The principle of CMOS is relatively straightforward. Its working process can generally be divided into four stages: reset, photoelectric conversion, integration, and readout.
When incident light passes through RGB color filters and reaches the CMOS sensor, the photoelectric effect occurs. Electrons are generated and read out, then converted into analog signals. These signals are amplified (ISO adjustment), and finally converted into voltage values. When these values are aggregated, they form the image we see.
A raw CMOS sensor cannot capture color images. After the shutter is pressed, the photodiode in each pixel converts the light intensity into a digital value (0–255).
This range represents brightness levels, from black (0) to white (255).
Color Filter Array (CFA)
To capture color images, a structure called a Color Filter Array (CFA) is required. This array functions similarly to the cone cells in the human eye. Specific wavelengths of light pass through corresponding color filters, allowing each pixel to capture color information.
Bayer Pattern
The most common CFA is the Bayer filter array. It consists of alternating rows of red, green, and blue filters. In this arrangement, 50% of the filters are green, while red and blue each account for 25%.
The Bayer pattern is also referred to as the RGGB pattern, representing the smallest repeating unit of the filter arrangement.
Over the past two decades, many optical manufacturers have explored alternative CFA designs. However, despite various innovations, the Bayer array remains the dominant standard today.
Single Pixel Structure
Each pixel in a CMOS sensor is measured in micrometers (µm) and includes both the photodiode and surrounding electronic components.
A CMOS pixel typically consists of:
- Photodiode
- Amplifier
- Reset Gate
- Transfer Gate
- Floating Diffusion
On top of these components is a microlens, which helps concentrate more light onto the photodiode, improving sensitivity.
Sensor Size, Resolution, and Pixel Size
The overall sensor size is determined by both resolution and pixel size. Generally, larger sensors either provide higher resolution or larger pixel sizes compared to smaller sensors.
However, due to space constraints in devices like smartphones, sensor development has taken alternative approaches within fixed sizes. Pixel-related optimizations will be discussed in more detail in later sections.
Summary
In essence, CMOS sensors convert light signals into electrical signals through a series of processes, and then use algorithms to output a readable image.
