Calibration pole distance

[NATHAN THAVARAJAH]

Good morning, 

I am currently analysing images in Agouti and notice that I have a number of sequences where animals are moving in shot some distance beyond the distance that has been calibrated. 

Please can you advise on whether these sequences are still okay to include for further analysis? In the training videos it is highlighted that one should calibrate as far back from the camera that the area is clear and an animal might be expected to be caught on the camera. If this is not done and an animal is caught beyond the distance of the stick, is Agouti able to extrapolate in the image based on the calibration information is has and if so, would there be a loss of accuracy in distance measure compared to animals tracked within the distance of the pole?

I am mainly checking at this point as we have another deployment this year and I wonder if we need to ask that people are more mindful of how far from camera they are calibrating.

Many thanks

Nathan

[Marcus Rowcliffe]

Hi Nathan

The calibration model can extrapolate to some degree but accuracy is likely to decline as you move beyond the calibrated area, so I would recommend that you reinforce training to ensure good coverage of calibration poles.

Best

Marcus

[NATHAN THAVARAJAH]

Hi Marcus,

Many thanks for all of your help, i realise i have had a few questions. With further regards to this one, for sequences that have been tagged and tracked but fall outside of the calibrated area, are these flagged during REM workflow in R and removed for analysis of speed and detection zone, or is it a case of flagging these during analysis in Agouti and not tracking these sequences?

Thanks again

Nathan

[Marcus Rowcliffe]

Hi Nathan

It is not (currently) possible to identify sequences or points beyond the calibrated area at the R stage, so any filtering has to be done through decisions to track or not at the Agouti stage.

Best

Marcus