What is a CMOS Image Sensor?

1 Answer
Can you answer this question?

- GoPhotonics

Apr 11, 2024

An image sensor is a device that converts optical images into electronic signals. The two main types of image sensors are the charge-coupled device (CCD) and the complementary metal oxide semiconductor (CMOS sensor). These sensors operate based on metal-oxide-semiconductor (MOS) technology, with CCDs using MOS capacitors and CMOS sensors employing MOSFET (MOS field-effect transistor) amplifiers. 

A CMOS image sensor is an electronic chip that converts photons to electrons for digital processing. This is a semiconductor device that captures visual information in digital form. It was invented in 1963 by Frank Wanlass. This sensor is used to create images in digital cameras, digital video cameras, digital CCTV cameras, smartphones, and various imaging systems.

Unlike CCD (charge-coupled device) sensors, CMOS sensors integrate amplifiers and A/D converters at each pixel, thereby enabling faster readout speeds and lower power consumption. The CMOS sensor operates on the principle of the photoelectric effect to change photons into electrical energy. In contrast to CCD sensors, CMOS sensors directly convert electric charge into voltage within the pixels. Currently, CMOS sensors are available with outstanding image quality and high frame rates, making them suitable for use in high-performance industrial cameras.

fig-2-ucsmva-lg

Within a CMOS sensor, the charge originating from the photosensitive pixel undergoes conversion into a voltage directly at the pixel site. Subsequently, the signal is multiplexed by row and column, directing it towards multiple on-chip digital-to-analog converters (DACs).

Components of CMOS Sensor

Fuji1

A typical CMOS image sensor is an electronic device that converts light intensity into a digital signal. It is an integrated circuit featuring an array of pixel sensors and supporting circuitry. Each pixel has three photodiodes and an operational amplifier. The supporting circuitry consists of an reset/sample pin, matrix selection switches, an output gain amplifier, and an analog to digital converter (ADC).

All of this is packed into a standard integrated circuit package and placed on printed circuit boards (PCBs). The components are all inside this package, and external pins help control the sensor. Usually, CMOS image sensors work with a shutter lens and a computer to display and process the images they capture.

The CMOS sensor contains four main parts: the color filters, the pixel array, the digital controller, and the analog to digital convertor.

Color Filter is a mosaic of small filters placed on the pixel sensors designed to capture color information. These color filters facilitate the individual measurement of red (R), green (G), and blue (B) photons. By filtering out unwanted wavelengths, the color filters allow only specific colors of light to reach a pixel sensor. To achieve this, each pixel is equipped with a red, green, and blue filter arranged in a specific pattern, such as the Bayer CFA pattern. The Bayer filter employs sub-mosaic 2x2 patterns containing one red, one blue, and two green filters. Given the human eye's heightened sensitivity to green light, two green filters are utilized in this arrangement.

fig-4-ucs

Pixels are the smallest unit in a digital display. It operates based on the photoelectric effect, with pixel sensors tasked with capturing the intensity of the light that passes through them. This intensity data is combined and then converted into an analog voltage signal, which is subsequently sent to an external circuit board for additional processing. A pixel is composed of three photodiodes accompanied by an operational amplifier. Each photodiode is equipped with a specific filter, permitting only one color of light to pass through. Within a pixel, these three photodiodes are selectively covered with red, green, or blue filters.

Parts of Pixel

  • Photodiode: A photodiode is a light-sensitive device that generates an electrical charge when exposed to light energy. This charge results in a small current, which corresponds to the intensity of light detected by the photodiode. However, this current needs initial signal conditioning from a transimpedance amplifier to be usable.
  • Transimpedance Amplifier: A transimpedance amplifier is a standard operational amplifier configured as a transimpedance circuit. It can receive a small current and convert it into an analog voltage. The analog voltage represents the input current according to a specific equation. While the information remains the same, the analog voltage is more compatible with other circuits.

