BirdWeather uses machine learning to detect and map different species of birds around the world. The platform continuously collects sound from active audio stations distributed across the globe and provides what is effectively an automated AI bird spotting map of the world.
Thousands of crowd-sourced audio stations around the world contribute audio data to BirdWeather. Using the BirdNET artificial neural network, a highly sophisticated machine learning algorithm, BirdWeather uses this captured audio data to accurately identify bird calls and songs. From this analysis BirdWeather is then able to automatically map the locations of recorded and identified bird species.
You can explore the results of the machine learning analysis of the audio data provided by thousands of audio stations on the BirdWeather interactive map. This map allows users to explore bird activity by location, species, and by date. Users of the map can also listen to recorded bird calls and songs, and watch live cam streams of selected nesting sites.
If you own a Raspberry PI and a USB Microphone or Sound Card you can also contribute to BirdWeather. Install BirdNET-Pi and you will be able to record and automatically identify bird songs and submit your results to the BirdWeather map.
Active 'twitchers' may also be interested in the eBird interactive map. eBird collects and documents data on bird distribution, abundance, habitat use, and trends. It has detailed information on more than 1,000 bird species around the world.
Select a species of bird on the eBird Status and Trends webpage and you can view an interactive map which shows the natural habitat of the selected bird. If you select the 'Weekly' option you can actually watch an animated map showing the species' relative abundance for every week of the year. This allows you to observe the migratory journeys undertaken by the selected species of bird over the course of the year.
Комментариев нет:
Отправить комментарий