Sample labeling method using image-matching

Abstract

A method and system for labeling and identifying a sample, including a rock core sample, after determining a physical primary label has been compromised is provided. The method includes labeling the sample by placing a physical primary label on the sample or sample container, creating a first rollout image of the sample for a secondary label, and storing the first rollout image in a secondary label digital database with a sample identifier. The method continues for identifying the sample after determining the physical primary label has been compromised by creating a secondary rollout image of the sample and searching the secondary label digital database using the second rollout image to create a plurality of matching scores using a machine learning or deep learning image matching technique. The matching scores are then evaluated to determine an identify of a sample and the sample is re-labelled using a replacement primary label.

Dependencies are inferred from the claim text.