Research
My research brings structural engineering, computer vision, and artificial intelligence together to improve how infrastructure is documented, inspected, and assessed.
I develop computer-vision and deep-learning methods for recognizing structural elements and damage in images and point clouds. The objective is to make inspection information more quantitative, repeatable, and useful to engineers.
This work uses UAV imagery, photogrammetry, LiDAR, and automated semantic processing to create digital representations of bridges and buildings. These representations connect geometry and inspection evidence for infrastructure decision-making.
My broader structural-engineering work includes vibration and dynamic response, condition monitoring, and probabilistic seismic hazard assessment for resilient civil infrastructure.
Projects
Published and ongoing work spanning the path from field data to inspection-ready engineering information.
BridgeELSpect is an automated element-level bridge-inspection workflow that uses UAV imagery, photogrammetric reconstruction, damage segmentation, 3D damage mapping, and geometry-driven quantification to produce condition-state information.
BridgeEQA-3D is an ongoing research direction exploring embodied question answering for bridge inspection in three-dimensional environments. Further details will be added as the work becomes public.
The project builds on the publicly available BridgeEQA research on virtual embodied agents for real bridge inspections.
This research program combines reality capture, bridge-element segmentation, and inspection data to develop useful digital twins of existing bridges. It includes field collection using UAVs and methods trained on synthetic bridge point clouds.
Transformer-based models trained exclusively on synthetic data identify individual reinforced-concrete bridge components in real point clouds, supporting automated creation of structured bridge representations.
Previous work includes earthquake-catalogue development, seismic source modeling, and probabilistic seismic hazard assessment for national and site-specific applications in Pakistan.