Ultra-low light cameras – features and applications
Embedded vision has seen a great transformation over the past decade. There have been huge strides in terms of achieving higher resolutions while retaining high frame rates. However, there was a huge challenge, while high resolution cameras ensured quality images under normal lighting conditions, embedded cameras struggled to capture clear images under a limited light supply.
This necessitated a more robust approach towards meeting the needs of low light applications such as smart surveillance. Specialized cameras had to be developed to attain the desired levels of quality while operating under challenging lighting conditions. In this blog post, we discuss what ultra-low light cameras are, their key features, and the embedded vision applications where they are most used.
What are ultra-low light cameras?
Low light cameras refer to camera systems that can effectively capture images of scenes with restricted lighting. But a question arises as to what ‘limited lighting’, or ‘low light’ is. The fact of the matter is, there is no universally accepted definition for what a low light environment is. However, anything under 1 lux can be considered a low light setting. Similarly, an environment with lighting of less than 0.5 lux can be considered ultra-low light. In general, ultra-low light cameras are those used in vision systems that need to operate at night or have inadequate light supply.
Why are ultra-low light cameras needed in embedded vision?
Embedded vision systems operate in a wide variety of lighting conditions. While some applications have intermittent light availability, others operate without sufficient light throughout.
For example, a smart traffic system works under bright sunlight during the day and operates in low light or no light conditions at night. This requires the camera to have the ability to perform imaging in two extreme light settings.
Next, let’s look at a camera system used in an underground parking lot. The location might not have adequate light even during the day. This is an example of a vision system that needs to leverage low light capability almost 24×7.
Most ordinary camera modules are not sufficient for imaging purposes in low light applications. This is especially true in applications where automated analyses need to be carried out using computer vision algorithms, where details of the image are very critical. While some level of enhancement can be done using the ISP (Image Signal Processor), that might not do the job in most of the modern embedded vision applications. These are the reasons why low light cameras have gained popularity in recent years.
Key features of ultra-low light cameras
The most important feature of a low light camera is sensitivity. The ability of a camera module to sense light is called sensitivity. The higher the sensitivity, the higher the ability of the camera to capture high-quality images under low light conditions.
Sensitivity is not the only parameter that determines the low light performance of a camera system. There are other factors such as:
- Sensor and pixel size
- Signal to noise ratio
- EMI performance
- Lens characteristics
- NIR sensitivity
Sensor and pixel size
Machine vision systems traditionally used large cameras to capture images under low light. One of the reasons for this is that larger sensors generally used larger pixels, allowing them to gather more light over their surface area. In embedded vision, cameras tend to be smaller, they had to come with a balance of resolution and pixel size.
For example, TechNexion’s VLS3-AR0822-CB – AR0822 8MP camera – comes with a pixel size of 2.1μm x 2.1μm and a high resolution of 8MP. This makes it a perfect fit for low light applications needing ultra-high resolutions.
It’s the pixel size that matters. A camera with a large pixel size can still be mid-size if the resolution is low. A large pixel size means that the camera can collect more light per pixel, which results in a higher sensitivity.
Signal-to-noise ratio
Signal-to-noise ratio or SNR is a measure of how much light can be converted into usable digital signals for processing. In other words, it is the ratio of quality image data to the noise in the output. The higher the SNR, the better the camera’s low light performance.
FPDLINK-III IP68 Camera with onsemi AR0822 8MP 4K Color Rolling Shutter with ISP + IR-Cut Filter with C Mount Body
VLS3-AR0822-CB
- onsemi AR0822 8MP Rolling Shutter Sensor
- 4K HDR Imaging Capabilities
- Near Infra-Red Enhancement for Outdoor Applications
- Designed for Low Light Applications
- C-Mount for Interchangeable Lenses
- FAKRA Automotive Connector
- Plug & Play with Linux OS & Yocto
- Comes with VizionViewer™ Configuration Utility and VizionSDK
Sensor | onsemi AR0822 |
Shutter | Rolling |
Megapixels | 8MP |
Chromaticity | Color, Monochrome |
Interface | FPD-Link III |
EMI performance
EMI (Electromagnetic Interference) is a quality criterion not just in low light cameras but for all embedded cameras in general. It refers to any type of electromagnetic interference from external sources that deter the proper functioning of the camera system. This is why different components in a vision system like cables and connectors often come with thick coatings and protections. The quality of the camera output will decrease with an increase in EMI. Hence, for superior low light performance, the impact of EMI must be minimized.
Lens parameters
One of the factors that should be considered while choosing a lens for your embedded vision system is the amount of light it is required to capture. To capture more light, you need a lens with a large aperture. Hence in low light applications, a lens with high aperture values is chosen.
