How do we account for recounts?
Essentially: the same animal/object can return to the video frame... how does BlueCounter's software take this into account in its total counts?
BlueCounter's software uses the counting metrics to take into account recounts: MaxN, MeanN, TotN and MinN are used by default, but you can request for other counting metrics when we are developing an ecological monitor for you.
Each counting metric can be better suited to counting one species than the other, deciding which counting metric to use can be done by consulting previous literature. The following resources could help you in your decision of which counting metric would be the best for your study:
Stobart B, Díaz D, Álvarez F, Alonso C, Mallol S, Goñi R (2015) Performance of Baited Underwater Video: Does It Underestimate Abundance at High Population Densities? PLoS ONE 10(5): e0127559. https://doi.org/10.1371/journal.pone.0127559
Stobart, Ben et al. “Performance of baited underwater video: does it underestimate abundance at high population densities?.” PloS one vol. 10,5 e0127559. 26 May. 2015, doi:10.1371/journal.pone.0127559
The ecological monitor also produces frame by frame data on what was identified for each frame of the video as a CSV file, which then can be exported to excel for any kind of data analysis you would like to conduct on it.
Additionally, a more advanced predictive model is under development at BlueCounter to enable the software to use previous positions of the animals to predict for recounts and take them into account while extracting data from the video.
Essentially: how accurately can BlueCounter's ecological monitors detect the right species off of the video and give the correct counts for each frame?
Identification accuracy largely depends on the amount of data used to train the software. This can be specified by you when we are making the ecological monitor for your project. Rule of Thumb: the more that data, the more accurate it is (generally).
After making your product, we will also provide a calculated percentage uncertainty which you can take into account for your study.