AI and bird population studies are helping scientists research how birds respond to fires and measures to help them. Scientists in California conducted a study of bird songs to determine how birds responded to wildfires in the Sierra Nevada mountain range and to determine which measures helped the birds recover faster. Researchers are currently analyzing a million hours of audio to gain insight into how birds responded to the fires and to find out which measures helped the birds recover.
The soundscape can also be used by scientists to analyze shifts in the timing of migration and to identify global patterns in population ranges, according to a recent news item in Scientific American. Increasingly, audio data is coming in from other projects as well, including a sound-based project to count insects and study the effects of light and noise pollution on bird communities that is underway. This method has helped with AI and bird population studies.
“Audio data is a real treasure trove because it contains vast amounts of information,” stated ecologist Connor Wood, a Cornell University postdoctoral researcher, who is leading the Sierra Nevada project. “We just need to think creatively about how to share and access that information.” Artificial intelligence is assisting in this, with the latest generation of machine learning AI systems being able to identify animal species in their calls, and can process thousands of hours of data in less than a day.
In her role as assistant director of the Cornell Lab of Ornithology’s Center for Conservation Bioacoustics, Laurel Symes is researching acoustic communication in animals such as crickets, frogs, bats, and birds. As part of her research, she has compiled many months’ worth of recordings of long-horned grasshoppers known as katydids (vocal long-horned grasshoppers that make up an important part of the food web) in the rain forests of central Panama. This audio is full of patterns of breeding activity and seasonal population variation that are hidden in it, but analyzing it would take a lot of time.
“Machine learning has been the big game changer for us,” Symes stated to Scientific American.
Symes and three of her colleagues spent 600 hours identifying katydid species based on just 10 recorded hours of sound. Her team’s machine-learning algorithm, called KatydID, performed the same task while its human creators “went out for a beer,” Symes said.
As part of Wood’s team’s analysis of the Sierra Nevada recordings, he will use a popular sound-recognition system called BirdNET, which can recognize the sound of birds. The BirdNET web application was developed by Stefan Kahl, a machine learning scientist at Cornell’s Center for Conservation Bioacoustics and Chemnitz University of Technology in Germany. Scientists are also utilizing BirdNET in order to document the effects of light and noise pollution on bird songs at dawn in France’s Brière Regional Natural Park.
The bird calls of avian species are diverse and complex. “You need much more than just signatures to identify the species,” Kahl stated. Machine-learning systems can detect the differences between the songs of birds that have more than one song, as well as those that have regional dialects – the white-crowned sparrows of Washington State sound very different from their Californian cousins. “Let’s say there’s an as yet unreleased Beatles song that is put out today. You’ve never heard the melody or the lyrics before, but you know it’s a Beatles song because that’s what they sound like,” Kahl stated. “That’s what these programs learn to do, too.”
BirdVox Combines Study of Bird Songs and Music
BirdVox, a joint project of NYU’s Music and Audio Research Laboratory and Cornell’s Lab of Ornithology, applies music recognition to bird song. According to a blog post at NYU, BirdVox is investigating machine listening techniques for the automatic detection and classification of free-flying bird species based on their vocalizations.
BirdVox’s researchers are developing a network of acoustic sensors to monitor the migration patterns of birds in real time, in particular, determining the precise timing of each species’ passage.
At present, bird migration monitoring tools rely on information from weather surveillance radar, which provides insight into bird movement density, direction, and speed, but not the species migrating. A majority of human observations are collected during daylight hours, so they are of limited use for studying nocturnal migrations, the researchers noted.
In addition to these methods, AI and bird population studies can be helped with methods like automatic bio-acoustic analysis is able to yield species-specific information, and is scalable. The researchers noted that such techniques are relevant to understanding biodiversity and monitoring migrating species in areas with buildings, planes, communication towers and wind turbines.
Duke University Researchers Using Drones to Monitor Seabird Colonies
Another team from Duke University and the Wildlife Conservation Society (WCS) is using drones and deep learning algorithms to monitor large seabird colonies. According to a press release from Duke University, the team is analyzing more than 10,000 drone images of mixed colonies of seabirds that nest in the Falkland Islands, off Argentina’s coast, according to the Duke University press release.
