NOISE REDUCTION TECHNOLOGY IN SECURITY CAMERAS

Vũ Khôi

Author at Wotech Electronics Blog

()

In a surveillance camera, when the light is dim, the camera's amplifier boosts the light to brighten the image. That amplification, however, introduces grainy noise into the image, which then calls for a filter to remove it.

This is exactly why digital noise reduction (DNR) technology was born as the answer to the problem. DNR (Digital Noise Reduction) technology processes images that turn blurry or grainy (collectively, "noise") when the camera has to observe in low-light conditions.

Today, noise reduction technology comes in two types — 2D DNR and 3D DNR — applied in surveillance camera manufacturing to:

• Increase brightness so images appear brighter and clearer.

• Reduce fine grain (known as Noise Reduction) for a smoother, sharper picture (details become easier to see).

2D-DNR is temporal noise reduction technology for low-light conditions. However, 2D-DNR is not very sharp and therefore cannot be applied to high-resolution images and video of 2K and above.

3D-DNR is spatial (3D) noise reduction technology for low-light conditions. 3D-DNR can remove noise from high-resolution video of 2K, 4K, and beyond, for smooth and sharp images.

2D-DNR Noise Reduction Technology

2D-DNR (2 Dimension – Digital Noise Reduction) is temporal noise reduction technology for cameras in dim light: the surveillance camera cannot see clearly, so grainy noise appears during video signal amplification.

2D-DNR technology uses an algorithm that analyzes and compares one frame against the next. The algorithm then finds and removes any anomalies between those frames. As a result, the image and video signal has its grain reduced (i.e., noise reduction).

The advantage of 2D-DNR is its fast processing time and the fact that the camera doesn't need a powerful graphics processor. This lowers the cost of the surveillance camera.

However, 2D-DNR is a temporal noise reduction method that cannot identify all noise points, so the image still retains a good deal of fine grain in dim conditions.

The 2D-DNR algorithm is practically "helpless" when processing high-resolution images; it is not suitable for cameras with higher resolutions such as Full HD (2.0MP), 4MP, and 8MP.

And it is precisely for this reason that 3D-DNR noise reduction technology was researched and developed.

3D-DNR Noise Reduction Technology

• What is 3D-DNR noise reduction?

3D-DNR is spatial (3D) noise reduction technology based on analyzing multiple frames simultaneously. The new algorithm in 3D-DNR compares different pixels and detects the pixels likely to be noisy before removing them.

• How 3D-DNR noise removal technology works:

3D-DNR technology doesn't just compare multiple adjacent frames — its algorithm can also infer which pixels are noisy based on the pixel matrix.

As a result, 3D-DNR noise reduction technology markedly improves image quality, proving especially useful for high-resolution camera lines: Full HD (2.0MP), 2K (4.0MP), and even 4K (8.0MP).

• Advantages of 3D-DNR technology:

3D-DNR digital noise reduction technology eliminates the blurry appearance of low-light images and handles moving objects without leaving motion blur trails, even in dim light.

The 3D-DNR noise reduction algorithm produces clearer, sharper images than unprocessed images (or those processed with the older 2D-DNR technology).

3D-DNR solves the noise reduction problem for ultra-high-resolution images, making it a fit for premium camera lines with resolutions of Full HD and above.

That said, the 3D-DNR algorithm demands more work from the camera's processor — deeper, more detailed analysis — so the camera needs a more powerful graphics processor to run the 3D Digital Noise Reduction algorithms well. This can indirectly increase the camera's price.

At KBVISION, the CVI 4K line, the PIR line, and all IP camera lines come with 3D-DNR noise reduction technology built in, so you can confidently choose a camera with 3D-DNR that fits your needs.

How useful was this article?

Click on a star to rate it!

Average rating / 5. Vote count:

No votes so far! Be the first to rate this article!

Related Articles