EXCITE Knowledge Base

After image acquisition, specialized image processing software can provide useful microstructural data. Image processing allows for accurate assessment and quantification of information contained within the image.

For an in-depth overview of image processing techniques in Earth science, visit: Renée Heilbronner & Steve Barrett (2014), Image Analysis in Earth Sciences (also useful for other scientific fields).

Image pre-processing is a technique used to prepare raw image data for analysis and model training with machine learning. The goal is to enhance the image quality and extract relevant features, making it easier for machine learning algorithms to learn and make accurate predictions.

Image post-processing in machine learning refers to the set of techniques applied to the output of a model to enhance the results, interpret them more effectively, or make them suitable for further use. These steps are crucial for improving the usability and presentation of the model’s predictions.

Image segmentation is a process in computer vision and image processing where an image is divided into multiple segments (phases or regions) to simplify or change the representation of the image into something more meaningful and easier to analyse.

The goal of segmentation is to separate an image into regions that are homogeneous with respect to one or more characteristics, such as colour, intensity (i.e. pixel intensity), or texture.

Here are a few image-processing tools and software that are available to download for free.

– Fiji (Imagej): Fiji is an image processing package distribution of ImageJ2, bundling a variety of different plugins that facilitate scientific image analysis.

• Fiji is easy to install and has an automatic update function, bundles a lot of plugins, and offers comprehensive documentation.
• Fiji is an open source project hosted in a Git version control repository, with access to the source code of all internals, libraries and plugins, and eases the development and scripting of plugins.

– HyperSpy: Open source Python framework for exploring, visualizing, and analyzing multi-dimensional data.

Open-source Image Segmentation Tool:
– ilastik: An interactive machine learning and segmentation toolkit. No machine learning expertise is required.

Image processing tools in Python:
– scikit-image: A collection of algorithms for image processing.

Digital Sreeni

Dr. Sreeni Bhattiprolu from ZEISS makes his own personal YouTube videos all about Machine Learning (ML), Deep Learning (DL), and Image Processing. These videos provide an overview and step-by-step guide to various types of image processing techniques. Some of the more relevant video playlists from his YouTube channel are included below:

• Python tutorials for image processing and machine learning
• Image processing with python
• Image filters in python

Advanced Materials Characterization Techniques (AMC-Tec)

The Advanced Materials Research Laboratory (AMReL), University of Peradeniya, is mainly engaged with research based on thin film solar cells, gas sensors, and value addition to natural graphite. On their YouTube channel, they discuss advanced analysis techniques, for SEM images, using ImageJ software which may also be applicable to Earth & Environmental materials research:

Analyzing and visualizing your imaging data provides a deeper understanding of your material’s structure, properties, and performance. Regardless of the scale, data modality, or organizational profile—whether a large industrial company, core imaging facility, national or local service laboratory, or academic institution—Thermo Scientific Avizo Software offers optimized workflows for advanced materials characterization and quality control within a single environment. Avizo Software serves as a universal, reliable, fully automatable, and customizable digital analytical lab.

Materials scientists and engineers can accelerate innovation, enhance the reliability and performance of materials and processes, and reduce the cost and time required for discoveries.

Access to the Avizo software is provided by some of our scientific partners for detailed analysis of your imaging data. We provide an in depth online workshop highlighting some of the capabilities of the Avizo Software:

Scanning Electron Microscopy (SEM) is widely used to investigate the surface morphology and composition of materials. SEM provides detailed images by scanning a focused electron beam across the sample surface and detecting the emitted secondary or backscattered electrons.

For background information, practical aspects, and a training guide about SEM and its applications, visit: MyScope SEM Basics and our Introduction to Electron Microscopy

To acquire the most accurate results, it is necessary to properly prepare your samples. Depending on your sample, preparation might include polishing, washing, drying, fixation, mounting, and coating.

