Systematic mapping study: automatic recognition and localization of volcanic seismic events

Active volcanoes represent a significant hazard to surrounding communities, requiring continuous and reliable monitoring. While seismic signals constitute the main source of information, each volcano produces massive datasets that are difficult to process and interpret manually. Consequently, automated recognition and localization systems have become essential for the real-time detection and assessment of volcanic activity. This article presents a systematic mapping of the state of the art in automatic volcanic seismic events recognition and localization systems. It reviews the different types of volcanic microearthquakes (e.g., Long Period, Volcano Tectonic, Tremor, among others) that can be identified and analyzes the main computational techniques employed, ranging from classical signal processing and statistical methods to modern machine learning and deep learning approaches. In particular, this study describes how these techniques are applied to volcanic seismic events detection, classification, and localization. Additionally, we identify the most studied volcanoes worldwide to serve as reference points for testing these systems. Finally, we discuss current challenges, the need for scalable real-time solutions, and outline possible research directions to address them.