The support circuitry accompanying a CMOS image sensor includes a reset/sample pin, matrix selection switches, an output gain amplifier, and an analog-to-digital converter (ADC). These elements facilitate the conversion of the analog voltage output from the pixels into a format usable by computing systems for image data processing.

Support Circuitry Parts of CMOS Sensor

  • Reset/Sample Pin: The reset/sample pin serves to initiate the capture of a new image by the camera. When a 3.3V signal is sent from the external computing system to this pin, the previous electric charge accumulated in the photodiodes dissipates, resetting the image sensor. Subsequently, the photodiodes capture a new sample (image). Additionally, this signal typically governs the operation of the external system's shutter lens, controlling the duration of light exposure on the photodiodes.
  • Matrix Selection Switches: Within the supporting circuitry lies selection circuitry, utilized for isolating specific pixel analog voltage values for further processing. Configured in a row-column layout, these selection switches allow individual addressing and access to the analog voltage values of each pixel. External selection of row and column numbers prompts the internal switches to open and close accordingly, pinpointing the desired location within the pixel matrix. These switches, constructed from transistors, operate under biased switch configurations.
  • Output Gain Amplifier: The output gain amplifier adopts a non-inverting operational amplifier configuration. It amplifies a small analog voltage input, yielding a larger analog voltage output based on the following equation.
  • Analog-to-Digital Converter (ADC) is a device primarily employed to convert signals. It begins by transforming an analog signal into a digital format. The ADC takes the analog voltage signals from the pixel sensor array and converts them into a digital signal. This conversion is achieved using a sample and hold algorithm.

The configuration of a CMOS sensor often integrates a rolling shutter, although the inclusion of additional transistors at the pixel site enables the attainment of a global shutter. Rolling shutter is a technique for capturing images or video frames in which the sensor scans across the scene either vertically, horizontally, or rotationally, instead of capturing the entire scene at once. This can lead to distortion effects, such as skew or wobble, especially when there is movement in the scene during capture.

On the other hand, global shutter is a technique for capturing images or video frames in which the entire frame is exposed to light simultaneously. This ensures that all pixels are captured at the same time, eliminating the distortion effects associated with rolling shutter, and is particularly useful for capturing fast-moving objects or scenes with rapid motion.

The operation of global shutter operates by exposing all pixels simultaneously and then reading them out sequentially. A notable advantage of CMOS sensors lies in their lower power consumption and dissipation compared to equivalent CCD sensors, owing to reduced charge flow or current. Moreover, CMOS sensors exhibit resistance to blooming, allowing them to handle high light levels adeptly, making them suitable for specialized high dynamic range cameras capable of capturing intricate details such as welding seams or light filaments. Additionally, CMOS cameras typically feature smaller form factors compared to their digital CCD counterparts, as digital CCD cameras necessitate extra off-chip ADC circuitry.

However, the multilayer MOS fabrication process used in CMOS sensor production precludes the incorporation of microlenses on the chip. As a result, the effective collection efficiency or fill factor of CMOS sensors is lower in comparison to CCD equivalents. This reduced efficiency, coupled with pixel-to-pixel inconsistencies, contributes to a diminished signal-to-noise ratio and overall image quality when compared to CCD sensors.

Comparison of CCD and CMOS Sensor

Sensor

CCD

CMOS

Pixel Signal (Chip Output)

Electron

Voltage

Chip Signal

Analog

Digital

Fill Factor

High

Moderate

Responsivity

Moderate

Moderate – High

Noise Level

Low

Moderate – High

Dynamic Range

High

Moderate

Uniformity

High

Low

Resolution

Low – High

Low – High

Speed

Moderate - High

High

Power Consumption

Moderate – High

Low

Sensitivity

High

Low

System Complexity

High

Low

Sensor Complexity

Low

High

Working of CMOS Sensor

The working of CMOS image sensor begins with an external signal activating the shutter lens and triggering the reset/sample pin on the image sensor. This action exposes the photodiodes within each pixel to incoming light. Subsequently, the photodiodes generate currents proportional to the intensity of light they receive from the environment. 

