The camera in your phone is part NASA
📐 Measuring the Earth
Question
How do you turn light into numbers cheaply enough to put a camera in a billion pockets?
In 1990, Eric Fossum joined NASA's Jet Propulsion Laboratory to run image-sensor research with one goal: interplanetary probes want thousands of tiny, cheap, low-power cameras, and the sensors of the day - CCDs - were big, thirsty and awkward. But shrink a pixel and the readout noise swallows the picture. The fix his group pressed: put the amplifier inside the pixel itself (an 'active' pixel), and sample twice - with light and without - so each pixel's personal offset subtracts away. In 1995 Fossum co-founded Photobit to commercialise it. The CMOS active-pixel-sensor lineage that followed is the camera in effectively every smartphone on Earth.
The math
Two samples per pixel, then subtract: signal minus the pixel's own dark offset. Noise that is fixed for a given pixel cancels out of the difference - which is what lets millions of tiny pixels on one cheap CMOS chip agree on the same picture.
Why it matters
Do the ubiquity maths: roughly 700 people have ever left Earth, while CMOS image sensors ship in the billions per year - the technology bearing NASA-born technology is, by orders of magnitude, the most widely distributed thing the agency ever touched. And the honest counterweight is the debunk card across from this one: NASA tuned this market, it did not invent the physics underneath - transistors and CCDs both predate the JPL work. What JPL actually did is narrower and better: solve a specific noise problem for spacecraft cameras, license it out, and watch it end up in a billion self-portraits. Status: September 2026.