ARBIMON (Automated Remote Biodiversity Monitoring Network) is a project developed out of the University of Puerto Rico by Carlos Corrada-Bravo, Mitchell Aide and their team that uses off-the-shelf technology in combination with advanced machine-learning algorithms to help analyze an area's biodiversity.
Traditionally, biodiversity is assessed using a variety of methods that are generally costly, limited in space and time, and most importantly, they rarely include a permanent record.....Long-term information is needed to understand the implications of land and climate change on biological systems. From both a conceptual and management perspective there is an urgent challenge to increase biological data collection over large areas and through time. - Full Article
At the core of the project are field monitoring stations that are created using solar panels, a 2nd generation iPod, 12v car battery, 900 MHz Radio/antenna, and a 20Hz-20kHZ microphone all wrapped up tightly in a waterproof case. The field systems capture one minute long recordings of an area every 10 minutes and then passes that data along to a base station (that can be located up to 40 kilometers away) to eventually reach the main projects server for analysis and backup.
Once captured and stored the audio is processed against an automated species identification system and algorithms that has been trained by researchers based on specific vocalization of a species.
The ARBIMON team has created and tested species-specific models for a range of amphibians, birds, mammals, and insects. Once created these models can be can be used to process previously recorded audio to generate daily and monthly patterns for a specific species or a "soundscape index" for an entire area that can be tracked over time.
You can check out a more detailed overview of the ARBIMON system at Arbimon.com/arbimon/ or read their latest research findings on the project at PeerJ.
Related: Shoal Pollution Monitoring, Acoustic landslide detection