Hi everyone! Check out this talk about reporting methods for drone wildlife research. @David Johnston is a co-author here. Thanks Dave!! https://bb.xcdcdn.com/file/xcdvideos/AQRG_2674483.webm_new.mp4?nocache=210113183107[…]nal_members&utm_content=This+Week%27s+eWildlifer+%26+TWS+Talks

Hi ! I know this channel has been dormant for a while, but I wanted to message folks interested in AI and seabirds in general about a workshop/symposium I am putting together for the World Seabird Conference in Hobart Tasmania, September 2026. This is my current description:

Seabird and waterbird colony monitoring programs around the globe are rapidly accumulating aerial datasets from drones, satellites, and automated sensors, but refined protocols lag behind, limiting comparability and slowing adoption of new tools across the regions. This session will bring together long-standing colony monitors and experts in applications of aerial imagery and machine learning in ecology to share successes, pitfalls, and future plans in integrating aerial imagery and machine learning into their work. The symposium will provide an opportunity for participants to learn about available tools and current work in this field. The workshop will then provide an opportunity for those interested to collaboratively draft best-practice guidelines to ensure robustness, accessibility, reproducibility, and cost-effectiveness of next-generation colony monitoring efforts.

If you're interested in being involved in the planning, I would love more input, assistance, presenters, or additional co-leads!

Is Barry Brook (or others from his lab) @ U Tas already involved? I have no idea whether he specifically works on birds or aerial surveys, but his group has done great work not only on AI for camera trap model-y machine-learning-y stuff, but on building tools that make model fine-tuning more accessible:

https://github.com/zaandahl/mewc

So even if they're not working on birds or aerial images, there might be some best practices to exchange about tools and fine-tuning, and maybe even a direct analogy to MEWC (the tool I linked to above), which helps users fine-tune models on MegaDetector crops; you and I chatted about allowing similar fine-tuning downstream of a hypothetical universal bird detector. Basically since he's local, it seems like it's a no-brainer to invite folks from his group. He hangs out on this Slack, though I don't know whether he monitors this channel.

Hello All,

Myself and my colleagues Richard Gibbons (Audubon Texas) and Krish Kabra (Rice University) have been working on the side for the last few years to develop open-source machine learning tools for digitizing birds in aerial imagery. We started entirely focused on the islands along the Texas coast and successfully published both a detection tool and a species identification tool back in 2022. We've been working on improving that tool since and are wrapping up that manuscript now.

We've also realized through these efforts and through conversations with others that the average seabird researcher using aerial imagery really needs to start with a well-trained detector tool - one that can auto-generate bounding boxes around each bird in the imagery, regardless of species (this is all some folks need if they're working in single-species colonies and it is a good first step for those working in mixed-species colonies). We're now gearing up to train a detection model on birds from a variety of different habitat types, species, and image types in an effort to make a "universal" bird detector readily available and easy-to-use for anyone who needs it. To do this, we're asking partners from around the world to send sample images to us to support the training effort. Ideally we would have 200 annotations of each bird x habitat x image type (for example, 200 COMU on vegetation from a plane, 200 COMU on rock from a plane, 200 COMU on vegetation from a drone, etc.), so the more variety of images the better. If you have aerial images of seabirds you're willing to share with us for this effort, we will enthusiastically accept and list you as a contributor of images on the final tool (we are really lacking imagery outside of the Americas at this point).

If you have any questions or thoughts, please don't hesitate to reach out either here or by emailing me at annacvallery@gmail.com

arXiv.org

Deep object detection for waterbird monitoring using aerial imagery

Monitoring of colonial waterbird nesting islands is essential to tracking waterbird population trends, which are used for evaluating ecosystem health and informing conservation management decisions. Recently, unmanned aerial vehicles, or drones, have emerged as a viable technology to precisely monitor waterbird colonies. However, manually counting waterbirds from hundreds, or potentially thousands, of aerial images is both difficult and time-consuming. In this work, we present a deep learning pipeline that can be used to precisely detect, count, and monitor waterbirds using aerial imagery collected by a commercial drone. By utilizing convolutional neural network-based object detectors, we show that we can detect 16 classes of waterbird species that are commonly found in colonial nesting islands along the Texas coast. Our experiments using Faster R-CNN and RetinaNet object detectors give mean interpolated average precision scores of 67.9% and 63.1% respectively.

Original URL: https://arxiv.org/abs/2210.04868

Hi Again,

I don't know if anyone is still using this channel, but I wanted to share that I am hosting a workshop/symposium at the upcoming World Seabird Conference in Hobart, Tasmania, September 7-11, 2026 that I think folks in this workspace would be interested in!

Workshop W02: Next-generation seabird monitoring: from technology and machine learning to conservation best practice
Conveners: Anna Vallery (Audubon Washington, USA), Wade Sedgwick (Island Foundation, USA), Matt Wood (University of Gloucestershire, UK), Alice Edney (National Trust Scotland, UK)
Emerging technologies are transforming how we monitor seabirds, providing critical insights into responses to threats such as climate change, habitat loss, and other environmental changes. Seabird monitoring programs around the globe are rapidly accumulating large datasets from drones, satellites, cameras, and other automated sensors, but refined protocols for managing and analyzing these datasets lag behind, limiting comparability and slowing adoption of new tools across programs and regions worldwide. This session will bring together experts in applications of these tools and machine learning in ecology to assess how technology can best support evidence-based management, sharing successes, pitfalls, and future plans while balancing innovation with realism about the limitations of technology in the field. The symposium will provide an opportunity for participants to learn about available tools and current work in this field. By bridging cutting-edge science and policy needs, the workshop will then provide an opportunity for those interested to collaboratively draft best-practice guidelines to ensure robustness, accessibility, reproducibility, and cost-effectiveness of next-generation seabird monitoring efforts, strengthening conservation planning while ensuring seabird monitoring remains robust into the next decade.

Abstracts for the conference are due January 15th! I am happy to answer any questions. We hope to see some of you all there! https://worldseabirdunion.org/world-seabird-conference-home/

World Seabird Union

World Seabird Conference Home

world seabird conference 7–11 September 2026Hobart, Tasmania, Australia Register for WSC4 A Truly Global GatheringThe World Seabird Conference (WSC) is an in-person event that will also include a f…

Original URL: https://worldseabirdunion.org/world-seabird-conference-home/

Hi! My name is Ash and I'm a PhD researcher working in avian bioacoustics, based in NZ. I'd lovr to know if anyone is working on seabird audio in this group?