There are the largest colonies of black-browed albatrosses (Thalassarche melanophris) in the world in the Falklands, and the second-largest colonies of rockhopper penguins in the world on the Malvinas, also known as the Falkland Islands. There are hundreds of thousands of birds breed on the islands in densely interspersed flocks.
Using deep-learning algorithms, the team reported that they were able to accurately identify and count the albatross species 97% of the time and the penguins 87% of the time. About 90% of the time, the automated counts continued to be within five percent of the manual counts.
“Using drone surveys and deep learning gives us an alternative that is remarkably accurate, less disruptive, and significantly easier. One person, or a small team, can do it, and the equipment you need to do it isn’t all that costly or complicated,” stated Madeline C. Hayes, a remote sensing analyst at the Duke University Marine Lab, who led the study.
Scientists used to count the number of species they could observe on a portion of the islands to monitor the colonies located on two rocky, uninhabited outer islands before this new method became available. These numbers would be extrapolated to get a population estimate for the colony. For better accuracy, counts often had to be repeated, a laborious process that could disturb the birds’ breeding and parenting behavior.
Over 10,000 individual images were captured by WCS scientists using a consumer drone. With image-processing software, Hayes created a large-scale composite visual. In this case, she then analyzed the image using a convolutional neural network (CNN), an AI approach that employs a deep-learning algorithm to analyze an image and count the objects it “sees” – two species of birds, in this case penguins and albatrosses. We compiled data on the total number of birds in colonies to come up with comprehensive estimates.
“A CNN is loosely modeled on the human neural network, in that it learns from experience,” stated David W. Johnston, director of the Duke Marine Robotics and Remote Sensing Lab. “You train the computer to pick up on different visual patterns, like those made by black-browed albatrosses or southern rockhopper penguins in sample images, and over time it learns how to identify the objects forming those patterns in other images such as our composite photo.”
Johnston, who is also associate professor of the practice of marine conservation ecology at Duke’s Nicholas School of the Environment, said the emerging drone- and CNN-enabled approach is widely applicable “and greatly increases our ability to monitor the size and health of seabird colonies worldwide, and the health of the marine ecosystems they inhabit.”
To Read the source articles and information mentioned in AI and bird population article here, please go to the following links –
Introduction: AI and bird population
AI and bird population studies are helping scientists research how birds respond to fires and measures to help them. Scientists in California conducted a study of bird songs to determine how birds responded to wildfires in the Sierra Nevada mountain range and to determine which measures helped the birds recover faster. Researchers are currently analyzing a million hours of audio to gain insight into how birds responded to the fires and to find out which measures helped the birds recover.
The soundscape can also be used by scientists to analyze shifts in the timing of migration and to identify global patterns in population ranges, according to a recent news item in Scientific American. Increasingly, audio data is coming in from other projects as well, including a sound-based project to count insects and study the effects of light and noise pollution on bird communities that is underway. This method has helped with AI and bird population studies.
“Audio data is a real treasure trove because it contains vast amounts of information,” stated ecologist Connor Wood, a Cornell University postdoctoral researcher, who is leading the Sierra Nevada project. “We just need to think creatively about how to share and access that information.” Artificial intelligence is assisting in this, with the latest generation of machine learning AI systems being able to identify animal species in their calls, and can process thousands of hours of data in less than a day.
In her role as assistant director of the Cornell Lab of Ornithology’s Center for Conservation Bioacoustics, Laurel Symes is researching acoustic communication in animals such as crickets, frogs, bats, and birds. As part of her research, she has compiled many months’ worth of recordings of long-horned grasshoppers known as katydids (vocal long-horned grasshoppers that make up an important part of the food web) in the rain forests of central Panama. This audio is full of patterns of breeding activity and seasonal population variation that are hidden in it, but analyzing it would take a lot of time.
Symes and three of her colleagues spent 600 hours identifying katydid species based on just 10 recorded hours of sound. Her team’s machine-learning algorithm, called KatydID, performed the same task while its human creators “went out for a beer,” Symes said.
As part of Wood’s team’s analysis of the Sierra Nevada recordings, he will use a popular sound-recognition system called BirdNET, which can recognize the sound of birds. The BirdNET web application was developed by Stefan Kahl, a machine learning scientist at Cornell’s Center for Conservation Bioacoustics and Chemnitz University of Technology in Germany. Scientists are also utilizing BirdNET in order to document the effects of light and noise pollution on bird songs at dawn in France’s Brière Regional Natural Park.
