Four kinds of GNSS work,
each on its own page.
PosFlow serves drone and UAV mapping operators, control and deformation surveyors, CORS and reference-station operators, and geoscience researchers.
A day's flight produces thousands of images and a rover log, and there's no way to know the ground coordinates are trustworthy until the orthomosaic is already built on top of them.
Apex solves your rover log against your own base station log with PPK — kinematic, static, or static-start — and returns a trajectory with per-epoch quality flags before you tag a single photo.
What won't produce a result
Worth checking before you upload, so a run doesn't cost you credits to tell you what your data was missing.
Only the rover log, with no base station observation file from the same flight
A base station RINEX file recorded over the same time window. PPK is differential — it resolves the rover's position relative to the base, so a flight with no base log has nothing to solve against.
Rover and base sessions that don't share a recorded time window
Overlapping observation spans. The solution only covers the window the rover and base recorded together.
A benchmark or monitoring point needs a coordinate you can defend years later, not just one that looks stable today.
Static PPP resolves a long-session position on its own, cross-checked between two independent engines, with full antenna calibration and a formal uncertainty on the result.
What won't produce a result
Worth checking before you upload, so a run doesn't cost you credits to tell you what your data was missing.
A short or frequently interrupted observation session
A long, continuous session. PPP resolves ambiguities through convergence, and a short or broken one may never settle on a stable coordinate.
Single-frequency-only observations
Dual-frequency tracking, at minimum. The solution needs two frequencies to cancel the ionospheric delay a single one carries no way to remove.
A reference station degrades quietly — a cable goes intermittent, an antenna radome ices over — and the first anyone hears about it is a downstream survey that won't converge.
Point the QC Monitor at a station's feed, schedule recurring checks, and get an alert when completeness drops or multipath climbs past a threshold you set, before it reaches anyone downstream.
What won't produce a result
Worth checking before you upload, so a run doesn't cost you credits to tell you what your data was missing.
An observation file with no matching broadcast ephemeris for its date
A dated, standard observation file. Elevation-dependent checks need satellite positions for that date, fetched automatically from the file's own timestamp.
A feed with gaps larger than the scheduled check interval
A shorter check interval than the gap you're trying to catch, or a continuous feed. A monitor can only flag what it has a chance to see.
A result that can't be reproduced or checked isn't a result a reviewer — or a co-author six months later — can trust.
Every threshold, scoring formula, and precise-product source we use is published, not asserted, so a reviewer can check the method against the number instead of taking our word for it.
What won't produce a result
Worth checking before you upload, so a run doesn't cost you credits to tell you what your data was missing.
A claim with no stated method behind it
The scoring formula and thresholds, published rather than described in general terms, so the number can be checked rather than just cited.
A result with no stated uncertainty
A formal uncertainty alongside the coordinate. A position without one can't be checked against independent control or propagated into further analysis.