Ocean Sensor Quality Control Transformation with AI

This document is a technical report detailing the development of a machine learning-based system for automated quality control (QC) of oceanographic sensor profiles. The report outlines the challenges faced by a large Federal Government defense agency that relies on accurate environmental intelligence for operations worldwide, including anti-submarine warfare and navigation safety. Koniag Government Services (KGS) created a solution that utilizes machine learning to detect anomalous sensor readings, thereby enhancing the accuracy of real-time environmental data. The system was trained on extensive historical QC decisions, allowing it to replicate expert judgment in identifying and removing anomalies. The report highlights the use of AWS SageMaker for scalable infrastructure, enabling efficient processing of sensor data and rapid model development. The outcomes indicate significant improvements, including a 43% increase in anomaly detection rates and an 85% reduction in false positives, while processing over 300,000 sensor profiles per minute, thus streamlining the QC process and improving operational efficiency.