Hey MarMam!
We are excited to announce that the Xcertainty R package is now available on
CRAN!
Xcertainty is an easy-to-use R package that uses a Bayesian approach for
predicting photogrammetric uncertainty in morphometric measurements of marine
mammals derived from drones.
The easiest way to install Xcertainty is via CRAN:
install.packages("Xcertainty")
library(Xcertainty)
Learn more:
GitHub: https://github.com/MMI-CODEX/Xcertainty
CODEX website:
https://mmi.oregonstate.edu/centers-excellence/codex/software-hardware/xcertainty
CRAN: https://cran.r-project.org/web/packages/Xcertainty/index.html
All morphological measurements derived using drone-based photogrammetry are
susceptible to uncertainty. This uncertainty often varies by the drone system
used. Thus, it is critical to incorporate photogrammetric uncertainty
associated with measurements collected using different drones so that results
are robust and comparable across studies and over long-term datasets.
The Xcertainty R package makes this simple and easy by producing a predictive
posterior distribution for each measurement. This posterior distribution can be
summarized to describe the measurement (i.e., mean, median) and its associated
uncertainty (i.e., standard deviation, credible intervals). The posterior
distributions are also useful for making probabilistic statements, such as
classifying maturity or diagnosing pregnancy if a proportion of the posterior
distribution for a given measurement is greater than a specified threshold
(e.g., if greater than 50% of posterior distribution for total body length is >
10 m, the individual is classified as mature).
Xcertainty is based off of previously published Bayesian statistical models. In
essence, measurements of known-sized objects (‘calibration objects’) collected
at various altitudes are used as training data to predict morphological
measurements (e.g., body length) and associated uncertainty of unknown-sized
objects (e.g., whales).
Xcertainty also includes functions that incorporate multiple measurements (body
length and width) to estimate different body condition metrics (i.e., single
widths, surface area, body volume, body area index) with associated
uncertainty, as well as combine body length with age information to construct
growth curves
Cheers,
KC Bierlich & Josh Hewitt
KC (Kevin) Bierlich, PhD, MEM
Assistant Professor Senior Research
Center of Drone Excellence
(CODEX<https://mmi.oregonstate.edu/centers-excellence/codex>)
Marine Mammal Institute,
Dept. of Fisheries, Wildlife, & Conservation Sciences,
Oregon State University
Pronouns: he, him, his
[email protected]<mailto:[email protected]>
_______________________________________________
MARMAM mailing list
[email protected]
https://lists.uvic.ca/mailman/listinfo/marmam
