Has anyone run thermal surveys with a Mavic 3T for feral hog abundance in mixed oak scrub? I’m modeling detection probability by canopy closure and converting detections per km to density; over 14 dawn flights at 50 m AGL we averaged 0.23 groups/km, but estimates swing when I adjust g(0). How do your counts compare to helicopter FLIR, and has the drone workflow informed removal prioritization on the ground?
At 50 m AGL, run opposite-angle passes; mark–recapture detections to estimate ‘g(0)’. Matched our helicopter FLIR within about 12%.
@OP Try pre-dawn by 30–45 minutes; mixed oak heat lags, boosting ‘g(0)’.
@OP I’ve bumped g(0) in oak scrub by flying one nadir transect and a quick oblique re-pass (gimbal about 10–15°) with 5–8 s hovers to verify in RGB; that put our M3T counts within about 15% of a helicopter FLIR on the same block. The oblique is like angling a flashlight under a porch, but watch warm rocks — split-screen check cuts false positives. What airspeed are you running?
Your 0.23 groups/km at 50 m AGL tracks us, but g(0) quit see-sawing when we locked the M3T isotherm/contrast (high‑gain, white‑hot) and did a manual NUC every about 2 min — auto adjust was smearing targets, . With that, we were within about 12% of the aircrew in the same blocks, though breezy leaves still hurt under dense oak. Are you running fixed isotherms or letting it float, @ryan_jk87?