Application of an Empirical Mode Decomposition (EMD) detection and classification process to environments for Naval monitoring and detection
Abstract
This project will quantify the robustness of a new algorithm process and apply it to detecting and classifying marine mammal vocalizations to ultimately estimate species densities. This semi-blind detection and classification (SDC) process is based on Empirical Mode Decomposition (EMD) and will be tested on four datasets that vary by location, water depth, and recording instrumentation. The proposed research aims to achieve the following objectives. The first objective is to refine the EMD algorithms so they are robust to signals in shallow and deep water in multiple ocean basins on a variety of recording platforms. This will be achieved by testing EMD performance against human analysis from four various datasets. The next objective is to compare the more robust EMD algorithms performance against other detection/classification methods. This will be achieved by comparing the precision and recall values of EMD against two currently used software packages in the bioacoustics community. The third objective is to develop a cue rate (counting) metric to inform density estimation models for many species of interest using the more robust EMD algorithms. The final objectives are to calculate the detectable range of classified vocalizations with propagation modeling and then transfer all attained knowledge to a wider professional audience. Anticipated outcomes of this research that will positively impact DoD capabilities, if successful, include a new method for calculating cue rates and known detectability ranges for each species and/or vocalization type in real-world habitats similar to the four test datasets. These outcomes will be available to the U.S. Navy to use in their take permitting and in expanding their long-term monitoring and density estimation capabilities for marine mammals.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 31, 2020
- Source ID
- N000142012620
Entities
People
- Nicholas Kirsch
Organizations
- Office of Naval Research
- United States Navy
- University System of New Hampshire