LuminPop: A Real-Time Satellite-Ground Collaborative System for Population Detection Based on Nighttime Imagery

With the accelerated urbanization process in China, real-time population monitoring has be- come increasingly important. This paper proposes a star-ground collaborative computing system named LuminPop, which utilizes nighttime light data to quickly and accurately predict population distribution in urban areas. The system employs neural network models to address traditional population censuses delays and resource limitations. By integrating various existing technolo- gies, LuminPop enhances prediction accuracy while reducing satellite bandwidth use and energy consumption. Experimental results show that the system performs well in real-time population forecasting across multiple regions, supporting government public resource allocation and emer- gency management. In the future, we plan to expand this system to more areas to further improve its practicality and accuracy.