This study introduces an innovative AI-driven approach for automated railway track condition monitoring by integrating vibration data from train-mounted sensors with high-resolution GPS and infrastructure element information. The primary objective is to accurately detect critical railway infrastructure components, such as turnouts and joints, enabling a transition from traditional disruptive and costly periodic inspections to a proactive predictive maintenance framework. Specifically, the proposed method leverages AI techniques to distinguish between infrastructure elements, normal track conditions, and potential anomalies based on vibration signal analysis. Two complementary vibration sensors, positioned on either side of a train operating in Sweden, capture dynamic responses that, when synchronized with precise GPS measurements and filtered infrastructure event data, enable robust spatial-temporal correlation despite challenges like misaligned data streams and GPS inaccuracies. A comprehensive data processing pipeline is introduced, consisting of stages for data acquisition, cleaning, filtering, feature extraction, and anomaly detection through AI techniques. Our integrated system will not only address the inherent complexities of sensor fusion and signal processing but will also significantly reduce false alarms by reliably separating genuine track discontinuities from background noise. This enhanced discrimination will minimize unwarranted alerts while ensuring swift detection of critical issues. Overall, the system will demonstrate substantial potential to optimize maintenance efficiency, improve operational reliability, and elevate safety in railway infrastructure management.