Security for Multi-Sensor Cameras: Focus Interview

1. For what purpose are multi-sensor cameras best used and what advantages do they offer?
2. Which model out of your offering shows the best performance?
3. Is it a challenge to run video analytics on a multi-sensor device?

Edwin Roobol Director Marketing EMEA, Axis Communications

1. At Axis we offer two different types of multi-sensor cameras. Our multi-directional Axis P37- and P47-Series and our panoramic multi-sensors, represented by Axis P38- and Q38-Series. The first comes in 2- or 4-sensors that can be individually positioned into a direction of choice and can cover up to 360 degrees. The second type exists out of three or four sensors that are seamlessly stitched together into one 180-degree horizontal view for complete situational awareness. The big advantage of these types of cameras is that they can cover wide areas in excellent detail. They could be a good alternative to single-senor cameras as both product types come with the big advantage that they require only one license, one cable, and one IP address which can substantially lower costs in an installation.

2. It is tough to name just one model as every camera has been built for a specific purpose or use case. Some customers want high resolution, like our 5-megapixel Axis P3719-PLE and Axis P4707-PLVE camera models, others value motorized positioning and high pixel density like on our Axis Q3819-PVE, versus a non-motorized Axis P38-Series. Therefore, every camera model is designed based on a customer’s needs and values. Providing options is one aspect, but for us most important is that our products are made to last, are cybersecure and be as low maintenance as possible for the customer. No matter which multi-sensor product is chosen, the customer knows it can handle the challenging conditions it will be placed in and have ample support both from a warranty and firmware perspective.

3. A camera’s primary goal is creating scene awareness and analytics can be a great complement to this. For the well-functioning of analytics, image quality under different circumstances is important. This, being one of Axis’ strongholds, caused us to optimize our own System on a Chip (SoC) including a Deep Learning Processing Unit, Artpec-8, for strong analytic performance without image compromise. Many of our new products, from multi-direction to panoramic multi-sensor, are being developed with our Artpec-8 chip as we speak. The first great results we can already see on our new Axis P47-Series, which represent dual sensor cameras, with analytics on both sensors. And soon more products will follow. That said, multi-sensor cameras are often more complex in design than their single-sensor counter parts and therefore take more time in development with regard to analytic features and the right balance of performance.

Kay Chan Product Manager Vivotek EMEA

1. Our camera model MA9322-EHTVL offers comprehensive video coverage from a single mount point. It captures every angle for comprehensive video coverage from a single IP address, helping to reduce the amount of cameras needed and saving on installation time and costs. For high quality unlimited coverage the camera features four independent 5 MP CMOS Sensors with IR illuminators. This camera can provide the most flexibility in surveillance monitoring. Each sensor utilizes a 3.7 to 7.7 mm motorized focus lens and 3-axis design along a circular track to enable full 360° coverage. By adding IR illumination up to 30 meters, no matter day or night, light or dark, it always provides complete video security.

2. The FD9391-EHTV-v2 is a newly revamped, outdoor fixed dome network camera that delivers 3840 x 2160 (4 K/8 MP) resolution at up to 30 fps. This camera features a built-in microphone, I/O functionality, and VIVOTEK SNV, WDR Pro, and AI-based Smart IR III technology to present the clearest image quality in both low light and high contrast environments.
It comes equipped with built-in cybersecurity capabilities, and AI-powered video analytics is able to differentiate type & colour of vehicle as well as people attributes. Its housing is IP66, IP67, IK10+, and NEMA 4X rated, and operational in temperatures ranging from -50 °C to 60 °C (-58 °F to 140 °F), making it ideal for any outdoor surveillance situation.

3. Yes, to some extent, it requires extensive experience in the surveillance field and resulting vast image databases to develop a neural-network powered detection engine for the human form and vehicle type. Vivotek’s deep learning AI engine offers advanced accuracy in people and vehicle detection compared with traditional rule-based designs. This helps reduce false alarms, as well as minimize the time and effort needed for setting up camera configuration.

Uri Guterman Head of Product & Marketing, Hanwha Techwin Europe

1. Multi-sensor cameras typically offer the capabilities of two, three or even four surveillance cameras in a single housing, offering cost savings as there is less cabling and mounting hardware required than if four cameras were used to do the same job. And, as they only have a single IP connection, only one VMS licence is required, and fewer network switch ports are needed resulting in energy savings. Each has a distinct role in an application: two channel multi-directional cameras are ideal for monitoring adjacent areas, for example stairways or L-shaped spaces, while three channel cameras help eliminate blind spots. Four sensor cameras offer a complete 360-degree view without any blind spots.

