Image Processing Techniques for Frequency Band Occupancy Detection in Radio Spectrum

The number of devices using the radio spectrum for communication has dramatically increased in recent years, creating a growing need for effective spectrum management. Spectrograms, which are graphical representations of the signal's frequencies spectrum over time, play a crucial role in visualizing and analyzing the spectrum usage. Thus, band occupancy detection is essential for efficient spectrum management. On one hand, it helps to identify which frequency bands are in use and to what extent, thereby minimizing interference. On the other hand, it facilitates the detection of unauthorized entities, that may be engaged in illicit activities, such as unauthorized spectrum use or signal interference. Consequently, the paper proposes image processing approaches for efficient RF frequency band occupancy detection. The results include spectrograms of signals from both the uplink and downlink communications of digital entities drawn from public databases. Moreover, timing consideration for the proposed methods are discussed, envisaging both CPU and GPU implementations.