Misconceptions about image sensor pixels

Image sensors are becoming more and more popular, especially in security, industrial and automotive applications. Many cars are now equipped with at least five or more image sensor-based cameras. However, there are some misconceptions about image sensor technology being different from standard semiconductor technology.

Moore's Law and Image Sensors

Some have assumed that the famous "Moore's Law" also applies to image sensors. Gordon Moore (founder of Fairchild Semiconductor, now part of ON Semiconductor) noted that the number of transistors on an integrated circuit (IC) doubles every two years. In order to fit twice as many transistors on a single device, the main way to achieve this is by shrinking the transistors. This trend has been going on for decades, but the rate of increase in transistor count has slowed in recent years. The increase in transistor density has also led to a decrease in the cost per transistor, so many electronic systems have become more and more versatile without increasing their prices.

But image sensors are different because some of their important components do not scale up as transistors get smaller. Specifically, image sensor optical components such as photodiodes (which convert incoming light into electrical signals) and some analog components (which convert electrical signals into digital images) cannot be easily scaled up like digital logic components. In sensors, image capture is mainly done using analog technology, and digital circuits convert the digital data from each pixel into an image that can be stored, displayed, or used for artificial intelligence machine vision.

If the number of pixels doubles every two years and the lens size remains the same, the pixels will become smaller, resulting in fewer received photons. (Think of a bucket on a rainy day; a smaller bucket collects fewer raindrops.) Therefore, the sensor must perform better in terms of sensitivity per unit area and noise reduction to produce the same high-quality images in low-light conditions. If it is not required by the product's application scenario, increasing the number of pixels is not only meaningless, but also forces an increase in bandwidth and storage space, increasing the cost of other components of the system.

Pixel size

Pixel size alone is not enough to determine pixel performance. We cannot assume that larger pixels will necessarily result in better image quality. While pixel performance under different lighting conditions is important, and larger pixels have more area to collect light, this does not necessarily improve image quality. Several other factors are also important, including resolution and pixel noise specifications.

If the application is impacted more by increasing resolution than by reducing the number of individual pixels, then a sensor with smaller pixels may outperform a sensor with larger pixels for the same optical area. It is important to ensure that the number of received photons is sufficient to form a high-quality image, so pixel sensitivity (photoelectric conversion efficiency) and application environment are very important.

Pixel size is a consideration when choosing a sensor for an application. However, its importance may be overstated; it is just one parameter among several that should be given the same careful consideration. When selecting a sensor, designers must consider all the requirements of the target application and then find the ideal balance of speed, sensitivity, and image quality to achieve a suitable design solution.

Large and small pixel design

In many applications, it is valuable to extend the dynamic range as much as possible to help correctly render shadows and highlights in the final image, but this can be very challenging for image sensors. Some companies have adopted a technique called "big and small pixels" to solve the problem of creating more capacity for the photodiode to collect electrons before the diode "saturates".

In the big and small pixel approach, the sensor area dedicated to a single pixel is divided into two parts: a larger photodiode covers most of the area, and a smaller photodiode uses the remaining area. Larger photodiodes collect more photons and can easily saturate in bright light conditions. Smaller photodiodes can be exposed longer without saturation because there is less area available to collect photons. A good analogy is to using a bucket and a water bottle to collect raindrops. A bucket is usually wider at the top than at the bottom, so it can collect raindrops very efficiently and fill up faster than a water bottle, which has a small opening and a wider body, so it collects raindrops more slowly. Using larger pixels in low light conditions and smaller pixels in bright light conditions can create an extended dynamic range.

Figure 1: Complex semiconductors, from photons to image output

ON Semiconductor solves this problem by adding an area to a single pixel where the excess signal or charge can overflow. Imagine using a bucket to catch raindrops, and then having a larger basin to catch the water that overflows from the bucket. The "bucket" signal is easy to read with high accuracy, allowing us to achieve good low-light performance, and the larger basin contains all the overflowing signal, extending the dynamic range. This way, the entire pixel area is used in low-light conditions, rather than saturating in bright light conditions. Saturation degrades image quality, such as distorting colors and reducing clarity. ON Semiconductor's super exposure technology provides better image quality in high dynamic range scenes, suitable for human vision and machine vision applications.

Figure 2: XGS 16000 is a 16-megapixel CMOS image sensor

The next time you choose an image sensor for your design, remember that “more and bigger is not necessarily better,” at least when it comes to pixels.