Join or Sign in

Register for your free asmag.com membership or if you are already a member,
sign in using your preferred method below.

To check your latest product inquiries, manage newsletter preference, update personal / company profile, or download member-exclusive reports, log in to your account now!
Login asmag.comMember Registration
https://www.asmag.com/project/dahua_wizcolor_low_light_cctv/
INSIGHTS

Product testing: Unisinsight’s AI network camera for ITS applications

Product testing: Unisinsight’s AI network camera for ITS applications
Unisinsight’s s AI checkpoint capture IP camera uses infrared strobe lights and a dual sensor architecture to effectively capture video at night. This makes the camera especially ideal in intelligent transportation systems (ITS).
Unisinsight’s s AI checkpoint capture IP camera uses infrared strobe lights and a dual sensor architecture to effectively capture video at night. This makes the camera especially ideal in intelligent transportation systems (ITS). This note takes a closer look, based on product testing by a&s China.
 
Increasingly, video plays an important role in smart transportation. However, nighttime capturing remains a challenge. Without proper illumination, video may become underexposed, blurry and ultimately usable.
 
In order to capture traffic footage clearly at night, video cameras generally use fill lights. Although they can provide sufficient illumination, they also cause “light pollution” – drivers may become blinded by the light and get themselves into accidents. Government agencies often receive complaints about this.
 
In response to this, Unisinsight has launched their AI checkpoint capture IP camera which, as opposed to similar products in the market, replaces the night strobe white light with infrared strobe light, thereby minimizing light pollution and, more importantly, effectively avoiding interference with the driver's line of sight.
 

Aluminum alloy shell

 
The camera adopts a traditional box design. The aluminum alloy shell not only ensures the camera’s stability and heat dissipation capacity, but also allows the camera to have an IP66 protection level, which enables the equipment to work properly even in harsh environments. The device’s front-end protective glass adopts a tilted design that effectively avoids the “ghosting” problem, which makes monitoring difficult and can cloud the AI’s judgment as it analyzes the video.
 
Two image sensors are used, both 1.1 inches in size. This enables larger pixel capacity, rich-color video capturing and recording 4096 × 2160 at 25fps. At the same time, the camera's two supporting lenses measure 30mm and 50mm. When used in smart transportation, the 30mm lens can cover three lanes and the 50mm lens two lanes.
 
At the base of the camera are six warm strobe lights. As mentioned earlier, the camera uses infrared light, which is invisible to the naked eye and prevents the driver from being blinded by the light.
 
The camera is also equipped with a TF card slot, which supports up to 256GB TF card, effectively avoiding video loss during network downtime.
 
The back of the camera is equipped with ports for network, USB, GPS, storage, data communication, I/O, alarms and power supply. This allows the camera to be used in a variety of scenes and environments.
 
It is worth noting that the camera also supports China’s famous Beidou positioning technology, which is in line with the country’s future development trend.
 

Dual-sensor architecture

 
Today, lowlight capturing typically employs a single-sensor dual-frame fusion technology, supported with fill lights. However, the brightness of the fill light is still high, causing light pollution. At the same time, the dual-frame fusion technology is not suitable for highway scenarios, primarily because the positions of objects in frame 1 and frame 2 may be different. Fusing them will cause the image to blur or smear. In complex situations, such as vehicles changing lanes, this problem will only become worse.
 
The checkpoint capture camera, on the other hand, adopts a dual sensor architecture – it has two built-in CMOS sensors. The lens uses a splitter to split light into visible light and IR light, each going to a different sensor. The resulting signals are then fused together via Unisinsight’s algorithm. This solution allows the camera to capture nighttime images with minimal light trails and noise. The use of low-level stroboscopic + infrared fill light makes lowlight capturing even more effective, with minimal light pollution.
 
The dual sensor architecture allows effective lowlight capturing.

 
In order to meet the needs of different environments, Unisinsight also provides an all-in-one fill light, which consists of infrared strobe, LED strobe and xenon lamp; the user can choose whatever mode that works best at the time. The LED strobe is the default setting.
 

AI capability

 
Last but not least, the camera’s AI capability is not to be ignored. a&s’s testing finds that the camera is equipped with Unisinsight’s self-developed algorithm, which can accurately detect and recognize pedestrians, motor vehicles and non-motor vehicles in the video.
 
It is worth noting that the AI can not only identify vehicle details, such as license plate number and vehicle type, make, model and color, but also vehicle behavior – for example speeding, under-speeding, pressing the white/yellow line, driving in the wrong direction and changing lanes illegally. The camera can also recognize other objects/scenarios such as thrown objects, roadblocks, accidents and congestion. It is also worth noting that the camera has a built-in deep learning chip, which can gradually enhance the accuracy of AI.
 

Product advantages:

 
1. Two 1.1-inch image sensors enable a dual-light fusion technology to comprehensively enhance night video capturing;
 
2. The body is equipped with 6 warm strobe lights and can be connected to an external all-in-one fill light to boost illumination;
 
3. Supports online algorithm upgrade and dynamic loading; video capturing will not be interrupted during this period;
 
4. Supports GPS and Beidou positioning functions to meet the future development needs of the industry;
 
5. Comprehensively meets the needs of traffic checkpoint users by extracting feature values for motor vehicles, non-motor vehicles and people.
 

a&s review:

 
Video has become a critical component in intelligent transportation. However, traditional monitoring systems suffer from certain shortcomings and limitations when it comes to lowlight capturing. Advanced solutions are needed in this regard
 
With an innovative image sensor structure, Unisinsight’s AI checkpoint capture camera effectively captures video in lowlight conditions, producing images with minimal light trails and noise. The camera’s low-brightness strobe + infrared flash technology, coupled with supplementary lights, can effectively reduce light pollution, protecting driver safety.
 
In addition, the camera is also equipped with Unisinsight’s self-developed algorithm, which can effectively identify pedestrians, motor vehicles and non-motor vehicles in the video and automatically record violations and report them to the backend. In ITS applications, the camera can also determine traffic flow, road construction and other scenarios to comprehensively help with traffic management and reduce traffic congestion.


Product Adopted:
Surveillance Cameras
Subscribe to Newsletter
Stay updated with the latest trends and technologies in physical security

Share to: