Abstract: Aerospace equipment structures are prone to hard-to-detect damages induced by diverse factors such as manufacturing processes, environmental conditions, and operational loads during fabrication and service, posing significant threats to operational safety. Ultrasonic guided waves (UGWs) are a promising method for thin-walled structure inspection, yet their reliance on dispersion curves and reference signals remains a major challenge. This study addresses UGW-based damage detection under prior-free conditions by establishing a perception matrix equation using sparse random measurement points. A frequency-wavenumber domain sparse reconstruction method is employed to extract dispersion information from direct waves, enabling accurate UGW dispersion curve reconstruction. Furthermore, a guided wave propagation data dictionary is constructed based on the UGW propagation model, and spatiotemporal sparse reconstruction is applied to extract spatial information from scattered waves for damage detection and localization. Experimental and numerical results demonstrate that the proposed method achieves dispersion curve reconstruction errors below 5% and damage localization errors of less than 1 cm. Compared to traditional UGW methods, the multi-domain sparse reconstruction framework eliminates dependencies on dispersion characteristics and reference signals, significantly enhancing detection capabilities and broadening application scope.