Biological Samples:
Biological samples should be fixated on a stub and completely dehydrated to reach higher vacuums for the best results. Samples that cannot be completely dehydrated can be imaged in a cryo-SEM.

Geological Samples:
Geological materials can be analysed in a range of different types, including rock chips that are polished and mounted to a stub, as well as polished thin sections. For analysing loose powder or particles, the particulate can be compressed in to a tablet, or spread out over a piece of carbon tape, followed by a flow of N2 gas to remove any access particles. Most geological materials are non-conductive and can trap electrons, which can influence measurements and imaging quality. A conductive coating, usually consisting of gold, platinum, or carbon, deposited as a thin film (10 nm) on the sample can prevent this, by conducting electrons away from the sample.

Electron Backscatter Diffraction (EBSD) is a technique used in SEM to perform quantitative microstructural analyses. EBSD can analyse crystal orientations, grain boundaries, and phase distributions on a millimetre to a nanometre scale, providing valuable insights into the material’s microstructure.

To learn more about EBSD, visit: EBSD Information and Training

To analyse EBSD data, we suggest using the following open access software: MTEX
– MTEX is a free Matlab toolbox for quantitative texture analysis, particularly for the analysis of EBSD data.
– Guides: MTEX forum

Energy Dispersive X-ray Spectroscopy (EDS or EDX) is an analytical technique used to determine the elemental composition of materials. It is often used in conjunction with a Scanning Electron Microscope (SEM) or a Transmission Electron Microscope (TEM). EDS/EDX is a powerful tool for material characterization, providing essential information about the elemental composition and distribution within a sample.

To learn more about EDS/EDX analysis: EDX/EDS information and training

The best open source software available for EDS/EDX analysis is DTSA-II, developed by NIST.
– NIST DTSA-II builds on the best available algorithms in the literature to simulate, quantify and plan energy dispersive x-ray analysis measurements.
– Introduction and tutorials for DTSA-II are available on YouTube.
– Relevant publication: Performing elemental microanalysis with high accuracy and high precision by scanning electron microscopy/silicon drift detector energy-dispersive X-ray spectrometry (SEM/SDD-EDS).

To analyse EDS/EDX maps, we suggest ImageJ.
– ImageJ tutorials

CL is an optic electromagnetic effect where photons are released after excitation of defect centres in the sample. CL is an effective method to analyse the chemical and structural variations of the material.

LumiSpy is an open-source HyperSpy python library aimed at helping with the analysis of luminescence spectroscopy data.

CL maps can be processed using ImageJ.
– ImageJ tutorials

FIB is a techniques that makes use of a focussed beam of ions that can be used for imaging, milling, or deposition of material. An advantage of FIB imaging is a higher canneling contrast compared to SEM imaging. However, a disadvantage of FIB imaging is its destructive nature. This destructive nature can be exploited for milling samples at nanoscales. A FIB can also be used to deposit material on the sample, by injecting an additional deposition material.

Applications of FIB:
– Cross sectioning
– FIB tomography (3D analysis)
– Sample preparation (e.g. TEM thin foils, Atom probe samples, and other sensitive materials)

For background information, practical aspects, and a training guide about FIB-SEM and its applications, visit: MyScope FIB overview, or watch the following video.

A TEM works similarly to an SEM, where a beam of electrons is used to image materials. However, a TEM works by tranmitting a focused beam of electons trough a very thin sample (< 100 nm). The electron beam passes through the sample which modifies and imprints its image in the electron beam. The beam is then magnified and focued on a detector, allowing magnification of very thin samples down to atomic resolution. FIB-SEM is capable of producing thin enough for TEM analysis.

For background information, practical aspects, and a training guide about TEM and its applications, visit: MyScope introduction to TEM, or watch the following videos.

An Electron Probe Microanalyzer (EPMA) is a sophisticated analytical instrument used for non-destructive chemical analysis of small volumes of solid materials. It combines an electron microscope with X-ray spectrometers to provide detailed information about the elemental composition of a sample.

