Abstract: This paper proposes an autonomous mobile platform for detecting defects in transmission line stay cables, integrating vibration-based diagnosis to achieve efficient inspection. The platform adopts a low-friction ball-bearing mobility mechanism together with a self-adaptive triple-jaw clamping device to ensure stable operation on the cable. A dual-mode excitation method, combining high-frequency electromagnetic stimulation with single-shot potential-energy impact, enables full-bandwidth signal acquisition through piezoelectric sensors. An FPGA-based real-time spectrum analysis algorithm accurately identifies cable defects, achieving a localization accuracy of ±0.5 m for six typical defect types. Field experiments demonstrate that the platform significantly enhances inspection efficiency while reducing detection costs.