Particle size distributions or PSD are predicted from a composite of imagery collected by GeologicAI’s core scans including high resolution color photography (RGB) and short-wave infrared (SWIR). Color and infrared channels are employed to extract pertinent pixel and textural features in the core images. The core photographs obtained from our scanning process are so detailed GeologicAI is able to determine grainsizes below 2 microns in size. To put that in perspective, sieve stacks used in many oil sands core programs are only used to down to the 44mm size. In order to obtain grain size distributions smaller than that requires wet sieving and laser analysis, a costly process with a number of potential procedural errors in the laboratory.
The final steps in GeologicAI’s PSD prediction utilizes the pixel and textural features along with available lab data to derive statistical models for the prediction of PSD for unsampled intervals in existing wells and for new wells in real time as the core is scanned.
Blind tests of our PSD predictions have yielded extremely promising results thus far and we are currently engaged in a core program that is utilizing our predictions exclusively over the laboratory analyses.