AbstractDistortion‐induced fatigue cracking is a concern for many aging steel bridges in the United States. Human visual inspections to characterize fatigue cracks have many drawbacks including inconsistencies in identification, significant time and monetary costs, and safety risks posed to the lives of both inspectors and the traveling public due to lane closures. Digital image correlation (DIC) is a vision‐based technology which has shown promise for identifying and characterizing fatigue cracks. Three‐dimensional DIC measurements can capture full‐field displacements and strains, allowing for detection and characterization of both in‐plane and distortion‐induced fatigue cracks. This paper describes an experimental study in which a half‐scale steel girder to crossframe subassembly was subjected to distortion‐induced fatigue loading to produce multiple geometrically complex cracks. A DIC‐based crack characterization methodology was applied to quantify the cracks, which was successful at characterizing cracks propagating in the girder web but struggled to characterize horizontal cracks. Additional work is needed to improve the accuracy of the DIC‐based crack characterization methodology to use as an automated bridge inspection tool.