iCore

We have developed a deep learning-based software capable of estimating geotechnical properties, such as Rock Quality Designation (RQD)  Joint Number (Jn), Ja and Jr from core images with  accuracy. This innovative software can also identify the dip and dip direction of  fractures from 2D core tray images and construct a Stereonet for further analysis.

iCore proceeds through two components: statistical and AI driven and mathematical driven. The detection of fractures are performed through AI training models and the data extracted from this components are used in mathematical models to construct the 3D fracture geometries. The internal fracture parameters such as Ja is analysed through statistical models.

The models have been already trained on thousands of tray images from east coat of Australia. We can also train individual models if the core tray images are provided by the companies.