Noise threshold estimation in spectrum monitoring data analysis application

With the rapid development of radio service and monitoring facilities, spectrum monitoring applications step into big data era. Based on the support of big data analysis technology, many algorithms are proposed to get valuable information through dealing with massive monitoring data. However, performances of these algorithms are limited by the usual methods of setting noise threshold. In this paper, we proposed a new method of setting noise threshold based on estimating the mean energy level of noise by searching spectrum hole. Its validity and applicability are verified by the real monitoring data. In the end we give some advices about setting noise threshold for relative applications.