An Imaging Science Case Study
I led the development of the oil industry's first comprehensive imaging standard (CRISP) for capturing visual data from “drilling cuttings” – rocks and dirt pulled up by drills. This standardized, quantitative approach to digitization enables oil companies to build more accurate AI models for predicting drilling outcomes and well productivity, while reducing data collection costs at scale.
Due to NDA, I cannot name the client, and am limited in the images and information I can share.
Left: Standard visible color image
Right: L*a*b* expanded image
This computational technique enhances subtle variations in rock color to visualize the value of high bit depth when encoding images for AI analysis.
The project bridged a critical knowledge gap between petroleum geologists and imaging scientists, creating a framework that transformed subjective practices into precise, measurable standards. By defining specific quality tiers aligned to different business needs, the standard enables petroleum companies to maximize the ROI of their cuttings imaging, and ultimately inform business decisions like where to invest in oil exploration.
Defined precise image quality metrics for AI model development versus human analysis
Established specialized UV fluorescence imaging protocols to detect hydrocarbon presence
Created practical quality control methodologies to encourage industry-wide adoption
Developed standards for light calibration, camera specifications, and color profiling
Developed novel visual data compression strategy to preserve critical mineral characteristics while reducing computational overhead for AI processing
The resulting standard enables consistent, comparable data collection across the industry, forming the foundation for next-generation computational modeling in petroleum exploration.
This website uses cookies to improve your experience.