Video surveillance technology has evolved considerably over recent decades. Analogue cameras connected on dedicated closed circuits have been replaced with network security cameras capable of recognizing objects and behaviours. Video images that used to be recorded on VHS tape are now stored in the cloud and across distributed systems that allow simultaneous access from various locations and with different access rights.
Opportunities and challenges
What it covers
The pace of technology evolution in the network camera market has resulted in dramatic benefits for physical security end-users. Security managers can remotely view locations on mobile devices, create alerts to various incidents or events, and clearly view activity with better image quality and resolution. The next evolution of the camera market will bring further artificial intelligence (AI) and connectivity benefits. As in any technology evolution, there will be opportunities and challenges.
Higher resolution cameras will require solutions to reduce bandwidth, government regulations will demand technologies to protect privacy and operational responses will be needed for future integration challenges. Increasingly, these solutions will need to be on the same camera. There is market opportunity for vendors that can meet these challenges.
- The evolution of network camera technology represents an opportunity to meet new physical security end-user challenges with new technology solutions. Challenges can be technological, network based, regulatory, related to image quality and the physical environment or operational. Solving these challenges is both a business and customer success opportunity for the video surveillance industry.
- Individual security cameras will be required to mitigate multiple end-user challenges and provide unique solutions all on the same device. This approach of combing technologies on one security camera will make installation more efficient and ensure a more integrated solution.
Network security camera market evolution
Video surveillance cameras have benefited from the rise of IoT (Internet of Things) market applications. Over the last decade, low resolution analogue cameras have been replaced with higher resolution networked cameras in almost every conceivable deployment scenario.
Furthermore, the shipment volumes for network cameras have increased as new technological features are launched. Omdia forecasts that one billion professional grade security cameras will be installed and operational globally by the end of 2022, with most of the new deployments in the network camera category.
The next stage of network camera technology is the artificial intelligence era. Primarily driven by object detection and behaviour analysis, the generated metadata will enable AI to provide other functionality. Examples include image coding enhancements, big data analysis and predictive crime centres, remote maintenance assessments and image quality enhancements.
The network camera market will also build on the IoT trend to provide more than just physical security functionality. In retail, business intelligence is already a recognized application type. In the safe and smart city markets, increasingly, the network camera will act as the vision of the system helping deliver new operational functionality in applications such as parking management, waste management, traffic management and building code enforcement.
Overall, these emerging technologies have clear benefits to a physical security solution. For end-users and systems integrators looking to install and operate this technology, it is critical to identify the key applications for each challenge, as well as how best to combine these technologies.
Security end-users face an extensive list of threats to their people, buildings and physical sites. These threats range from identifying and intercepting intruders, managing risk and safety, protecting remote locations from attack and regulated responsibilities to meet government legislation or insurance requirements.
In response, end-users have turned to video surveillance equipment to help mitigate these threats. Network cameras provide a range of solutions to meet each threat type. The following section discusses some of the challenges end-users face followed by examples of how network camera technology can help to make a security end-user or systems integrator’s life easier.
As network camera technology has evolved, the average selling price of a camera has declined. In turn, this trend has driven adoption across new customers and end-user industries. One consequence of the increased installed base of cameras is that it has become harder to search through the data and images created by these cameras. Manually searching video feeds has become labour intensive and expensive.
A key evolution in this technology has been the move from solely centralized analysis, on the server, to increased use of compute at the edge. AI metadata analytics can filter out data that is not important and ensure human operators are as efficient as possible in post-event identification and analysis. AI adoption in security cameras is forecast to be a key technology driver over the next five years and this application is core to its success.
It is worth noting that AI and deep learning algorithms can also be used to mitigate other difficulties such as false alarm rates. In many regions, central monitoring stations and police response is based on a verified alarm in order to avoid costly false responses. AI algorithms embedded in network security cameras are well placed to support this requirement.
The increasing number of cameras installed has created another challenge for end-users: networking and bandwidth capability. Cameras are increasingly higher resolution. In fact, Omdia predicts that 65 percent of cameras shipped in 2024 will be 4-megapixel resolution or higher. Bandwidth is therefore crucial in supporting these camera feeds as they are networked to the storage solutions and operational centres.
