Topic status: We're looking for students to study this topic.In order to monitor and understand the rapid pace of environmental change due to global warming and human development, we have researched and deployed sensor networks in various sites which are ecologically interesting. These sensor networks have been collecting large volumes of data but placing much stress on the capacity of ecologists to store, analyse and visualise that data. The acquisition and analysis of acoustic recordings of the environment is one part of the great challenges of environmental monitoring.
Sound segmentation plays an important role in sound analysis. To date, common algorithms are based on different parameters such as short-term energy, zero crossings, frequency etc. All these algorithms have their own suitable scope and limitations. What parameters can accurately describe properties of sound, especially wildlife sounds, have not yet resolved.
The aim of this research is to develop new algorithms and techniques to extract most effective segmentation of the sound, so as to help ecologists understand all aspects of the acoustic environment via computational technologies.
Granular Computing is a computing paradigm of information processing using granules. It is mainly used to deal with imprecision, fuzzy, uncertainty, partial truth and massive information. It is a promising approach for massive sound data analysis and visualization. This project will therefore explore the applications of granular computing in sound analysis.