EPMA is widely used in various fields such as geology, materials science, metallurgy, and semiconductor research due to its ability to provide precise and localized elemental analysis with high spatial resolution. Watch the video lecture below for a more detailed introduction

Software:
For quantitative WSD analysis we recommend the Probe for EPMA software. It offers unique and powerful options, such as non-linear background corrections for trace element analysis, multi-point background corrections for very complex matrices, and time-dependent intensity corrections for volatile elements or Na migration.

Sample preparation:
To ensure optimal data aquisitions, proper sample preparation is essential. The EPMA specialists at Utrecht Univeristy have developed a document outlining sample requirements for EPMA:

X-ray Microscopy is a non-destructive imaging technique that uses X-rays to obtain high-resolution images of the internal structures of materials. This technique allows researchers to visualize and analyze the internal composition of samples without the need for physical sectioning.

Watch the following video for a brief overview of XRM. Or, read our short introduction to X-ray microscopy.

X-ray Computed Tomography (XCT), including X-ray Micro-Computed Tomography or Micro-tomography (μCT), is a specific application of X-ray Microscopy that generates 3D images of a sample by compiling a series of 2D X-ray images taken from different angles. XCT provides detailed information about the internal features and spatial arrangements within a sample, making it invaluable for research in Earth and Environmental sciences.

For an in-depth understanding read more here: An Overview of 3D X-ray Microscopy, or watch the following videos:

Relevant Publication for Applications in Geoscience:
Please also refer to this comprehensive review of X-ray Tomography in Geoscience by Dr. Verlee Cnudde and Dr. Matthieu Boone from our partner facility, Ghent University: Cnudde, V., & Boone, M. N. (2013). High-resolution X-ray computed tomography in geosciences: A review of the current technology and applications. Earth-Science Reviews, 123, 1-17.

Data visualization is crucial for interpreting and analyzing the intricate details captured in images. Tools for visualizing and analyzing data enable researchers to better understand the structure and properties of materials, facilitating advances in academic research and industrial applications.

Open-source software for 3D data visualisation & analysis:

Fiji (Imagej) is a useful and free-to-download tool that aids in image processing. In this case, Fiji can be used to assist with 3D data visualisation.

Dragonfly offers a comprehensive suite of features for quick and meaningful results, including straightforward data visualization and AI-based segmentation. Dragonfly allows for multi-scale and multi-modality segmentation and analysis capabilities. Dragonfly supports a wide range of imaging modalities such as X-ray tomography, SEM, FIB-SEM, ion beam, and confocal microscopy and has many video tutorials to follow. A free 30-day trial is available for users to explore its capabilities.

Avizo Software by ThermoFisher, similar to Dragonfly, helps users investigate material structures, properties, and performances across different scales and data modalities. Avizo Software offers optimized workflows for materials characterization and quality control, making it suitable for industrial companies, imaging facilities, service laboratories, and academic institutions. A free trial is also available.

Raman spectroscopy is a powerful tool for providing chemical and structural information of molecules. Besides a Raman fingerprint, Raman spectra contain other information like fluorescence background, Gaussian noise, cosmic spikes and other effects dependent on experimental parameters. These undesired effects have to be removed to ensure that analysis is based on the Raman fingerprint and not something else.

Relevant papers on data processing:
– Guo, S., Popp, J., & Bocklitz, T. (2023). Key Steps in the Workflow to Analyze Raman Spectra.
– Bocklitz, T., Walter, A., Hartmann, K., Rösch, P., & Popp, J. (2011). How to pre-process Raman spectra for reliable and stable models?. Analytica chimica acta, 704(1-2), 47-56.

Open access software tools to process Raman data:
– RamanSpy
– OpenRAMAN
– Origin
– CrystalSleuth

The fundamental principle of LIBS involves using a high-energy laser pulse to vaporize a sample into a high-temperature plasma, causing electrons to excite and emit photons as they return to lower energy levels. Each element emits photons with unique wavelengths, creating a distinct spectral fingerprint that is collected and analyzed by a spectrometer to produce a LIBS spectrum.

