Steel bridge deterioration presents critical challenges for inspection and load rating, particularly at girder ends where complex corrosion patterns exceed the capabilities of traditional assessment methods. This study establishes a comprehensive framework for digital evaluation through three contributions. First, systematic classification of 553 inspection reports across six New England states confirms regional consistency of corrosion topologies, supporting standardized assessment protocols. Second, a platform-independent methodology transforms terrestrial LiDAR and photogrammetry data into engineering-ready thickness contour maps through resolution-constrained optimization that balances geometric fidelity with computational efficiency. Third, full-scale testing of eight naturally corroded girders validates both scanning methodology and finite element models, achieving 5% average prediction errors for stiffness and capacity while successfully reproducing observed failure modes. Validation of Massachusetts capacity equations using scan-derived parameters demonstrates 5.14% mean prediction errors. The integrated framework enables geometry-informed capacity evaluation that advances beyond discrete measurement approaches, establishing a foundation for implementing digital assessment technologies in bridge engineering practice. The framework supports practical engineering applications including bridge rating, load posting decisions, and prioritization of rehabilitation across large bridge inventories.