Our transportation research applies computer vision and machine learning to improve roadway safety, traffic modeling, and transportation infrastructure assessment. Recent work includes beta distribution learning for reliable roadway crash risk assessment, omnidirectional depth estimation via stereo matching from multi-cylindrical panoramas (MCPDepth), and multi-scale satellite imagery for fatal crash risk estimation. We also develop methods for dynamic traffic modeling from overhead imagery, remote estimation of free-flow speeds, and fully automated roadway safety assessment (FARSA) using LiDAR and overhead imagery. Our research addresses critical challenges in transportation safety and infrastructure monitoring.