The bird calls of avian species are diverse and complex. “You need much more than just signatures to identify the species,” Kahl stated. Machine-learning systems can detect the differences between the songs of birds that have more than one song, as well as those that have regional dialects – the white-crowned sparrows of Washington State sound very different from their Californian cousins. “Let’s say there’s an as yet unreleased Beatles song that is put out today. You’ve never heard the melody or the lyrics before, but you know it’s a Beatles song because that’s what they sound like,” Kahl stated. “That’s what these programs learn to do, too.”
BirdVox Combines Study of Bird Songs and Music
BirdVox, a joint project of NYU’s Music and Audio Research Laboratory and Cornell’s Lab of Ornithology, applies music recognition to bird song. According to a blog post at NYU, BirdVox is investigating machine listening techniques for the automatic detection and classification of free-flying bird species based on their vocalizations.
BirdVox’s researchers are developing a network of acoustic sensors to monitor the migration patterns of birds in real time, in particular, determining the precise timing of each species’ passage.
At present, bird migration monitoring tools rely on information from weather surveillance radar, which provides insight into bird movement density, direction, and speed, but not the species migrating. A majority of human observations are collected during daylight hours, so they are of limited use for studying nocturnal migrations, the researchers noted.
In addition to these methods, AI and bird population studies can be helped with methods like automatic bio-acoustic analysis is able to yield species-specific information, and is scalable. The researchers noted that such techniques are relevant to understanding biodiversity and monitoring migrating species in areas with buildings, planes, communication towers and wind turbines.
Duke University Researchers Using Drones to Monitor Seabird Colonies
Another team from Duke University and the Wildlife Conservation Society (WCS) is using drones and deep learning algorithms to monitor large seabird colonies. According to a press release from Duke University, the team is analyzing more than 10,000 drone images of mixed colonies of seabirds that nest in the Falkland Islands, off Argentina’s coast, according to the Duke University press release.
There are the largest colonies of black-browed albatrosses (Thalassarche melanophris) in the world in the Falklands, and the second-largest colonies of rockhopper penguins in the world on the Malvinas, also known as the Falkland Islands. There are hundreds of thousands of birds breed on the islands in densely interspersed flocks.
Using deep-learning algorithms, the team reported that they were able to accurately identify and count the albatross species 97% of the time and the penguins 87% of the time. About 90% of the time, the automated counts continued to be within five percent of the manual counts.
“Using drone surveys and deep learning gives us an alternative that is remarkably accurate, less disruptive, and significantly easier. One person, or a small team, can do it, and the equipment you need to do it isn’t all that costly or complicated,” stated Madeline C. Hayes, a remote sensing analyst at the Duke University Marine Lab, who led the study.
Scientists used to count the number of species they could observe on a portion of the islands to monitor the colonies located on two rocky, uninhabited outer islands before this new method became available. These numbers would be extrapolated to get a population estimate for the colony. For better accuracy, counts often had to be repeated, a laborious process that could disturb the birds’ breeding and parenting behavior.
Over 10,000 individual images were captured by WCS scientists using a consumer drone. With image-processing software, Hayes created a large-scale composite visual. In this case, she then analyzed the image using a convolutional neural network (CNN), an AI approach that employs a deep-learning algorithm to analyze an image and count the objects it “sees” – two species of birds, in this case penguins and albatrosses. We compiled data on the total number of birds in colonies to come up with comprehensive estimates.
“A CNN is loosely modeled on the human neural network, in that it learns from experience,” stated David W. Johnston, director of the Duke Marine Robotics and Remote Sensing Lab. “You train the computer to pick up on different visual patterns, like those made by black-browed albatrosses or southern rockhopper penguins in sample images, and over time it learns how to identify the objects forming those patterns in other images such as our composite photo.”
Johnston, who is also associate professor of the practice of marine conservation ecology at Duke’s Nicholas School of the Environment, said the emerging drone- and CNN-enabled approach is widely applicable “and greatly increases our ability to monitor the size and health of seabird colonies worldwide, and the health of the marine ecosystems they inhabit.”
To Read the source articles and information mentioned in AI and bird population article here, please go to the following links –
Scientific American
Blog post at NYU
Press release from Duke University.
Share this: