TerrAlert Team · 2026-09-09
Satellite monitoring means using pictures taken from space, at regular intervals, to keep track of a place without sending anyone there. It sounds futuristic; the underlying mechanics are closer to a very patient, very high-altitude camera on a fixed schedule.
Earth-observation satellites like those in the European Space Agency's Sentinel-2 mission orbit the planet on a repeating path, capturing images in strips as they pass over. Because the orbit repeats, a given point on the ground gets photographed again every few days: for Sentinel-2, roughly every two to three days at most latitudes, using two satellites working together. That is the "revisit time," and it is a hard physical limit. No software can make a satellite pass over a point more often than its orbit allows.
Each pass captures light across several bands, not just the visible red-green-blue humans see. Near-infrared is one of them, and it is unusually good at distinguishing healthy vegetation from bare ground or water, because live plant material reflects near-infrared light very differently than soil or concrete does. This is why satellite-based vegetation and water analysis is often more reliable than it looks from an ordinary color photo.
"Resolution" describes the smallest area one pixel in the image represents. Sentinel-2's optical bands run at 10 metres per pixel, meaning each pixel covers a 10×10 metre patch of ground. That is enough to clearly see a new building, a cleared field, or a flooded area, but not enough to identify a specific vehicle or read a sign. Higher-resolution commercial satellites exist and go down to under a metre per pixel, at a correspondingly higher cost per image. The right resolution depends entirely on what you are trying to see, not on "more is always better."
The single biggest practical limitation of optical satellite monitoring is not the satellite. It is the atmosphere. A satellite pass that happens over a cloudy sky simply cannot see the ground that day; the image exists but is unusable for the area under cloud. Across a typical monitored area, a large share of passes are affected by cloud to some degree, which is why the effective observation frequency (how often a genuinely clear, usable image exists) is meaningfully lower than the satellite's raw revisit schedule.
Radar-based satellites (like the Sentinel-1 mission) work differently. They send their own signal and measure what bounces back, which means they see through cloud cover entirely. The trade-off is that radar images are harder to interpret directly and are typically better suited to detecting specific kinds of change, such as new construction, than to general-purpose visual inspection.
Satellite monitoring is well suited to large or hard-to-access areas, comparing conditions over time, and catching changes nobody was specifically looking for. It is poorly suited to anything requiring fine detail below its resolution, anything happening under persistent cloud with no radar coverage, or anything that needs to be understood the moment it happens rather than within days.
Can satellites monitor at night? Optical satellites need daylight to capture a visible-light image; radar satellites do not depend on sunlight and can operate day or night.
Why not just use higher-resolution imagery everywhere? Higher resolution generally means smaller coverage per image and a higher cost, and most monitoring use cases do not need it. 10-metre resolution is enough to see a new building or a cleared field.
TerrAlert is built on this exact Sentinel-2 imagery. See How TerrAlert Works for how it turns these passes into alerts.
All posts