There is a phenomenon where from one flight to the next, the location information seems to be shifting relative to the objects in the image (i.e. fixed nests are sliding to new coordinates from weekly image to the next).
Examples can be seen by following these steps:
Checkout the exp-nest-detect-imagery branch of the EvergladesTools repo
Run server.R in to load the Shiny app
In the app, click "Predicted Nests"
On the Predicted Nests page, select Site = JetportSouth, Year=2020, SampleId=1
the image has a yellow circle. @ethanwhite set this circle to stay at a fixed location across the time series of images. In this first image, there is a nest inside the circle. Above the image, there is a toggle that says "Select Date". Slide that toggle and watch the bird jump. If you use the thick branch pointing to at 9'oclock it can serve as a good reference point for seeing how the circle is jumping from one nest to the other.
Other examples include:
Site Year ID
JetPortSouth 2020 11
JetPortSouth 2020 19
JetportSouth 2021 22 (this one is pretty good most of the time, but jumps notably on 3/08)
Joule 2021 41
Joule 2021 41 (date 3/12 is particularly notable)
I can provide more examples but this is a good starting point.
There is a phenomenon where from one flight to the next, the location information seems to be shifting relative to the objects in the image (i.e. fixed nests are sliding to new coordinates from weekly image to the next).
Examples can be seen by following these steps:
Other examples include: Site Year ID JetPortSouth 2020 11 JetPortSouth 2020 19 JetportSouth 2021 22 (this one is pretty good most of the time, but jumps notably on 3/08) Joule 2021 41 Joule 2021 41 (date 3/12 is particularly notable)
I can provide more examples but this is a good starting point.