Artificial Intelligence/Machine Learning

We have a unique set of skills applying machine learning and artificial intelligence to clients’ natural resource questions. WEST has a large staff of statisticians, computer coders, database experts and administrators, and biologists who collectively can design studies and collect necessary data, manage large datasets, and develop machine learning neural networks as well as application-specific software solutions.

Automated Wildlife Detection and Remote Sensing

  • Many wildlife species are cryptic by nature. While keeping a low profile can help wildlife avoid potential predators, this also makes it difficult for field biologists to monitor wildlife populations. Monitoring wildlife populations or their habitat conditions in person often requires repeated visits to multiple sites covering large study areas, imposing cost and logistical constraints on monitoring efforts. WEST scientists are active in pioneering new methods to detect and monitor wildlife and aspects of their environment using remote cameras, aerial or satellite imagery, environmental DNA, drones, passive acoustic recorders, and other field-deployed sensors. These solutions can increase the amount of information wildlife managers can gain with limited budgets and often shed light on previously unobserved or unobservable biological phenomena.
 
Data-Driven Pattern Recognition

  • Pattern-driven recognition is one of the most innovative and disruptive technologies of the last 20 years. While most pattern recognition tasks entail some element of human activity, WEST is one of the first firms to recognize the amazing utility of these technologies in wildlife applications. WEST is truly a worldwide leader in customizing neural network architectures to detectors and classifying wildlife in still and video photography.

Machine Learning Applications Experience

  • Dolphin Identification From Fin Characteristics
  • Bird and Bat Carcass Detection at Renewable Projects Using Drones and Machine Learning
  • Surveys Using Drones and Infrared Cameras
  • Bird and Bat Collision Detection System for Offshore Wind Energy
  • Sagebrush Cover Modeling
  • Smart Curtailment Algorithms for Reducing Bat Mortality
  • Multi-Use Unmanned Aerial Systems Surveys for Environmental Compliance at Solar Projects
  • Environmental eDNA Applications
  • Sampling Design and Data Collection Protocol Development
  • Biological Data Collection
  • Machine Learning Classification System Development and Coding
  • Software Application Development

Publications

Automated classification of bat echolocation call recordings with artificial intelligence
Michael A. Tabak
, Kevin L. Murray, A. M. Reed, J. A. Lombardi, and Kimberly J. Bay 
2021 

Remote Sensing of White-Tailed Prairie Dogs in the Pinedale Anticline Project Area Using Aerial Imagery and Artificial Intelligence
Jason D. Carlisle
, Joel Thompson, Terri Harvey, Kristen Klaphake, Ryan Anderson, Chad W. LeBeau, and T. L. McDonald
2020

Effects of tones associated with drilling activities on bowhead whale calling rates
S. B. Blackwell, C. S. Nations, A. M. Thode, Mandy E. Kauffman, A. S. Conrad, and R. G. Norman
2017

Project Highlights

Ready to connect with us? Let's work together.

Aleutian Tern Population Estimates

The Aleutian tern is an uncommon seabird that annually migrates between nesting areas in coastal areas of Alaska and wintering grounds in Polynesia, Thailand, and Korea. At least one study has documented precipitous declines in Alaska of nesting birds between 1960 and the present. In 2017, a multi-stakeholder group composed of the US Fish and Wildlife Service, US Forest Service, Alaska Department of Fish and Game, National Fish and Wildlife Foundation, Audubon, Oregon State University, Conservation Metrics, and WEST initiated a series of studies to better understand population levels of the Aleutian tern in Alaska. 

Estimating Aleutian tern population sizes is difficult due to their ephemeral nesting colonies and patchy distribution. In 2018 and 2019, we used low-altitude photography obtained by unmanned aerial systems to obtain density and abundance estimates using a computer-assisted photo recognition routine that counted both Aleutian tern and Arctic tern at six Aleutian tern colonies in coastal regions of southern Alaska. We have documented the utility of the unmanned aerial systems method and are working toward statewide surveys.