At the same time, you also need to take into consideration the size of the sensor. The lens aperture has to be such that it is compatible with the sensor size.
Learn more about the factors to consider while choosing a lens for your embedded vision system here: How to Choose Camera Lens for Embedded Vision Applications.
Lens parameters
One of the factors that should be considered while choosing a lens for your embedded vision system is the amount of light it is required to capture. To capture more light, you need a lens with a large aperture. Hence in low light applications, a lens with high aperture values is chosen.
At the same time, you also need to take into consideration the size of the sensor. The lens aperture has to be such that it is compatible with the sensor size.
Learn more about the factors to consider while choosing a lens for your embedded vision system here: How to Choose Camera Lens for Embedded Vision Applications.
NIR sensitivity
When visible light is scarce, imaging is usually carried out under NIR (Near InfraRed) lights. An NIR camera works similarly to the eyes of animals that can see in the dark such as mosquitoes, vampire bats, and some snakes. The sensor is sensitive to not just the visible spectrum but the NIR spectrum too. With this capability, these camera systems will be able to capture images even with zero ambient light.
NIR sensitivity is a critical feature in low light cameras since many of them operate during the night. TechNexion’s portfolio has many cameras that are enhanced for Near IR performance. VLS3-AR0522-CB, an AR0522 Near IR camera, is a great example.
FPDLINK-III IP68 Camera with onsemi AR0522 5MP Rolling Shutter with ISP + IR-Cut Filter with C Mount Body
VLS3-AR0522-CB
- onsemi AR0522 5MP Rolling Shutter Sensor
- Near Infra-Red Enhancement for Outdoor Applications
- Designed for Low Light Applications
- C-Mount for Interchangeable Lenses
- FAKRA Automotive Connector
- Plug & Play with Linux OS & Yocto
- Comes with VizionViewer™ Configuration Utility and VizionSDK
Sensor | onsemi AR0522 |
Shutter | Rolling |
Megapixels | 5MP |
Chromaticity | Color, Monochrome |
Interface | FPD-Link III |
Applications of ultra-low light cameras
The development of ultra-low light applications in embedded vision has been a two-way process. While the need for low light imaging led to the development of new camera technologies in the early days, today, product developers are coming up with newer use cases since high-quality low light imaging is possible.
In this section, let us look at a few next-gen embedded vision applications where ultra-low light cameras are making a huge impact.
- Smart traffic – Smart traffic systems use cameras for tasks such as ALPR (Automatic License Plate Recognition), vehicle counting, and people analytics. Doing this at night requires the camera system to have low light capabilities. In some areas, the light intensities can be lower than 1 lux, which in turn demands a high sensitivity camera module.
- Smart surveillance – Any kind of surveillance system that has to operate in low light needs an ultra-low light camera. We earlier spoke about why a parking lot management system (which is a type of smart surveillance) needs a high sensitivity camera. Similarly, surveillance systems used in streets, warehouses, mines, wildlife sanctuaries, etc., require a camera module that comes with excellent low light performance.
- Medical microscopes – Certain types of microscopic applications (e.g. dark field microscopy) use the NIR spectrum to accurately analyze the specimen. The cameras used in them should have high QE (Quantum Efficiency) values in the NIR spectrum. Most ultra-low light cameras come with this feature. Also, some microscopes have an environment with a light intensity lower than 0.1 lux.
- Patrol robots – Patrol robots automate surveillance by moving around the premises like a human. Since they have the requirement to operate 24/7, the cameras embedded in them require good low light capabilities.
A medical microscope
Final words about ultra-low light cameras
Camera technology is improving every day. Sensor manufacturers like Onsemi are coming up with innovations to consistently elevate the performance of camera modules. Recent improvements such as the BSI (Back Side Illuminated) technology are a great example of a huge leap in enhancing the sensitivity of image sensors.
We are likely to see better ultra-low light cameras in the coming years to the market, especially with the increased quality demands of AI-based embedded vision applications. Automation in image analysis is encouraging scientists and innovators to come up with newer ways to improve image quality, leading to a smarter and faster world. Camera manufacturers like TechNexion will also be a part of this revolution by introducing newer low light cameras suitable for a wide variety of industrial, medical, retail, and smart city applications.
TechNexion – pioneering embedded vision
TechNexion has been in the embedded systems industry for two decades. Our engineering expertise, product quality, and world-class manufacturing capabilities have given us an edge over our competitors for a long time. After venturing into embedded vision, our key focus has been to design, develop, manufacture, and integrate high-quality camera modules into your customers’ vision systems. From low light cameras to HDR cameras and global shutter cameras, we have every type of camera your embedded vision system needs. Learn more about our embedded vision cameras here.
Related Products
Get a Quote
Fill out the details below and one of our representatives will contact you shortly.