Using White-Light for Camera Analytics

18 October 2022

Camera analytics have improved hugely over recent years and many cameras now have at least some degree of built-in motion or cross-line detection technology. We are also seeing the growth of ‘self-learning’ analytics which have huge potential. However, one thing that holds back the performance of analytics is grainy or sub-par images; even a minimal level of image noise can degrade the accuracy of a detection system. Dedicated, high-quality illumination is used to combat these issues and plays a crucial role in improving images and analytics accuracy.

When specifying lighting for surveillance systems, there is a general understanding that greater distances can be achieved by using Infra-Red (IR) illumination compared to White-Light (WL). This is due to security cameras being more sensitive to 850nm IR. Let’s take a look at the specs of two products from our VARIO2 range of illuminators to demonstrate this;

VARIO2 w8-1 White-Light illuminator
Wavelength: White-Light
Number of LEDs: 24
Consumption: 42W
Max Distance: 180m / 591f

VARIO2 i8-1 Infra-Red illuminator 
Wavelength: 850nm Infra-Red
Number of LEDs: 24
Consumption: 46W
Max Distance: 350m / 1,148ft

As we can see, the equivalent IR illuminator achieves far greater distance – but maybe distance should not be the only factor to consider when specifying lighting. Below we compare the performance of these two light types when used in conjunction with camera analytics.

Comparing Wight-Light and Infra-Red for Camera Analytics

We recently carried out some testing in conjunction with Axis Communications to compare Axis Object Analytics performance when used in conjunction with IR and WL during low light hours. Staged in a typical urban scene using an AXIS P3265-LVE camera supported with illumination from Raytec’s VARIO2 Hybrid 6 illuminator. VARIO2 Hybrid illuminators deliver both White-Light and Infra-Red illumination from a single illuminator.

We tested the performance of the camera analytics using the illuminator’s IR module, then replicated the testing using only the WL module to provide a direct comparison between the two types of light. One might expect the analytics to achieve better detection results under IR illumination, but let’s take a look at the footage side by side to see the results;

What may be surprising here is that the WL achieved comparable results to the IR despite lower claimed distances. So why is this?

Camera analytics rely on contrast to perform effectively. When using Infra-Red illumination, the camera is providing monochrome images and the camera is relying on shapes and motion alone to aid detection. When White-Light is deployed, allowing the camera to capture colour images, the analytics software receives additional information (i.e. different colours) to support detection.

In addition, colour images are often more compelling when being used to support the prosecution of an intruder. The colour of an individual’s clothing or other important details can act as important evidence when identifying and prosecuting an individual.

While we cannot say that specifying White-Light for camera analytics will deliver a specific performance increase over Infra-Red, it may certainly be worth testing the use of a dedicated, professional White-Light system when designing your next install where camera analytics are deployed.

Sign up to the Raytec newsletter