What Is an Active Pixel Sensor (APS)?

An active pixel sensor (APS) is a solid-state image sensor in which each pixel contains both a photodetector and at least one active transistor that amplifies the photogenerated signal before it is read out, distinguishing it from a passive pixel sensor (PPS) in which the signal is transferred unamplified to a peripheral readout circuit. The inclusion of an in-pixel amplifier reduces the effective noise and capacitance at the readout node, enabling higher signal-to-noise ratios and faster readout speeds than passive alternatives. The technology was developed concurrently at Mitsubishi Electric and at NASA's Jet Propulsion Laboratory in the early 1990s, and its implementation in standard complementary metal-oxide-semiconductor (CMOS) processes gave rise to the CMOS image sensor industry that has largely displaced charge-coupled device (CCD) sensors in consumer and scientific applications.

The APS designation reflects both the active signal amplification at each pixel and the addressability that allows any pixel or region of interest to be read out independently without clocking the entire array. This random-access capability is particularly useful in high-speed tracking and adaptive imaging applications. The IEEE community has documented the sensor's development extensively, including foundational work on CMOS active pixel image sensors for highly integrated imaging systems, which established the framework for on-chip integration of image processing functions alongside the pixel array.

Pixel Architecture

The canonical 3T APS pixel consists of a photodiode, a reset transistor, a source-follower amplifier transistor, and a row-select transistor. During integration, the photodiode collects charge generated by incident photons, and the voltage on the diode node falls in proportion to the collected charge. At readout, the source-follower transistor buffers this voltage to the column line without drawing charge from the photodiode, preserving the signal. The reset transistor initializes the photodiode to a known voltage at the start of each integration period. A more advanced 4T pixel adds a transfer gate and pinned photodiode, improving dynamic range and reducing fixed-pattern noise through correlated double sampling. The 4T architecture is now standard in most commercial CMOS image sensors.

CMOS Integration and On-Chip Processing

Because APS arrays are fabricated in standard CMOS processes, analog and digital circuitry can be integrated on the same die as the pixel array without the specialized process modifications required by CCD fabrication. This integration enables column-parallel analog-to-digital converters, timing controllers, image signal processors, and output interfaces to be placed alongside the pixel array, reducing system size and power consumption. NASA's JPL pursued APS development specifically to exploit this integration benefit for space instrument miniaturization, and the high-performance active pixel sensor work in advanced CMOS color imager technology from IEEE demonstrates subsequent advances in per-pixel performance within a standard CMOS flow.

Performance Characteristics

Key performance metrics for an APS include quantum efficiency (the fraction of incident photons that generate collected electrons), read noise (the uncertainty in the pixel readout introduced by the amplifier and readout circuit), dark current (thermally generated charge that accumulates during integration), and dynamic range (the ratio of the maximum unsaturated signal to the noise floor). CMOS APS sensors historically lagged CCDs in noise performance, but ongoing process scaling and pixel architecture improvements have narrowed this gap substantially. Back-side illumination (BSI) structures, which flip the pixel so that light enters from the opposite side of the substrate, remove metal interconnect layers from the optical path and significantly improve quantum efficiency, as covered in research on CMOS active pixel image sensor design from pixels to systems.

Applications

Active pixel sensors have applications in a range of fields, including:

• Consumer digital cameras and smartphone image sensors
• Scientific and medical imaging including microscopy and X-ray detection
• Space-based instruments and planetary exploration cameras
• Machine vision systems for industrial inspection and quality control
• Distributed sensor networks requiring low-power, compact imaging nodes