2. Best performance can be hard to quantify, as it depends on the specific application for the camera. However, life can be made much easier for installers if models equipped with motorised PTRZ gimbals are specified. These will reduce on-site times as they enable engineers to remotely pan, tilt, rotate and zoom the lens’ positions in order to set the camera’s field of view. It also means installers can install the camera even during darker hours of the day and then set up the field of view later in daylight. Therefore, the installation of PTRZ cameras takes less effort and less labour than a standard camera. Hanwha Techwin recently enhanced its premium Wisenet P series with the addition of three new PTRZ 4-channel, multi-sensor cameras.
The 8 MP Wisenet PNM-9084QZ features motorised varifocal lenses for precise control of focal length, angle of view and zoom of each of its sensors, as do the 8 MP PNM-9084RQZ and 20 MP PNM-9085RQZ, which are also equipped with built-in smart IR illumination for each sensor. This illuminates objects at a distance of up to 30 m by focusing the beam in line with zoom.

3. Our three new 4-channel multi-sensor cameras effortlessly deliver built-in Intelligent Video Analytics (IVA), including Tampering, Loitering, Directional Detection, Defocus Detection, Fog Detection, Virtual Line, Enter/Exit, Appear/Disappear, Audio Detection, Face Detection and Motion Detection. Our panoramic multi-sensor cameras include AI object tracking in the stitched image, allowing users to track across the images of all four physical sensors.
For AI to work, sufficient processing power and memory are essential. The application and user interface also need to be good enough to help users apply the right AI applications on the right channel and get effectively notified about an AI event. In addition, integrating with the leading VMS solutions to allow a forensic search on any of the multi-sensor channels is a must-have. Finally, it also helps if the multi-sensor camera supports PTRZ: this ensures the best field of view and angle is set and helps deliver the best forensic images for AI to work with.

Wanda Nijholt Manager, Marketing & Communications i-PRO EMEA & APA

1. Multi-Sensor cameras are best used for environments that require a wide range of visibility, such as street crossings, shopping malls, parking lots or airports. Not only do multi-sensor models cover up to 360° of vision, but also eliminate blind spots compared to several individual cameras. As one single multi-sensor camera basically includes four cameras in one, the number of cameras needed for projects is reduced massively. Thanks to the slim design – the world’s thinnest – i-PRO Multi-Sensor models blend in seamlessly into any surrounding and fit perfectly into the indoor environment, basically as an advanced fisheye camera. In addition, our cameras are designed with a special hook structure which makes the installation quick and easy. All of the above aspects do not only save time, but also money.

2. Our best performing Multi-Sensor model is our GIT SECURITY AWARD 2023 winner WV-S8574L, equipped with four sensors in 4 K resolution. The AI functionality is integrated directly in the camera via the built-in Deep Learning AI processor which offers, in addition to the pre-installed free applications, also the possibility to integrate 3rd party applications – with all sensors able to be configured individually. For use in dark environments, the WV-S8574L has IR LEDs. In addition, the camera is fully NDAA compliant and FIPS 140-2 Level 3 certified. The WV-S8574L is part of our recently launched Multi-Sensor Range, available with three or four imaging sensors in 4 K, 6 MP and 4 MP resolution, ensuring exceptionally detailed image capture for 180° or up to 360° vision adaptable to the installation environment.

3. All i-PRO Multi-Sensor models are designed for edge computing and are equipped with a powerful built in AI processor in line with our mission to make AI the standard in the security industry. This allows video and image data to be processed and analysed on the cameras themselves and eliminates the need for additional server infrastructure and reduces the load on the network. Our AI-capable cameras classify the image data and then only pass the metadata via the Active Guard Plug-in on to the Video Management System. To make best use of the metadata that is generated, our multi-sensor cameras are provided with a one cable only connection and require only one VMS license. Active Guard is already supported by leading VMS providers, such as Genetec and Milestone.

Hundt Huang Product Manager at Hikvision

1. Multi-sensor cameras are best used for applications that demand broader coverage, where both panoramas and detailed, zoomed-in views of the same large site are needed simultaneously, for example, parking spaces, squares, stadiums, and other outdoor areas. As these cameras have multiple lenses built into one camera body, they can reduce the total number of cameras required, helping minimize both cost and time spent on installation and maintenance. Integrate two or more imaging lenses into a single camera, there is only one IP channel to manage, and only one license is needed on the video management software platform. Besides, the smart-linkage between multiple sensors and lenses makes the configuration and management of multi-sensor cameras easier, compared to a system with several single-sensor cameras.

2. That would be the Hikvision 8-inch TandemVu PTZ DS-2SF8C442MXS-DL(14F1)(P3). It combines a 4 MP fixed-bullet camera with ColorVu technology and a 4 MP PTZ camera with DarkFighter technology. This means that the camera can monitor large areas and zoom in on specific security incidents while maintaining focus on both ‘viewpoints‘ simultaneously.
Featuring Hikvision’s signature ColorVu and DarkFighter technologies, the cameras are able to provide high-quality, full-color video 24/7, even in extreme low light conditions. The camera also supports IP67 and IK10, which guarantees good water and vandal protection in various scenes.

3. It is indeed. Firstly, data fusion and processing from multiple sensors require higher performance in both hardware and software, compared to conventional single-sensor cameras. Secondly, for multi-sensor fusion, it is essential to perform the online calibration of sensors to accurately recognize changes in sensor pose or changes in the positions of the sensors. Thirdly, the intelligent and persistent analytics working within the smart linkage between different lenses is very complicated and challenging.