

The CobWeb work package in cooperation with APS Antriebs-, Prüf- und Steuertechnik GmbH under the SUGAR III undertakes the the development a Graphical User Interface (GUI), and to analyze high resolution X-ray computer tomography (XCT), Acoustic Emission and Ultrasound data analysis.
The XCT module of the CobWeb software is finalized and tested. The main advantage of CobWeb to commercially available XCT software’s is the using of Machine Learning libraries for 3D grayscale segmentation. The new workflow technique renders high accuracy and precision towards digital rock physics analysis.
The current version is capable to read and process (reconstructed) XCT files in .tiff and .raw format. Tools to zoom in, zoom out, cropping, color scale assist in understanding the XCT data better. Noise filters such as, non-local means, anisotropic diffusion, median and contrast adjustments are implemented to increase signal to noise ratio. The user has a choice of five segmentation algorithms namely, Kmeans, Fuzzy C-means (unsupervised), Support vector machine (supervised), Bragging and Boosting (ensemble classifiers) for accurate segmentation and cross-validation. Material properties like relative porosity's, pore size distribution, volume fraction (pore, matrix, mineral phases) can be calculated and visualized. The software runs stable on windows 7 and 8.
Contact
Name | Contact | |
---|---|---|
![]() Picture: Ingo Sass
| Prof. Dr. Ingo Sass | sass@geo.tu-... +49 6151 16-22290 B2|02 134 |
Further Information
- Funding Period: 36 Months (15.12.2014 – 31.12.2017)
- Grant Authorities: BMBF (German Federal Ministry of Education and Research)
- (opens in new tab) A new software collection for 3D processing of X-ray CT images
- (opens in new tab) Phase segmentation of X-ray computer tomography rock images using machine learning techniques: an accuracy and performance study
- Processing of rock core microtomography images: Using seven different machine learning algorithms
- Comparison of Micro X-ray Computer Tomography Image Segmentation Methods: Artificial Neural Networks Versus Least Square Support Vector Machine