Ultrasonic testing is a cornerstone of Structural Health Monitoring (SHM), providing a powerful non-invasive approach for assessing defects and material properties in aerospace, maritime, and civil infrastructure. Nevertheless, conventional ultrasonic inspection methods face significant challenges when applied to realistic engineering structures such as railway tracks and composite materials. Key challenges include achieving fast imaging with practical hardware, extracting coherent information from complex and noisy wavefields, and establishing reliable links between sensing data and mechanical performance. This seminar highlights recent advances in ultrasonic SHM that address these challenges through the integration of experimental mechanics, computational modeling, signal processing, and optimization. First, ultrasparse imaging techniques are developed for both active and passive sensing within the Synthetic Aperture Focusing Technique (SAFT) framework. Two prototype systems demonstrate real-time 3D imaging of internal rail flaws with enhanced speed and resolution. Second, data-driven signal processing methods are introduced to adaptively separate defect signals from noise. The frequency-coherent matched-field beamformer achieves super-resolution, high-contrast damage imaging in stiffened composites, while an iterative eigenfilter effectively suppresses artifacts and preserves true defect features in ultrasonic scanning videos. Finally, a non-contact ultrasonic testing system is developed to identify 3D viscoelastic properties of composite materials from 1D guided wave data via an inversion-based digital-twin approach. This enables in-situ monitoring of curing in automated composite manufacturing, as well as quantitative evaluation of impact damage in built-up structures.