Deep Learning for Automated Detection and Classification of Waterfowl, Seabirds, and other Wildlife from Digital Aerial Imagery | USGS 19 FA 0203

Opportunity Information: Apply for USGS 19 FA 0203

The grant opportunity titled "Deep Learning for Automated Detection and Classification of Waterfowl, Seabirds, and other Wildlife from Digital Aerial Imagery" is a Department of the Interior, U.S. Geological Survey (USGS) cooperative agreement focused on modernizing how marine and coastal wildlife are surveyed using aerial photos. The core idea is to develop and improve deep learning and computer vision methods that can automatically find and identify animals such as waterfowl, seabirds, and other marine wildlife in digital aerial imagery, reducing the heavy manual effort typically required to review and annotate large image collections.

This effort is being jointly supported and coordinated by multiple federal partners with overlapping missions in wildlife monitoring and offshore environmental stewardship. The Bureau of Ocean Energy Management (BOEM) is a primary funder and has prioritized the use of Outer Continental Shelf Program funds through USGS across fiscal years 2019, 2020, and 2021 to push forward two key building blocks: (1) expanding and refining an imagery and annotation database, and (2) advancing deep learning algorithms (DLA) that can use those labeled images to automate detection and classification. The U.S. Fish and Wildlife Service (USFWS), specifically the Division of Migratory Bird Management (DMBM) and its Branch of Migratory Bird Surveys, is also collaborating, reflecting the practical need for accurate bird survey tools that can support population assessments, migration monitoring, and conservation decision-making. The partnership framing signals that the project is not just a research exercise, but intended to produce methods and resources that can be applied in real survey programs and environmental assessments.

The technical scope centers on applying deep learning to the challenging conditions found in aerial wildlife imagery, where animals can appear small, partially obscured, clustered, or visually similar across species, and where lighting, sea state, glare, altitude, and sensor differences can all affect image quality. By investing in improved training data (imagery plus high-quality annotations) and algorithm development, the project aims to strengthen automated workflows that can scale to large geographic areas and long time series of flights. In practical terms, success would mean survey teams can process imagery faster, more consistently, and potentially with measurable accuracy metrics, enabling broader monitoring coverage and quicker turnaround for management needs.

From an administrative standpoint, this is a discretionary funding opportunity offered as a cooperative agreement, meaning the federal partners expect to have substantial involvement during the project rather than simply issuing a hands-off grant. The opportunity number is USGS 19 FA 0203, and it falls under the Natural Resources funding activity category with CFDA number 15.808. Eligibility is limited to public and state-controlled institutions of higher education, positioning the work for universities that can combine ecological expertise with machine learning, remote sensing, and data management capabilities. The maximum award amount listed is $85,000, with one expected award, indicating a targeted, single-recipient project rather than a broad multi-award competition. The posting lists a creation date of June 7, 2019, and an original closing date of June 17, 2019, which implies a short application window typical of some specialized cooperative research calls.

Overall, the opportunity is aimed at building practical deep learning tools and supporting datasets that help federal agencies detect and classify birds and other wildlife from aerial imagery more efficiently. The broader value is tied to improving environmental monitoring and decision support, particularly in offshore and coastal contexts where BOEM, USFWS, and USGS rely on accurate wildlife distribution and abundance information for stewardship, planning, and impact assessment.

• The Department of the Interior, U. S. Geological Survey in the natural resources sector is offering a public funding opportunity titled "Deep Learning for Automated Detection and Classification of Waterfowl, Seabirds, and other Wildlife from Digital Aerial Imagery" and is now available to receive applicants.
• Interested and eligible applicants and submit their applications by referencing the CFDA number(s): 15.808.
• This funding opportunity was created on Jun 07, 2019.
• Applicants must submit their applications by Jun 17, 2019. (Agency may still review applications by suitable applicants for the remaining/unused allocated funding in 2026.)
• Each selected applicant is eligible to receive up to $85,000.00 in funding.
• The number of recipients for this funding is limited to 1 candidate(s).
• Eligible applicants include: Public and State controlled institutions of higher education.

Apply for USGS 19 FA 0203

[Watch] Creating a grant proposal using the step-by-step wizard inside the applicant portal: