Imperial College London


The Brazilian Amazon is a global epicentre of biodiversity, crucial for regulating the Earth’s climate. However, its harsh environment inhibits consistent, close monitoring of its ecosystems’ health. Therefore, there is a need to improve methods of monitoring bioindicators within the rainforest, to allow for earlier interventions against degradation such as pollution and deforestation.

The Amazon Multichannel Autonomous Recording Unit (AMARU) provides an accessible, robust method of enabling ecoacoustic monitoring within the Amazon rainforest. By engaging with local rainforest communities, long-term deployment has been achieved, where limited internet connectivity and solar coverage otherwise make Amazon-based deployments difficult.

The AMARU also offers improvements to existing bird call detection algorithms. This improvement was achieved owing to the AMARU’s six-microphone array, which enables acoustic beamforming: a technique that enhances sounds from directions of interest, whilst reducing background noise. Studies showed that beamforming significantly improved the precision and recall rate of BirdNET (a commonly used bird vocalisation detection algorithm).

Related Project

Fully automated biodiversity monitoring in the tropical forests of Borneo

Safe Acoustics

find out more

Related Publications

No items found.