What is CMOS Sensor? How does it work?

Introduction

An image sensor is a device that converts light into electrical charges, which can then be processed to create an image. In digital cameras, the image sensor is one of the most important components, as it plays a crucial role in determining the quality of the final image. There are several types of image sensors, each with their own strengths and weaknesses. This article will provide an overview of the CMOS image sensor, its working principle, and a comparison between CMOS and CCD image sensors.

What is a CMOS sensor?

A CMOS (Complementary Metal-Oxide-Semiconductor) image sensor is a type of image sensor that converts light into electrical charges. The electrical output signal can then be processed to create an image. It is commonly used in digital cameras, smart-phones, and other imaging devices.

Working principle of CMOS image sensor

The CMOS image sensor is made up of an array of tiny light-sensitive cells also known as pixels, each of which is connected to a transistor that acts as a switch. When light strikes the pixels, it is converted into an electrical charge which is then amplified and read out by the sensor’s on-chip readout electronics. The output is then converted into a digital signal, ready for further image processing or storage.

Capture and readout process

CMOS image sensors capture image data by reading out each pixel individually. This process is known as “parallel readout” and results in a faster readout time. Additionally, the readout circuit for each pixel is integrated on the same chip as the pixel which makes the manufacturing of CMOS image sensors cost effective.

Conversion of light into electrical charges

The conversion of light into electrical charges is a key process within a CMOS image sensor. The image sensor is made up of an array of tiny light-sensitive cells or pixels. These pixels are typically made up of photodiodes which convert light into electricity. Each photodiode is connected to a transistor that acts as a switch, allowing the charge to be read out from the diode. The charge output from each photodiode is then collected by the on-chip readout electronics and is used for constructing the image.

What does camera sensor size mean?

CMOS sensors are often characterized by their physical dimensions. The amount of light that a CMOS image sensor can capture is a function of its physical size. Hence, the resolution and pixel size of the CMOS sensor are also a function of its size which is typically specified in inches. The sensor size or optical format of the CMOS sensor is calculated through the multiplication of sensor’s diagonal size by 3/2.

Another important factor related to the size of CMOS image sensor is Crop factor. Crop factor also known as focal length multiplier is a measurement that is used to compare the field-of-view of different image sensors. The crop factor is determined by dividing the diagonal size of a full-frame image sensor (35mm) by the diagonal size of the sensor in question. A crop factor of 1 indicates that the sensor has the same field-of -view as a full-frame sensor while a crop factor greater than 1 indicates a smaller field-of-view.

CMOS sensors versus CCD sensors

CMOS (Complementary Metal-Oxide-Semiconductor) and CCD (Charge-Coupled Device) image sensors are both used to convert light into electrical signals that can be processed to create an image. However, there are some major differences between these two image sensor types.

The first difference between CMOS and CCD image sensors lies in the way the image data is captured and read out. A CCD image sensor captures image data by moving charge packets in a linear sequence from one pixel to the next, which is referred to as “charge transfer.” This method results in improved image quality, but it also has drawbacks such as a slower readout speed. On the other hand, CMOS image sensors capture image data by reading out each pixel individually, a process known as “parallel readout.” This results in a faster readout time but generally lower image quality compared to CCD sensors.

Another major difference is in the production cost. CMOS image sensors are generally less expensive to produce than CCD image sensors due to their more complex electronic circuitry and manufacturing process.
In terms of power consumption, CCD image sensors are known to be more power-hungry as compared to CMOS image sensors. This is because CCD image sensors need active consumption of power to gather light as opposed to the CMOS image sensors which don’t have such a requirement.

When it comes to flexibility, CMOS image sensors have a distinct advantage as the readout circuit for each pixel is integrated on the same chip as the pixel. This feature allows for a wide range of enhanced functions such as image processing and signal amplification to be added to the sensor.

Lastly, CCD image sensors are known for their high image quality, lower noise, and greater light sensitivity while CMOS image sensors have the advantage of being faster, less power-hungry and more cost-effective.

Why CMOS sensors are leading the world of embedded vision?

CMOS (Complementary Metal-Oxide-Semiconductor) image sensors are dominating the world of embedded vision for several reasons. The first major reason is the manufacturing cost as CMOS sensors are less expensive to produce than the CCD image sensors.

Another key factor is the power consumption, as CMOS sensors consume less power as compared to CCD sensors. This factor is particularly important for embedded vision systems that have strict power consumption requirements.

Furthermore, advancements in CMOS sensor technology have significantly improved their imaging capabilities, closing the gap with CCD sensors in terms of sensitivity, noise and image quality. For example, Sony’s STARVIS series includes a wide variety of sensors with superior low light performance and NIR (near infrared) sensitivity.

In conclusion, CMOS sensors are winning the race in embedded vision due to their cost-effectiveness, power efficiency, and constantly improving imaging capabilities. All these factors make CMOS sensors an ideal choice for embedded vision applications resulting in their growing popularity and increasing adoption rate.
How to choose the right CMOS sensor size for an embedded vision application?

Following factors are taken into consideration while selecting the CMOS image sensors for an embedded vision application:

• Resolution
• High frame rate and global shutter
• Sensitivity
• Lens mount selection
• Image circle diameter
• Low light performance

Resolution is a critical factor, particularly for applications that require precise 3D depth measurement. Large pixel sizes tend to have higher resolution, and it’s important to ensure that the resolution of the lens matches the pixel size of the sensor to achieve high-quality images.

High frame rate and global shutter features are important for applications such as automated license plate recognition, gesture recognition, robotic vision, drones, and autonomous mobile robots (AMR).
Sensitivity is also an important factor, as large sensors tend to have large pixel sizes which indicate higher sensitivity. This is crucial for applications that require high image recognition and detection performance, such as smart city, surveillance, and traffic monitoring systems.

The lens mount selection depends on the sensor size. For example, a C-mount is a suitable option for a 1/1.5 inch sensor, while an S-mount lens, commonly used in industrial applications, is appropriate for sensors that are 1/2 inch, 1/3 inch or smaller in size.

Industrial cameras can experience problems such as lens vignetting or lens shading, which causes a decrease in brightness or color intensity from the center of an image to its edges. This is often caused by a lens with an image format (or circle) that is too small for the sensor being used. To avoid this problem, it is important to ensure that the diameter of the image circle is either the same size or is larger than the sensor.

A larger sensor is better equipped to capture images in low-light conditions because it contains bigger photosites that are more sensitive to light. This is better than a smaller sensor in low light situation. Two popular sensor sizes that are specifically designed for low light performance are 1/1.2 inches (e.g. the Sony IMX485 4K-resolution CMOS image sensor) and 35mm full-frame.

Conclusion

CMOS image sensors are one of the most widely used types of image sensors. Due to their on-chip readout circuitry and parallel readout process, CMOS sensors offer faster readout times as compared to CCD sensors. Moreover, CMOS sensors are more cost-effective and power-efficient as compared to the CCD image sensors. In this article we have discussed the operating principle of CMOS sensors, CMOS sensor sizing, CMOS sensor selection for embedded vision applications, and comparison between CMOS and CCD sensors.