Learn · Foundations
A satellite does not take a photograph. It measures how much light of each wavelength a piece of ground sends back. Almost everything else we do is built on that one idea.
Sunlight arriving at the ground is a mixture. Here is one place under that mixture: a lake, a stretch of forest, a ploughed field, a town at the edge. The view you would recognise from a plane window is already a measurement, because a true colour image is nothing more than the red, green and blue bands laid on top of each other.
When light hits a surface, some is absorbed and some is reflected, and the split depends on the wavelength. Measure that split right across the spectrum and you have a fingerprint for the material. A leaf answers light differently from water, and differently again from bare soil, and each of them does it consistently enough to rely on.
Above, only green light is left, and the colour has gone with it. What remains is one number for each patch of ground. Look at the forest and the field: eleven percent against fourteen. Almost the same. When two things you need to separate come back with the same value, you cannot map either of them, and you cannot tell when one has turned into the other. That is the gap between having an image and having a measurement you can act on.
Chlorophyll absorbs red light to run photosynthesis, so a healthy canopy sends almost none of it back. The forest drops to about four percent and becomes one of the darkest things in the picture, darker even than the bare field beside it.
The lake is darker still. The town barely moves, because concrete has no pigment doing any work at all.
Move less than two hundred nanometres and the same forest jumps to about fifty percent. Nothing about the trees changed. Chlorophyll has no use for infrared, and the internal structure of a healthy leaf scatters it hard instead of absorbing it.
That jump happens between about 700 and 750 nanometres, and it has a name: the red edge. A stressed, drying or dying plant loses it first, before any colour change is visible to somebody standing in the field. This is the single most useful fact in optical remote sensing.
Out here water itself is what registers. These wavelengths are absorbed strongly by water, so the wetter a surface is, the darker it comes back. Pigment no longer has much to do with it.
That reorders the scene one last time. The forest is full of water, so it dims. The ploughed field is dry, so it overtakes the forest and becomes the brightest surface here. The lake, black since we left the visible, stays black. This is the band you reach for when the question is drought, irrigation or soil moisture.
Everything you have just scrolled past is the same ground, on the same day, untouched. Only the wavelength changed.
In green light the forest and the field were hard to separate. In the near infrared they are nothing alike. Satellites carry infrared bands for that reason: the surfaces we care about are far easier to tell apart there than anywhere in the range your eye covers.
Infrared is not a special effect and it is not false data. It is ordinary light with a longer wavelength, arriving in the same sunbeam as the rest, carrying information the eye was never built to collect. Choosing a band is really choosing which differences you want to be able to see.
Every deliverable we produce traces back to this. Deforestation screening works because a forest and a cleared plot have different signatures. Water mapping works because water is the darkest thing in the infrared and nothing else behaves like it. Built up area detection works because concrete is stubbornly flat where everything natural is not.
None of it is image interpretation in the everyday sense. It is measurement, made against physics that holds whether or not anyone is looking, and then checked on the ground. The rest of this series follows that chain forward, one step at a time.
Reflectance values here are representative figures for each surface type, not measurements of a specific place.