EDGE-BASED ACOUSTIC MONITORING OF NATURAL SIGNALS USING RASPBERRY PI AND DIGITAL SIGNAL PROCESSING TECHNIQUES

Abstract

The article presents the implementation of an acoustic monitoring system based on the Raspberry Pi microcomputer, which provides local signal processing in real time. The developed algorithm combines methods of acoustic data preprocessing, statistical and spectral–wavelet analysis, as well as event detection based on the short-term to long-term energy ratio (STA/LTA).

The system was implemented using Python libraries (NumPy, SciPy, PyWavelets, SQLite3) and tested on both synthetic and real acoustic datasets. Experimental results confirmed the stable operation of the Raspberry Pi device during more than two hours of continuous recording without memory leaks, with an average CPU load not exceeding 30%.

The obtained results demonstrate the potential of the proposed architecture for developing autonomous bioacoustic and seismoacoustic monitoring stations operating under limited computational and power resources. In the future, the system is planned to be enhanced by integrating wireless communication modules and machine learning algorithms for intelligent classification of acoustic events and further expansion of its functionality.