article / 30 November 2021

Low-cost underwater camera trapping and deep learning

This study breaks ground in outlining a methodology for a system of low-cost, long-term camera traps (Dispersed Environment Acquatic Cameras) that can be deployed over large spatial scales in remote marine environments. Using machine learning to classify a large volume of collected images (over 100,000, collected over five months) and an ResNet-50-based deep learning model which achieved 92.5% overall accuracy in sorting images with and without fish, the study represents the first successful method for broad-scale underwater camera trap deployment. The methodology can be applied to address marine animal behavior, distributions, and large-scale spatial patterns.

Title: A low-cost, long-term underwater camera trap network coupled with deep residual learning image analysis

Authors: Stephanie M. Bilodeau, Austin W. H. Schwartz, Binfeng Xu, V. Paúl Pauca, Miles R. Silman

Journal: BioRxiv

Citation: Stephanie M. Bilodeau, Austin W. H. Schwartz, Binfeng Xu, V. Paúl Pauca, Miles R. Silman (2021). "A low-cost, long-term underwater camera trap network coupled with deep residual learning image analysis." BioRxiv, Department of Biology, Wake Forest University. https://doi.org/10.1101/2021.03.08.434472

Open access: Yes 

Abstract:

  1. Understanding long-term trends in marine ecosystems requires accurate and repeatable counts of fishes and other aquatic organisms on spatial and temporal scales that are difficult or impossible to achieve with diver-based surveys. Long-term, spatially distributed cameras, like those used in terrestrial camera trapping, have not been successfully applied in marine systems due to limitations of the aquatic environment.
  2. Here, we develop methodology for a system of low-cost, long-term camera traps (Dispersed Environment Aquatic Cameras), deployable over large spatial scales in remote marine environments. We use machine learning to classify the large volume of images collected by the cameras. We present a case study of these combined techniques’ use by addressing fish movement and feeding behavior related to grazing halos, a well-documented benthic pattern in shallow tropical reefscapes.
  3. Cameras proved able to function continuously underwater at deployed depths (up to 7 m, with later versions deployed to 40 m) with no maintenance or monitoring for over five months, and collected time-lapse images during daylight hours for a total of over 100,000 images. Our ResNet-50-based deep learning model achieved 92.5% overall accuracy in sorting images with and without fish, and diver surveys revealed that the camera images accurately represented local fish communities.
  4. The cameras and machine learning classification represent the first successful method for broad-scale underwater camera trap deployment, and our case study demonstrates the cameras’ potential for addressing questions of marine animal behavior, distributions, and large-scale spatial patterns

Key words:

Behavior, camera trap, image classification, long-term, machine learning, marine, underwater, reefscape, deep learning, landscape of fear


Add the first post in this thread.

Want to share your own conservation tech experiences and expertise with our growing global community? Login or register to start posting!