The sensor is the light-sensitive electronic component at the heart of a digital camera, replacing film as the medium that records the image. It is a flat rectangle of silicon divided into millions of tiny photosites, each one converting incoming photons into electrical charge. That charge is read out, amplified, digitized, and assembled into the final image file. The sensor’s size, photosite count, microlens design, and color filter array set the fundamental limits of resolution, dynamic range, noise, and color reproduction for the entire camera system.
Sensors are not natively able to distinguish color. Each photosite simply counts photons. Color is added by a filter array bonded to the surface, with the Bayer pattern being by far the most common. The Bayer array uses a repeating grid of two green, one red, and one blue filters, biased toward green because human vision is most sensitive to luminance in the green portion of the spectrum. Fujifilm’s X-Trans pattern uses a more randomized 6×6 grid intended to reduce moire without an optical low-pass filter. Either way, the missing color information at each site is reconstructed by demosaicing during raw conversion.
Above the color filters sit microlenses, one per photosite, that focus incoming light onto the photodiode and compensate for the gaps between sites caused by readout circuitry. Without microlenses, fill factor would drop and effective sensitivity with it. Below the photodiodes runs the readout architecture, which differs between CCD and CMOS designs. CMOS dominates today because each photosite has its own amplifier, allowing fast parallel readout, on-chip noise reduction, and the integration of phase-detection autofocus pixels directly into the imaging array.
Sensor size determines the crop factor relative to 35mm film. Full frame measures 36x24mm, APS-C runs roughly 23.5×15.6mm depending on manufacturer, Micro Four Thirds is 17.3x13mm, and medium format digital backs reach 44x33mm or larger. Larger sensors gather more total light at a given f-stop and ISO, which is why full frame typically outperforms smaller formats for low-light noise and shallow depth of field, all else equal.
Readout speed determines whether the sensor uses a rolling shutter, where rows are read sequentially, or a global shutter, where all sites freeze simultaneously. Stacked sensors with on-chip DRAM and backside illumination dramatically reduce rolling shutter artifacts and enable features like blackout-free burst shooting, pixel shift multishot, and high-speed video. Modern sensors also support dual-gain ISO, where two read paths at different gain values are combined to extend dynamic range.
Cleaning the sensor is part of routine maintenance. Dust on the cover glass shows up as soft spots in images, especially against bright, even areas like sky at small apertures. Most cameras include an ultrasonic shake mechanism, but periodic wet cleaning with proper swabs and cleaning fluid is sometimes required. Sensors are not actually being touched during cleaning; the cleaning is performed on the cover glass and any anti-aliasing filter stacked above the silicon.