Identification of earthquake types based on seismogram data

Abstract

Indonesia is one country that has many active volcanoes, and it can have a high potential for earthquakes due to volcanic activity. Central Java is an area that has an active volcano, the Mount Slamet. Various kinds of earthquake signals, earthquake strength, and frequency are recorded using a seismogram. The process of recognizing seismic signal patterns using short-wave transformations has a higher chance of success. Several types of earthquakes that have occurred on the Mount Slamet. There are three types of earthquake data tested, Shallow Volcanic Earthquakes (VB), Gust, and Tremors. This study aims to identify the type of earthquake vibration signals recorded on the seismograph. Earthquake image processing system consists of several parts. The image is normalized to get the image in the time domain. Then the image is processed with two processes to determine the characteristics of the earthquake. The Fast Fourier Transform process is used to determine the strength of earthquake signals based on the frequency. The quantization process is used to take samples of each data in the time domain. In this study, the method used for identification is pattern recognition and decision trees. The identification system can recognize signals that are approached using the Root Mean Square, Average Power Spectrum, and statistical features. The results of tests carried out obtain 100% accuracy of each method.