The Bird Was Hit in Mid-flight. It's Not Because It Wasn't Paying Attention, but Because the Sun Was Too Dazzling.

The Argonne National Laboratory, a subsidiary of the U.S. Department of Energy, is developing technology to identify and monitor the interactions between birds and solar facilities. This project will have positive implications for bird protection and the planning and layout of solar facilities.

According to statistics, billions of birds die every year in the world, among whichNearly one billion were killed in the collision.

In April 2019, a report from the Cornell Lab of Ornithology at Cornell University in the United States pointed out thatApproximately 600 million birds die each year from flying into tall buildings in the United States.

But this result is not caused by the poor eyesight of birds, but because the glass curtain walls, glass windows and plants placed inside the buildings can cause flying birds to "go astray" and collide with the buildings.

However, few people know thatThere is another very lethal killer of birds - solar equipment.

Solar power kills tens of thousands of birds every year in the US

Given the advantages of solar energy such as being pollution-free, having large reserves and being renewable, people will adopt solar energy facilities more frequently.

But as more solar systems are installed, people are beginning to pay attention to their impact on wildlife and the environment.For example, the mirror reflection of solar energy may affect the vision of wild animals and birds.

Argonne National Laboratory published a study in 2016 that estimated the number of bird collisions with U.S. photovoltaic panels based on limited data.

Among them, the data shows thatAll AmericaUtility-scale solar facilities kill between 37,800 and 138,600 birds per year.

Note: Due to uncertainty regarding the causes of bird deaths, the estimated figures range widely

While this number is relatively small compared to the hundreds of millions of birds that die each year from building and vehicle strikes, if left unchecked, the number of birds dying from solar energy will increase as solar energy facilities increase around the world.

Assessment of U.S. solar energy potential

Therefore, a deeper understanding of how and when these deaths occur can help people effectively prevent them. On the one hand, this is to protect birds; on the other hand, reducing the activities of birds around solar facilities,It will also greatly reduce the cleaning work of staff.It's a win-win situation.

Using computer vision to quantify impact

To assess the interaction between birds and solar energy, scientists are quantifying the mutual impact of each, but current methods of manual data collection are time-consuming and labor-intensive.

Recently, the Argonne National Laboratory under the U.S. Department of Energy announced a solution.Using computer vision and artificial intelligence to monitor bird interactions with solar infrastructure.

Specifically, Argonne's new project will combine computer vision technology and artificial intelligence models toSupervisorObserve the activity of birds over solar panels, collecting data as they fly over, perch on, or collide with solar panels at a lower cost than humans.

“Collecting all this information requires people to be out in the field, walking around in facilities,Searching for bird carcasses is very time-consuming, labor-intensive, and very costly.”

Moreover, the human-generated collection method is also limited in frequency and span, making it difficult to obtain behavioral data on birds around solar panels.

"While there has been speculation about how solar infrastructure might affect bird populations,But we need more data to understand scientifically what’s going on.”

Argonne National Laboratory is one of the largest and oldest scientific research institutions in the U.S. government. The three-year bird-solar interaction monitoring project has been funded by the U.S. Department of Energy's Solar Energy Technologies Office with $1.3 million and was launched this spring.

Identifying birds is more complicated than identifying drones

According to reports, achieving this project goal involves three tasks:

• Detect moving objects near solar panels;
• Identify which objects are birds;
• Classify the event (such as perching, flight, or collision).

Scientists will also use deep learning to build models to identify birds and their behaviors.

In an earlier project at Argonne, researchers trained computers to recognize drones flying overhead.

Adam Szymanski, software engineer and developer of the drone detection model, said:The project of birds interacting with solar energy will be based on the technology used to identify drones, but will be more complex.

For example, a solar device’s camera will be pointed toward the panels, not upward, so the background will be more complex, and the system will need to distinguish between birds and other moving objects in the field of view, such as clouds, insects or people.

Initially, the project will involve installing cameras at one or two solar farms to record and analyze video. To train the computer model, hours of video will need to be manually processed and categorized.

Additionally, because collisions are relatively rare, the researchers say they could be simulated with objects like toy birds so that the system has initial information that can be used as training examples.

Once the model is trained,It runs live video from inside the camera, classifying bird-solar interactions as it flies.This, of course, is another challenge involving edge computing, in which information processing centers are located closer to where the data is collected.

Szymanski added: “We don’t have the financial resources to record a lot of videos and send them back to the lab for analysis. So, we have to design a more efficient model so that it can be executed in real time at the edge.”

Far-reaching significance: protecting more wildlife

According to reports, in addition to detecting, monitoring and reporting bird activities around solar facilities, the projectThe system also notifies solar facility staff when a collision occurs.

These results will help people better understand the interaction between birds and solar energy.It can assist in the siting of solar facilities and reduce impacts on wildlife.

Hamada said the technology will then be ready for more large-scale field trials at solar facilities, where data collected can be used to test whether all aspects of the model are set up optimally..

At the same time, these data will also answer questions such as:

Are certain types of birds more vulnerable to attack at solar facilities?

Do collisions increase at certain times of the day or year?

Does the geographic location of solar panels play a role in the various interactions?

Can solar facilities provide viable habitat for birds?

In addition, the technical framework can also be used to monitor other wildlife by retraining the AI with appropriate data.

“Once the patterns are identified, this knowledge can be used for wildlife conservation programs,” Hamada said.

-- over--