models , presented by Thomas Blenkinsop. Using drones one can make 3D surface models, which can be combined with cross sections to create a 3D MOVE (TM Midland Valley) project. It starts with the regular manual work, but adding the digital models improves 3D thinking as well as it allows students to check their own cross sections.

Deforming ice with students, presented by Dave Prior. A low-cost ice deformation rig, designed to be used by student teams. It brings Dave’s own research into the class room. Through a questionnaire teams were designated by Dave to get the right mix of skills to eventually present a poster. Some results are of high enough quality to publish.

 

Thomans Blenkinsop explaning about using drones, Lidar measurements and 3D models for undergraduate teaching. Credit: Anne Pluymakers.

Category 2: How to make life easier for a teacher?

Jupyter notebooks, presented by Florian Wellmann: software for those who can’t program. It makes it easier to create exercises, and to allow students to play around with parameters. Automatic and manual exercise grading are both easy.

STEREOVIDEO, presented by Jose A. Alvarez-Gomez. This is a (currently Spanish only) channel with various Youtube video instructions on how to use stereonets. The channel will bring more subjects later. It is very popular in Latin America.

 

Category 3:  Can we add extra information using the virtual world?

Zappar, presented by Friedrich Hawemann. Using an icon on a poster and a free download app one can add extra layers of information to images. It also recognizes objects such as a polished rock, allowing the teacher to add arrows, circles etc. to highlight features

 

By Anne Pluymakers (just a visitor) and Hans de Bresser (session convener)