Acoustic monitoring system for endangered Philippine raptors

Abstract

In ecosystems, avian species, especially apex predators such as the Philippine Eagle and Pinsker's Hawk Eagle, play a crucial role in regulating other species and fostering biodiversity. These bird species actively manage populations of smaller carnivores, thereby making significant contributions to the overall health and balance of the ecosystem. If these species disappear, there would be no suitable replacements, negatively impacting the ecosystem. Protecting and monitoring birds and their habitats are crucial for ecosystems and human well-being. Therefore, the researchers developed an acoustic monitoring system for endangered Philippine raptors using a recorder and a computer program. In a rainforest environment, specifically at the Philippine Eagle Center, the researchers employed a portable, battery-powered recorder to capture bird calls continuously. Subsequently, a computer program was utilized to automatically filter, detect, and classify these calls. The analysis results were exported into a spreadsheet for monitoring purposes. To predict bird species, delta features of Mel Frequency Cepstral Coefficients (MFCC) and Gammatone Cepstral Coefficients (GTCC) were utilized as features. The Cosine k-nearest neighbors (KNN) algorithm was employed for classification, resulting in an overall accuracy performance of 97.42%. Specific accuracy rates for each bird species were also achieved: Brahminy Kite (100.00%), Philippine Serpent Eagle (98.41%), Pinsker's Hawk Eagle (99.58%), Philippine Eagle (95.83%), and White-bellied Sea Eagle (91.86%).