Track Geometry Data-Driven Interlayer Defect Detection of High-Speed Rail Slab Track Based on PSO-SVM Model

With the development of high-speed rail, the decline in track performance has led to a high incidence of track geometry defects, affecting the safety of train operations. Analyzing track geometry data will greatly improve the efficiency of track maintenance and repair. Based on the track surface irregularity data of the Shanghai-Hangzhou high-speed railway, features of track defects are extracted and input into a defect identification model based on Particle Swarm Optimization-Support Vector Machine (PSO-SVM) for identification. Results show that the extracted feature indicators comprehensively reflect the defect information hidden in the data; using the principal component analysis method for dimensionality reduction can improve the efficiency of defect identification. The PSO-SVM model achieves 100% recognition accuracy for the selected test set, providing a technical foundation for early warning of track structure defects.