AI Automates Critical Risk Classification for Navigational Chart Changes

Hydrographic offices face a significant challenge in manually reviewing and classifying changes to Electronic Navigational Charts (ENCs), which are crucial for maritime safety. This manual process is labor-intensive, struggles to scale with the volume of updates, and can lead to inconsistencies between analysts. A new research paper introduces an automated solution to address these issues. The proposed method leverages machine learning to classify ENC changes, determining whether they pose a critical or non-critical risk. It features a baseline encoding scheme that translates complex vector data changes into a structured tabular format. Key components of this scheme include a spatial context encoder, which enriches change representations with surrounding geographic features, and an ENC attribute encoder, capturing nuanced attribute-value descriptions of modified objects. Evaluated on two operational datasets comprising over 100,000 individual chart modifications, the approach achieved accuracies of 90% and 94% using tuned gradient-boosted trees. This represents a 5-7% improvement over models without the enhanced encoding, proving the effectiveness of integrating machine learning into geospatial workflows to enhance maritime safety and streamline ENC maintenance.