Internal defects in rails and degrading ballast support conditions in railroad crossties (sleepers) are some of the primary causes of train derailments. Current ultrasonic rail inspection techniques can operate up to speeds of 25 mph. Lower testing speeds result in traffic disruptions and are, therefore, not a sustainable inspection solution. Moreover, existing techniques for detecting degraded tie-ballast interface conditions (e.g., center-binding) require transducers mounted on individual ties and do not support in-motion testing. Techniques introduced in this dissertation have the potential of operating at revenue speeds using the concept of a smart train, with onboard transducers, a data-acquisition system, and real-time data processing, to flag defective rails and crossties during regular service runs. The first part of this study presents a high-speed non-contact rail inspection technique that utilizes an array of capacitive air-coupled ultrasonic transducers in an output-only mode to extract a transfer function for a rail segment in a passive manner. The passive approach utilizes the ambient excitation of the rail induced by the different mechanisms of rail-wheel interaction. The extracted transfer function is sensitive to the presence of internal rail flaws such as transverse defects. Features from the transfer function are tracked using a statistical outlier analysis with an adaptive baseline to compute a parameter called the Damage Index (DI) that determines if the probed rail segment has a discontinuity. Performance of the system, in terms of receiver operating characteristic curves, are presented from field tests conducted at the Transportation Technology Center Inc. (TTCI) at testing speeds of up to 80 mph. The second part of this study introduces a non-contact railroad tie inspection system using an ultrasonic sonar-based ranging technique for detecting ties with center-binding support conditions. A transducer array placed parallel to the railroad tie, is used in pulse-echo mode, and full-field tie deflection profiles are computed using a reference-based cross-correlation. Proof-of-concept tests for deflection measurements in a 3-point bending test are presented. In-motion field tests were performed on a replica railroad track at the Rail Defect Testing Facility and a BNSF yard by mounting an inspection prototype on a train car.