Skip to content

You are using an outdated browser

Internet Explorer is not supported by this site and Microsfot has stopped releasing updates, therefore you may encounter issues whilst visiting this site and we strongly recommend that you upgrade your browser for modern web functionality, a better user experience and improved security.

Upgrade my browser

Research shows AI camera technology greatly reduces bird fatalities at wind farms

4 min read

Inspiring People
Research shows AI camera technology greatly reduces bird fatalities at wind farms
Source: NREL|Illumination|Duke Energy

Wind farms using IdentiFlight artificial intelligence (AI) camera technology can detect over 5 times more bird flights than human observers alone — the system is able to identify birds from a distance with an accuracy rate of 94%, and shut the turbines down before they approach.

Research shows IdentiFlight camera technology greatly reduces wind turbine eagle fatalities

New study by researchers from The Peregrine Fund finds that using IdentiFlight, an automated computer vision system that shuts off wind turbines when it detects eagles, can reduce eagle fatalities by 82%. The findings are published in Journal of Applied Ecology and reported by the British Ecological Society.

The wind site is the first to use IdentiFlight, a network of cameras and software that can detect golden and bald eagles (shown here) in flight and stop the turbines from spinning before the birds get too close.
Duke Energy Sustainable Solutions’ Top of the World Windpower Project, Glenrock, Wyoming. The wind site is the first to use IdentiFlight, a network of cameras and software that can detect golden and bald eagles (shown here) in flight and stop the turbines from spinning before the birds get too close. Source: Illumination/DukeEnergy

Use of the new IdentiFlight system resulted in an 82% reduction of eagle fatalities

Wind energy is an important technology as society moves toward identifying green sources of energy. However, points out a Peregrine Fund press release published by the British Ecological Society, wind turbines can cause a negative impact on local wildlife, including bird and bat fatalities, which occur when they collide with rotating turbine blades.

In a recently released study, “Eagle fatalities are reduced by automated curtailment of wind turbines,” in the Journal of Applied Ecology, researchers from The Peregrine Fund, Western EcoSystems Technology, Inc., and the U.S. Geological Survey, tested an automated computer vision system called IdentiFlight. This system is designed to use cameras and machine learning to detect eagles near turbines, so turbines can be shut off, reducing fatalities of eagles at wind energy sites. Use of the new IdentiFlight system resulted in an 82% reduction of eagle fatalities.

IdentiFlight is an automated curtailment system consisting of cameras and software programed for detecting flying objects, classifying them, and stopping turbines from spinning when an eagle is at threat of collision. The study to test IdentiFlight was performed at Duke Energy Renewables’ Top of the World Windpower facility in Wyoming, USA and included before and after studies as well as control and treatment sites.

The research determined that the number of fatalities at the treatment site declined by 63% between the before and after periods of testing IdentiFlight while fatality numbers increased at the control site by 113%. These estimates are corrected for seasonal and site-specific variation in rates of detection of eagle carcass and rates of removal of carcasses by scavengers.

Dr. Chris McClure, Director of Global Conservation Science at The Peregrine Fund and lead author on this study, concludes, “These results show that using the IdentiFlight system can lessen numbers of fatalities of eagles at wind energy facilities, reducing the conflict between wind energy and raptor conservation.” 

Read in full the Peregrine Fund press release published by the British Ecological Society.

Source: British Ecological Society

When they detect an object, the software, powered by artificial intelligence and machine learning, determines if it's an eagle and tells the turbines to stop spinning. An independent study in 2020 showed an 82% reduction in eagle fatalities at Top of the World, Wyoming.
IdentiFlight’s cameras constantly scan the sky. When they detect an object, the software, powered by artificial intelligence and machine learning, determines if it’s an eagle and tells the turbines to stop spinning. An independent study in 2020 showed an 82% reduction in eagle fatalities at Top of the World, Wyoming. Source: Illumination/DukeEnergy

IdentiFlight protects species via targeted curtailment of turbines

5-Step Journey to Curtailment

1. Detection

A bird is flying and comes within range as the IdentiFlight constantly analyses wide-field images with its eight camera array,  detects movement and points the stereo cameras at the bird, locking onto it. 

2. Classification

IdentiFlight determines the distance of the bird with the stereo cameras then classifies it based on a neural network and the bird’s physical characteristics. The system then sends the data to a local database and the Base Station. Note: These first two steps occur in milliseconds!

3. Data Delivery

The base station sends the data to the graphical display on site, the curtailment module for analysis, and to the IdentiFlight Dashboard for real-time operational and data analysis.

4. Flight Analysis

The curtailment module analyses bird flight path in relation to wind turbines and determines whether to curtail the turbine. Curtailment may be triggered by proximity, trajectory, or flight characteristics. 

5. Curtailment

If the curtailment module detects a bird is in danger it will send off a curtailment signal to customer SCADA Network via OPC or other means. Additionally, the IdentiFlight Dashboard reports turbine curtailments as well as imagery that can be analysed by qualified individuals to determine if the correct decision was made.

Source: IdentiFlight

This is accomplished using a proprietary hybrid of machine vision and artificial intelligence technologies. IdentiFlight also implements machine learning via convolutional neural network technology. In this technology, millions of images gathered from IdentiFlight’s data set are used to train a neural network. Leveraging the convolutional neural network data allows the system to continuously learn and improve as the data set grows.
Blending AI, Machine Learning and High-Precision Optics: the overwhelming advantage of IdentiFlight over other approaches is the ability to determine species in real time. This is accomplished using a proprietary hybrid of machine vision and artificial intelligence technologies. IdentiFlight also implements machine learning via convolutional neural network technology. In this technology, millions of images gathered from IdentiFlight’s data set are used to train a neural network. Leveraging the convolutional neural network data allows the system to continuously learn and improve as the data set grows. Source: IdentiFlight
The IdentiFlight towers can be positioned to cover multiple turbines in a single wind farm. When installed as a network with overlapping aerial coverage, the systems work together to provide the most protection possible for avian activity in the area.
The system detects a bird as far as one kilometre away, classifying it as a protected species such as an eagle (or not) in real time. The IdentiFlight towers can be positioned to cover multiple turbines in a single wind farm. When installed as a network with overlapping aerial coverage, the systems work together to provide the most protection possible for avian activity in the area. Source: NREL

The IdentiFlight system blends artificial intelligence with high-precision optical technology to detect eagles and other protected avian species. Proprietary software and neural network technologies process the images to determine 3D position, velocity, trajectory, and protected species of interest, all within seconds of detection. IdentiFlight towers operate as an autonomous system detecting, classifying, and curtailing specific turbines that could pose a risk to the bird. Source: Vimeo/IdentiFlight
Make an Impact

DEBUNKING THE 4 MAIN MYTHS ABOUT WIND ENERGY FOR BETTER UNDERSTANDING

While stories of bird fatalities at wind farms are actually true, the problem is nowhere near as bad as some people claim. While wind energy might be the simplest renewable energy to understand, there remain some major misconceptions about the wind industry—here we try to uncover the truth so you can make informed judgments for yourself.