Automated Quality Control of Cutting Photos Using Deep Learning

Quality checking and correcting is a crucial task for the data in the oil and gas industry. Conventionally, cutting images quality checking task is manual/visual where the human subjectivity might greatly impact the final result causing high uncertainties/risk on operational decision making. Quality checking manually all cuttings is out of reach. Developing an automated solution is the only conceivable solution.

Quality checking and filtering out bad quality cutting images, has a significant role in assisting geologists toward having more accurate and trustworthy geological analysis of the available data which leads to improving decision-making. To achieve our goal of having a smart quality control system, we developed a neural network-based transfer learning system to provide support to geologists on recognizing and filtering out bad quality cutting images and concentrating on the ones with good quality. Moreover, the developed solution controls the follow of the bad quality images by sorting the cutting images along with the quality issue of them.