Each element thus provides a unique pattern of spectral lines. The Atomic Spectra Database interface for LIBS, offers easy access to a convenient dataset, enabling users to plot spectra of various chemical mixtures under typical laser-induced plasma conditions and compare them with experimental spectra.

For more background information about LIBS read our introduction to LIBS, or watch the following video.

Our partners at CEITEC provide access to a Google Colab script (“Classification of LIBS spectra”) for the fundamental elaboration of LIBS spectra. This script allows for advanced processing of LIBS spectra, using machine learning algorithms. The Google Colab scirpt provides all the necessary data, and guides the user with easy to follow step-by-step instructions, for straightforward use.

Atomic force microscopy (AFM) works by scanning a sharp tip attached to a force-sensitive cantilever across a sample surface, measuring atomic-scale attractive and repulsive forces to create detailed topographic images. The deflection of the cantilever is detected by a laser beam reflected off its top, which records variations on a photodiode detector. AFM operates in three modes: Contact mode, offering high resolution but potential sample damage; Non-contact mode, reducing damage at the cost of resolution; and Tapping mode, balancing resolution and sample preservation by oscillating the cantilever.

For background information, practical aspects, and a training guide about AFM and its applications, visit: MyScope introduction to AFM, read our introduction to AFM, or watch the following video.

AFM data processing is performed using software that loads the topography images from scans (TIFF) and enables the simultaneous generation of 3D images with up to two superimposed magnitudes (e.g., Topography and Phase) and the obtaining of roughness data, either quadratic or arithmetic, over the entire image or a region of interest.

For analysing AFM data, we suggest our proprietary PARK SYSTEMS software called XEI, which also allows for the analysis of F-D curves from force spectroscopy. This software is available to all TNA users free of charge, through our partners at UGR. Additionally, there is another free software called Gwyddion for data analysis.

UGR also offer additional modules for measurements, not only of topography but also of:
– Measurement of conductive properties (e.g. EFM, C-AFM, and I/V curves)
– Mechanical property mapping (e.g. Adhesion, Stiffness, Elastic Modules, and F-D Curves)

NanoSIMS, short for Nano-scale Secondary Ion Mass Spectrometry, is a cutting-edge analytical technique used in various scientific disciplines, notably in geology, biology, and materials science. It enables researchers to visualize and quantify the distribution of isotopes and trace elements within samples at the nanometre scale, by using a focused beam of ions.

For more information, please read our short introduction to NanoSIMS, visit the MyScpope introduction to SIMS, or watch the following webinar by CAMECA.

Look@NanoSIMS (LANS) is a free software tool developed and actively maintained by Lubos Polerecky, who operates the NanoSIMS facility at Utrecht University. LANS is distributed as MatLab code and thus requires a working version of MatLab to run. Key features of LANS include:
– Loading of secondary ion counts (SIC) image data.
– Drift-correction and accumulation of planes.
– Regions of interest (ROIs).
– Quantification and export of isotope and element ratios.
– Statistical analysis of data in ROIs.
– Processing of multiple nanoSIMS datasets (aka “metafile processing”).
– Import and alignment of external images (e.g., TEM, SEM, AFM, fluorescence).

Because Lubos Polerecky is in charge of the development of LANS as well as running the NanoSIMS facility. Full support, tutorials, and even new software features can be provided, should the user desire.

Relevant publication: L. Polerecky, B. Adam, J. Milucka, N. Musat, T. Vagner, M. M. M. Kuypers (2012). Look@NanoSIMS – a tool for the analysis of nanoSIMS data in environmental microbiology. Environmental Microbiology 14 (4): 1009–1023.

Optionally, OpenMIMS is an ImageJ plugin, that does not require MatLab as proprietary software.