The CMOS active-pixel sensor
1993 (NASA Jet Propulsion Laboratory)
Eric Fossum's team at NASA's JPL put an amplifier in every pixel of a chip made like ordinary computer chips. Cheap, fast and frugal, CMOS sensors now sit in nearly every camera and phone.
How it began
NASA wanted smaller, lighter, lower-power cameras for spacecraft. At the Jet Propulsion Laboratory in Pasadena, Eric Fossum and his team showed in 1993 that an image sensor could be made with the same CMOS processes used for computer chips, if each pixel had its own amplifier and a way to transfer its charge.
How it works
As in a CCD, light frees electrons in each pixel. But instead of shifting charge across the chip, each pixel converts its own charge to a voltage and is read out row by row, with timing, digitising and processing circuits on the same chip. That uses far less power and allows very fast readout. See how light becomes an image.
Why it mattered
CMOS made image sensors cheap enough to put in phones, laptops and cars, and fast enough for high-speed bursts and video. Canon’s use of its own CMOS sensors helped launch affordable full-frame and APS-C digital SLRs. Back-illuminated (2009) and stacked designs later improved low-light quality and speed.
Limits and what replaced it
Early CMOS sensors were noisier than CCDs, and reading rows one after another causes “rolling shutter” distortion with fast motion. Stacked sensors and, from 2023, global-shutter CMOS sensors in some cameras reduce this.
Today
Nearly all cameras and phones use CMOS sensors, many of them back-illuminated or stacked.