This paper introduces a novel system for obtaining regional population data using mobile signaling data to overcome the limitations of traditional methods like censuses and surveys. The system leverages real mobile signaling data from the Guangzhou platform and employs a base station clustering method based on traffic communities to address signaling drift and ensure accurate location data. It enhances population statistic accuracy using eigenvalue curve fitting and optimization methods to determine feature thresholds. Compared to the seventh national census data, the system achieves over 95% precision for the working population and 88% for permanent residents. Additionally, it validates its large-scale application capability through the analysis of temporal and spatial population distribution in the Greater Bay Area. This approach offers a cost-effective and efficient solution for regional population data acquisition, supporting better-informed policy-making and urban planning.