A low-Earth-orbit satellite moves at roughly 7.5 km/s to stay in orbit against gravity, so its ground track sweeps across any given region in only a few minutes. During that brief overpass window, each ground target is only reachable while it sits inside the sensor's field-of-regard — the shaded band around the satellite's nadir track. Once the satellite's own orbital motion carries that band past a target, the target is gone for this pass, imaged or not.
Imaging a target costs two things: the time to slew the camera (or spacecraft body) from its current aim point to the new target, and a short fixed dwell/exposure time to actually capture it. A slow, non-agile satellite spends most of the window slewing and can only reach a handful of targets before the band sweeps past them. A highly agile satellite — fast reaction wheels or a gimbaled sensor — spends only a sliver of the window on each re-aim, so it can dash between many more targets in the same window.
t_capture = t_now + |aim − target| / slewRate + dwell
target imaged if t_capture ≤ t_exit (still inside field-of-regard)
target missed if t_capture > t_exit (band already swept past)
- Slew / re-aim speed — how quickly the sensor's aim point can move between targets; this is pointing agility, not camera resolution.
- Ground targets — candidate imaging sites scattered along this overpass's ground track.
- Window remaining — simulated time left before the satellite's orbital motion carries the whole region out of range.
- Throughput — targets successfully imaged versus the total in view; this, multiplied across ~14 orbits/day, is what actually limits a constellation's daily tasking capacity — not sensor quality.
Real-world relevance: modern Earth-observation operators (Planet, Maxar, ICEYE, BlackSky) advertise agile-satellite tasking precisely because re-aim speed, not megapixels, is what decides how many customer requests one overpass can actually fulfil.