Lead Institution(s): Western Ecosystems Technology, Inc.
Project Lead: Riley Knoedler
Collaborator(s): Thomas Preybl
Focal Species: All Sea Ducks, Bufflehead (Bucephala albeola), Barrow’s Goldeneye (Bucephala islandica), Common Goldeneye (Bucephala clangula), Harlequin Duck (Histrionicus histrionicus), Long-tailed Duck (Clangula hyemalis), Common Eider (Somateria mollissima), King Eider (Somateria spectabilis), Spectacled Eider (Somateria fischeri), Steller’s Eider (Polysticta stelleri), Common Merganser (Mergus merganser), Hooded Merganser (Lophodytes cucullatus), Red-breasted Merganser (Mergus serrator), Black Scoter (Melanitta americana), Surf Scoter (Melanitta perspicillata)
Project Description: This project used deep learning computer vision methods to automatically detect, identify, and count sea ducks in aerial images. Project objectives included: 1) Identify and annotate (or ‘paint’) sea ducks in aerial images that were provided by SDJV; 2) Train computer vision models, including those using computer vision, to automatically detect, identify, and count sea ducks in aerial images; and 3) Create an end product of an R package that will deploy the computer vision system, as well as a Shiny Application (a user-friendly HTML-based interface) to easily deploy the system.
Interim Report FY21
Final Report
Data Products:
SeaDuckDetectorR is software designed to identify sea duck and other avian species in aerial photographs. To run the SeaDuckDetectoR Program, first download and extract SeaDuckDetectoR_Manual.zip to a location on your computer. Once extracted, open the “SeaDuckDectectoR_Manual.docx” file that will provide further installation instructions.
You will also need to download the SeaDuckDetectoR_0.1.0.zip file but you do not need to extract the contents of this file.