Organizations specialized on disaster prevention take pictures of Popocatepelt volcano in Mexico every minute. Proper and accurate investigation of these pictures could be of great interest for disaster prevention and civil protection.
Using the methodology of pattern recognition, we apply concepts of discrete dynamic systems and non-linear time series to investigate a set of digital images of Popocatepetl volcano to identify different stages of its eruptive activity. We use packages 'ripa', 'fractal' and 'tseriesChaos' of "R" for data analysis.
Results allow us to identify the most relevant changes in volcano's behaviour and characterise the type of process corresponding to its signals. While this methodologies are well known in photography, music and energy industry, their application for investigating volcanoes represents a novel contribution.
|