Creating realistic virtual training environments is critical for domains like transportation security, where spatial accuracy and cognitive fidelity support effective training transfer. However, generating such environments often demands significant time, expertise, and hardware resources. This case study explores a hybrid 3D scanning workflow that combines Light Detection and Ranging (LiDAR)-based environmental capture with high-resolution photogrammetry (FARO®) and Computer Aided Design (CAD) models to efficiently construct immersive, cognitively valid virtual spaces. We compare LiDAR and FARO® scans of a TSA checkpoint and its workstations using both objective (surface density, RMSE) and informal subjective metrics of realism. Results show that FARO® scans produce higher object fidelity despite lower average point density, while LiDAR offers rapid full-scene capture with acceptable spatial accuracy. Importantly, we demonstrate that integrating CAD models into LiDAR-derived environments provides a scalable, resource-conscious solution that supports perceptual realism where it matters most, on the objects users attend to and interact with. Our findings reinforce that point density alone does not determine visual realism and underscore the importance of aligning visual fidelity with cognitive demands. We also highlight the promise of emerging tools such as neural radiance fields (NeRFs), which offer photorealistic reconstructions from image data, as future complements to hybrid workflows. This work provides a grounded framework for designing cost-effective, high-fidelity Virtual Reality (VR) training environments, particularly in operational settings where development resources are limited.