Each pixel's current corresponds to red, green, or blue light, facilitated by the inclusion of filters within each pixel. The minute currents are then directed to an operational amplifier (op-amp) configured in a trans-impedance setup, transforming the current into a low-level analog voltage signal. Consequently, each pixel now possesses three distinct low-level analog voltage values, representing the intensity of red, green, and blue light.

Utilizing a matrix of switches, each pixel is individually selected by specifying a row and column. The analog voltage values of the selected pixel are then routed to an output gain amplifier, which elevates the voltage levels to a usable range for the analog-to-digital converter (ADC). The ADC subsequently converts these analog voltage values into digital equivalents.

This conversion process recurs for each pixel's analog voltage values until all pixels' analog signals are digitized. Ultimately, the CMOS image sensor outputs digital voltage values, reflecting the red, green, and blue light intensities captured by each pixel within the sensor's matrix. These digital voltage values are transmitted to an external computing system, commonly a processor, where they can be further manipulated or displayed as a digital image on a screen.

Types of CMOS Sensor

The distinction between types of CMOS sensors typically arises from the quantity of transistors per pixel, which affects the fill factor. The fill factor refers to the portion of the pixel sensor that is sensitive to light.

RollingShutter_GlobalShutter_Animation

  1. Rolling Shutter Type: This type features a limited number of transistors, resulting in a high fill factor. However, because lines of pixels are exposed at different times, movement in the target can cause image distortion.
  2. Global Shutter Type: In contrast, this type has a high number of transistors, leading to a low fill factor. All pixels are exposed simultaneously, eliminating the movement artifacts associated with rolling shutter type sensors.
  3. Mechanical Shutter: A mechanical shutter is a physical mechanism in a camera that controls the amount of light reaching the image sensor and the duration of exposure. It typically consists of one or more blades or curtains that open and close to allow light to pass through the camera's lens and reach the sensor. When the shutter is closed, it blocks all light from reaching the sensor, and when it opens, it allows light to expose the sensor for a specific period of time, determining the exposure duration. Mechanical shutters are commonly found in DSLR (Digital Single-Lens Reflex) cameras and traditional film cameras.

Applications

CMOS can also be found in astronomical telescopes, scanners and barcode readers. The optical technology is used in machine vision for robots, in optical character recognition (OCR), in the processing of satellite photographs and in the enhancement of RADAR images, especially for meteorology. The low-cost manufacturing of CMOS makes it achievable to generate low-cost consumer devices. 

  • Astronomical Telescopes: Utilized for capturing high-resolution images of celestial objects, CMOS sensors offer sensitivity to low light levels, aiding astronomers in studying faint stars and distant galaxies.
  • Scanners and Barcode Readers: In retail, logistics, and office environments, CMOS sensors power fast and accurate scanning, essential for document scanning, image capture, and barcode reading.
  • Machine Vision for Robotics: Integral to machine vision systems in robotics, CMOS sensors enable robots to perceive and interact with their environment, vital for tasks like industrial automation, quality control, and autonomous navigation.
  • Optical Character Recognition (OCR): Integrated into OCR systems, CMOS sensors convert printed or handwritten text into digital data, used in document digitization, text extraction, and document management.
  • Satellite Photography Processing: In environmental monitoring, land surveying, urban planning, and disaster management, CMOS sensors process satellite photographs, providing high-resolution imagery for detailed analysis of Earth's surface.
  • Enhancement of RADAR Images: In meteorology, CMOS technology enhances RADAR images, improving weather forecasting and atmospheric phenomenon tracking by combining RADAR data with optical imagery.
  • Low-Cost Consumer Devices: With its cost-effective manufacturing, CMOS enables affordable consumer devices like digital cameras, camcorders, webcams, and smartphones, democratizing access to advanced imaging technology.