A new integrated testing and modelling paradigm based on full-field imaging and finite element (FE) analysis that utilises full-field data fusion is proposed for structural evaluation and model validation at substructure level. The approach is developed for the assessment of a composite wind turbine blade (WTB) substructure subjected to multiaxial loading, mimicking in-service conditions, using a new reconfigurable loading rig. A steel mock-up equivalent to the WTB substructure was used to demonstrate the new experimental, numerical, full-field imaging, and data fusion approaches. Digital Image Correlation (DIC) and Thermoelastic Stress Analysis (TSA) were used to obtain the complex load response of the substructure. Strains and displacements derived from DIC were fused with numerical predictions obtained using a FE-based stereo-DIC simulator, which provided unparalleled like-for-like data comparisons. A numerical FEA solution for TSA was also obtained that accounts for heat transfer and allowed an independent means of structural evaluation. The challenges of deploying full-field imaging on the substructure scale are highlighted alongside procedures for mitigating multiple deleterious effects that are concatenated in large structures testing. It is demonstrated that high quality and fidelity experimental data can be obtained and fused with numerical models to provide a comprehensive and quantitative structural assessment at the substructure scale. It is shown that the proposed full-field data fusion efficiently reveals uncertainties in both the models and experiments. The work provides important steps to support virtual testing at higher length scales and their integration into the design, development, and certification programs of next generation, high-performance structures.