Estimation of the degree of recrystallisation of carbonates based on machine learning using thin sections

Recrystallisation processes in carbonates can affect the chemical and isotopic composition, the preservation of organic matter and fossils, change the filtration-capacity properties of the rock. Consideration of these processes seems to be critical for correct interpretation of the results of a wide range of analytical studies. On the basis of computer vision and machine learning technology the models for qualitative and quantitative express estimation of the degree of recrystallisation of sedimentary carbonate rocks are developed on the basis of thin-section images. The models have been trained on the basis of 300 images and 45000 fragments of thin-section images. The achieved accuracy of the models exceeds 90 %. The results of application of models and software based on them can be used for comparison of geochemical and isotopic information, as well as for express selection of the least recrystallised samples for analytical studies.