Abstract: This study aims to provide a method for rapidly and accurately capturing the evolution laws of the drastic changes in subsurface structures to provide reliable technical support for key fields such as ground collapse early warning and urban underground space detection. Taking the key area traversed by the shield of Shenzhen Metro Line 11 as an example, real-time transmission node acquisition equipment and a real-time calculation platform were employed in this study. By combining this with the spatial autocorrelation (SPAC) technology of ambient noise, a theoretical model for spatial autocorrelation imaging was constructed, and a full-process real-time monitoring platform was established. The measured data showed that the background noise signals collected by this system exhibited extremely high stability and reliability. Additionally, the abnormal areas of the captured underground velocity structure exhibited a high degree of spatiotemporal coupling with the advancing trajectory and time process of shield construction. More crucially, the monitoring system achieved a technical breakthrough, with the spatial resolution reaching the meter level and temporal resolution up to the hourly level. Moreover, the high-precision and high-frequency real-time dynamic monitoring of subsurface structures within a depth of 50m was possible even with a short survey line layout. In addition to opening up an innovative path for the safety monitoring of subway shield construction, the research achievements also provided important theoretical support and engineering practice examples for works such as the early warning of the risk of urban ground collapse, refined census of underground space, and emergency response.