Raghad Rabah – 2M CCTV
Generally, when taking pictures or videos in the dark, you will notice dark and light spots on the image. This is prevalent in closed circuit television (CCTV) cameras, digital cameras, camcorders, basically any type of camera in low light settings. Fortunately for surveillance cameras, most of them feature digital noise reduction. It removes the specks on the photo (noise) and saves you space on your hard drive and makes the image clearer and easy on the eye.
Here is an example of Digital Noise Reduction in action using the Veilux VS-70CDNRD DNR Color Box Camera
What is Noise?
Have you ever taken a picture or video outside in the dark (without flash)? You’ve probably noticed that it can sometimes turn out to be very grainy. In technical terms, this is called noise. Noise is any unwanted interference in the signal. It can be random, white noise, or coherent noise produced by the devices algorithms. It can be caused by low lighting situations, a nearby power interference, heat, or a defect on the CCD (Charge Coupled Device).
Different Types of Digital Noise
To understand noise, we must first explain how an image is captured. The image sensor, located on the CCD, processes each image and sends it to the DVR (digital video recorder). When the area the camera is located in has poor lighting, the image sensor picks up what is called “chroma”, or variations in hue, and luminance, variations in brightness. See Figure 1
Image noise is caused when there is not enough lighting to illuminate the area. It occurs when there is insufficient light reflecting off objects, so it cannot distinguish different colors or different contrast.
Other types of noise include salt and pepper noise, when pixels are of different hues than the pixels around them, causing the image to have dark and white dots, hence the name. See Figure 2
Gaussian noise, or amplified noise, is caused by random interference and causes the every pixel to be changed from its original color. The name is derived from the Gaussian distribution, or the probability density function which is equal to a normal distribution. Gaussian noise amplifies every pixel in the image, particularly blue pixels, which causes image distortion.
Film grain is a another type of noise. It is dependent on the signal of the video. It is the normal grain that is found on videos taken in low lighting. It is given a uniform texturing.
In a nutshell, noise is caused by the sensor when it does not pick up adequate lighting. This causes blending in the image, which results in a grainy effect. This obscures the image and causes it to be blurry, or ghostly.
What is Digital Noise Reduction?
Some surveillance camera companies may include their brands in their innovative DNR technology, i.e. SSNRII (second generation of Samsung’s Super Noise Reduction) or XDNR (eXcellent Dynamic Noise Reduction by Sony). Sony’s XDNR delivers the finest noise reduction in low lit areas and eliminates motion blur. Samsung’s SSNRII removes much more noise than traditional DNR.
When it comes to CCTV, DNR is crucial for clearer images. The image sensor on the CCD eliminates the grainy effect on the images, and consequences in a richer image. This is essential to identify any movement or objects on the screen. It helps to have DNR on surveillance cameras located in parking lots, and it comes in handy for forensic use.
How Does DNR Function?
Digital Noise Reduction utilizes software in the CCD to digitally remove any noise found in each image. It has an algorithm that analyses two consecutive frames and removes any grains that do not match to the previous frame.
When the CCD eliminates the noise from the photo, the image is then transferred to the DVR/NVR. When it is stored on the hard disk drive (HDD), the picture size is decreased by 70%, and is clearer and crisper than before.
However, the image is typically only processed in the foreground of the image. Objects in the background tend to appear grainy. Newer developments have found a solution for this. The most advanced type of digital noise reduction is called 3D-DNR, or as it is sometimes written, 3DNR.
The latest form of Digital Noise Reduction is 3D-DNR. It compares every pixel with the pixels surrounding it in addition to every frame with the next. This process of matching is called spatial noise reduction.
With 3D-DNR, a thorough processing of the image is applied, including the background. This results in a clearer image than traditional DNR and less space is taken on the hard disk drive.