image sensor

An image sensor is the electronic organ of digital capture: the part of a camera that turns light into numbers. It is the modern successor to photographic emulsion, occupying the same place in the camera and the same sense of the word film — the recording medium — but storing the image as electrical charge rather than as silver.

How it captures light

A sensor is a grid of photosites, each a tiny well that accumulates electrical charge in proportion to the light falling on it — the photoelectric effect at industrial scale. At the end of an exposure the charge in every photosite is measured, amplified, and converted by an analog-to-digital converter into a number. The array of numbers is the raw image. In a CMOS sensor (now dominant) each photosite has its own amplifier and can be read out in parallel; the older CCD shifts charge across the chip to a single readout.

Color and the Bayer array

A bare photosite is color-blind: it counts photons regardless of wavelength. To record color, most sensors place a color-filter array over the grid so each photosite sees only red, green, or blue light. The standard Bayer pattern uses one red, one blue, and two green filters per 2×2 tile, because human vision is most sensitive to green. Each pixel therefore captures a single color; the two missing colors at each site are estimated from its neighbors in a step called demosaicing.

Compared with photographic film

The sensor and the emulsion solve the same problem by different physics. Emulsion stores an image chemically, as developed silver (or dye) in grains of random size and position; a sensor stores it electrically, as charge in a regular grid of photosites. Film’s grain is stochastic and organic; digital’s pixels are uniform and addressable. Both are the recording medium — sense 1 of film — realized in two different centuries.