A camera sensor is an electronic imaging component that detects light and converts the detected optical energy into electrical signals that can be processed into a digital image. It is the digital camera's primary light-sensitive surface and performs a role broadly analogous to the light-sensitive film in a traditional camera, although the physical mechanisms differ. The sensor contains an array of photosensitive elements, often called photosites, which measure light arriving from different areas of the optical image. The resulting measurements are read electronically, converted into digital values and processed into pixels in the final image. Common digital camera sensor technologies include CMOS and CCD, although CMOS dominates most modern cameras because of manufacturing, integration and power advantages.
A sensor's photosites detect photons using semiconductor materials. When light enters a photosite, photons can generate charge carriers through semiconductor photoelectric processes. The accumulated charge represents information about the amount of incident light, subject to the sensor's response characteristics and operating limits. Readout circuitry measures this charge and converts it into a signal that can subsequently be digitised. The maximum charge a photosite can hold before saturation is related to full-well capacity, while the smallest useful signal is constrained by sources of noise. These characteristics contribute to dynamic range, which describes how effectively the sensor can distinguish relatively dark and bright information within a single exposure.
Camera sensors differ in physical size. Common categories include small sensors used in compact devices and smartphones, Micro Four Thirds sensors, APS-C sensors, full-frame sensors and larger medium-format sensors. Sensor size affects several aspects of photographic behaviour, including the amount of light that can be captured for a given exposure and the field of view produced by a lens of a particular focal length.
Larger sensors can make it easier to achieve shallow depth of field at equivalent framing and can provide advantages in noise performance when other factors are comparable. However, larger sensors require larger lenses and camera bodies for equivalent optical systems, and sensor size alone does not determine image quality.
Sensor resolution is determined by the number and arrangement of photosites, but photosites are not necessarily identical to final image pixels in every camera. A conventional colour sensor with a Bayer filter array measures only one principal colour component at each photosite, and the image processor estimates the remaining components through demosaicing. Sensor design can also include different colour-filter patterns, stacked architectures or alternative colour-detection methods. Pixel size is determined by the physical sensor area divided by the number of photosites. Smaller photosites can permit high pixel counts in compact physical areas but may have different noise, diffraction and full-well characteristics from larger photosites.
Noise is an unavoidable consideration in electronic imaging. Read noise arises from electronic measurement and amplification processes, while shot noise results from the statistical nature of photon arrival and is particularly relevant at low light levels. Dark current and thermal effects can also contribute to noise, especially during long exposures or at elevated temperatures. Signal processing can reduce visible noise, but aggressive noise reduction can remove fine detail or texture. Sensor performance therefore depends on the balance between light collection, conversion efficiency, electronic noise and image-processing algorithms. This is one reason why a larger sensor with fewer but larger photosites can sometimes perform differently from a smaller high-resolution sensor despite similar total pixel counts.
Sensors also differ in how they capture motion. Many CMOS sensors use rolling shutters in which image rows are read sequentially. Fast motion can therefore produce geometric distortion because different parts of the scene correspond to slightly different capture times. Global-shutter sensors can capture a frame with substantially more simultaneous exposure timing, reducing certain motion distortions. Sensors can also support different frame rates, readout speeds, bit depths and autofocus technologies. Faster readout is valuable for video, burst photography and reducing rolling-shutter artefacts, but achieving high-speed readout without excessive noise or power consumption requires sophisticated sensor and processing design.
Sensor characteristics are closely linked to lens and exposure characteristics. A sensor cannot recover detail that the lens fails to resolve, and a high-resolution sensor does not automatically create a high-detail photograph if focus is inaccurate or motion blur occurs. The sensor's dynamic range can also be limited by the amount of light captured and its saturation capacity. Exposure settings determine how the sensor is used, while image processors determine how its raw measurements are transformed into visible images. Consequently, sensor specifications such as megapixel count, pixel pitch and sensor format should be interpreted together rather than treated as independent measures of photographic quality.
A camera sensor is therefore an electronic light-detection array that forms the central measurement stage of digital imaging. Its physical size, photosite count, photosite size, quantum efficiency, read noise, full-well capacity, dynamic range, readout architecture and colour-filter arrangement all affect the information available to the image processor. CMOS and CCD describe sensor technologies rather than image-quality grades, and megapixels describe spatial sampling rather than every aspect of sensor performance. The sensor captures measurements of light; the lens determines how that light is projected, and subsequent electronics and software convert those measurements into the final photograph.
