Sales :  Support :

Registered in England and Wales company # 03449181.
This website does not use cookies.

© Copyright: Conwy Valley Systems Limited  2000 - 2026

Conwy Valley Systems Limited,

West Acre, Llys Helyg, Deganwy

CONWY LL31 9BN, United Kingdom.

VAT# GB 741 5190 48

A Geoscience Wales  founding member.

PETROG includes patent protected technology.

PETROG was originally designed to collect data from a live microscope image.  It has since evolved to capture and work not only with live images, but also to import data from a variety of sources - individual photomicrographs, imported SEM images, mosaics of photomicrographs that form a larger image, or whole slide scanned images created by external devices.

This gives you the capability of working away from the microscope - and in the case of multi-user setups with a shared PETROG database and access to the software on multiple workstations, then one user can be working on one sample with virtual images, while another is point counting from a physical sample on the microscope with the SteppingStage.

PETROG provides a standard structure for data and analysis. That in itself provides a valuable resource, but this resource can also be used to provide training data for Machine Learning / AI, so you can then use AI to assist when collecting data from new samples. Modern AI systems for automated mineral identification depend on large, structured, expertly labelled datasets—exactly the kind of data PETROG is designed to produce.

Each PETROG dataset combines precise compositional labels, exact spatial coordinates, and high-quality microscope imagery, giving Machine Learning models the rich contextual information they need to learn how minerals truly appear in thin section. Recent research highlights this need clearly: the 2025 LITHOS benchmark for automated petrography states that “scalable mineral classification is only possible when training data are both abundant and consistently annotated” [LITHOS 2025], while a 2025 Scientific Data release emphasises that “large, diverse, and expertly curated datasets are essential for accurate mineral classification” [Scientific Data 2025].

PETROG’s structured workflow ensures that every point-counted data point, and the corresponding saved image(s), are captured with the consistency and precision required to provide the input to create and train new AI models.

For more detail, please follow the links below:

Scanned Images, Archives and AI