Efficient Broad-scale Wildlife Monitoring

Submitted by atripp on Thu, 05/20/2021 - 08:35

The FWS Division of Migratory Bird Management is integrating remote sensing and machine learning technologies to improve safety, data quality, and efficiency of broad-scale migratory bird surveys. The Division uses manned Department of the Interior fleet aircraft to monitor migratory bird populations over vast regions of North America. Data from these annual monitoring programs are critical to sustainably manage hunting programs and inform conservation and other regulatory decisions. The need for species-level data for small-bodied birds, over very large geographic regions, and under tight phenological and regulatory time constraints has forced continued reliance on low-level aerial surveys with human observers. Remote sensing solutions for migratory bird surveys enable flight at higher altitudes and can improve personnel safety and data quality. However, conducting these surveys using high-resolution remote sensing technologies onboard manned fleet aircraft generates a tremendous data volume that cannot be manually processed in a cost-effective manner. The Migratory Bird Program is collaborating with other agencies that have overlapping wildlife monitoring needs, such as the Bureau of Ocean Energy Management and the U.S. Geological Survey, to improve the cost-efficiency of broad-scale wildlife surveys using remote sensing. First, the program works with academia and industry to foster computer vision and machine learning methods for automated detection, species identification, and counting of birds and other wildlife from imagery. The program is also working to integrate hardware and software on manned fleet aircraft to enable near-real-time, in-flight processing of data and robust handling of large data volumes in the field.

Observations of common eider (Sometaria mollissima) at Nantucket Shoals, Massachusetts, are marked and labeled in the imagery to train and validate deep learning models for automated bird detection and species classification. Photo credit: Mark Koneff, USFWS.

Author Name
Mark Koneff
Author Email