Increasingly, coding is important in mitigating bandwidth concerns. For a full HD (1080p) camera, coding evolution has brought down the bitrate requirement from 16 Mbit/s for H.263 to 4 Mbit/s for H.265 high efficiency video coding. This means bandwidth costs can be managed and more cameras can transmit over a limited network connection.
AI technology can also play a part in mitigating this challenge. Algorithms can identify and prioritize what is changing in each video frame. Areas of an image which do not change or are not deemed important do not need to be re-encoded in each video frame. Depending on the activity in the scene and the amount of new information required to be encoded, using AI powered encoding can dramatically reduce the required bandwidth.
Privacy and protection of personal data has received more attention in the past decade and is therefore a relatively new challenge for physical security end-users. Regulations such as the European Union’s General Data Protection Regulation (GDPR) addresses the storage and transfer of personal data. It is designed to give the individual control over what information is stored about them and how it is used. Video surveillance system owners and operators, under jurisdiction of GDPR and similar privacy laws, have a duty of care not to store personal information unless they have a lawful purpose for doing so.
Images of a person’s face is considered personal data. This type of image is regularly collected and stored by security cameras posing challenges to surveillance owners and operators. In some scenarios, solutions that can redact the face image from the scene can be valuable in helping to ensure a video surveillance solution is compliant with GDPR regulations. Artificial intelligence algorithms have the ability to identify faces and apply some form of masking to protect an individual’s privacy in these stored video images.
As already mentioned, security camera resolution has improved throughout the history of the industry. Presently, most network cameras installed are capable of HD (High Definition) resolution. This is great in day light. However, threat incidents often occur at night when it is much more difficult to view, recognize and categorize objects and people. Image sensors have a trade-off between higher resolution and low light performance. As more pixels are squeezed into sensors to provide higher resolution, the pixels are smaller and less sensitive to light.
Many network cameras are shipping with advanced low light functionality to help end-users in these conditions. Technologies can be designed to improve either the optical signal received or reduce the noise on the sensor – optimizing the light path at the lens and filter. In the past few years, the marketing of advanced low light functionality has been characterised by the solutions ability to provide colour imaging in low light rather than traditional monochrome low light images. Finally, illuminators, such as onboard or remote LED lighting can be used to help the security camera detect and deliver a clearer image in difficult conditions. These solutions combine to improve the image quality in challenging conditions ultimately improving the security solution.
Some physical security end-users are responsible for securing sites with a large geographic area. In the energy, transportation and industrial verticals, these locations can be vast and might not have defined fencing or perimeter security equipment located at all locations. This makes it challenging to protect, but it also creates another problem; how do you respond to an incident that is a significant distance from your operations team but requires a quick response?
One solution is increasingly being built into the network camera. Audio capability gives the operations team sitting in the command and control centre the ability to directly communicate with an intruder in real-time.Bi-directional audio capability is now found on most network cameras shipped globally. This type of system allows the operator to directly interact with the intruder as well as to hear their response. Uni-directional systems allow the audio signal to move in one direction only. In some cases, these cameras have in-built audio capability but still require separate speakers and microphones. However, incorporating the audio on the camera means a less complicated integration and system operation.Another important approach is the use of a warning light to respond during an incident and to warn off intruders. Compared with white light, flashing blue and red lights are generally more visible in mist and fog scenarios. Again, integrating this on the camera can simplify the install and operation.
As video surveillance systems add more technology features, they also become more complex to integrate. Furthermore, systems from various equipment manufacturers, software platform vendors and other channel partners may be needed to support with video analytics, image improvement technology, audio features and other functionality.
Including multiple technology solutions on one device has several benefits. First, the integration is pre-installed and is more likely to work effectively. Second, the installation time and complexity are lower, meaning security systems can be deployed more quickly. Third, new functionality can be created through combining these technology features on the same device. Equipment vendors have the potential to build new solutions by combining technology that already exists on the security camera to meet new and emerging applications.
Dahua’s TiOC approach
Technology features on one device
Many of the technology features discussed in the white paper are provided by video surveillance equipment vendors active in the market. However, Dahua has a built a new camera that uniquely combines three key technology features on the same device: full colour imaging to improve vision; deep learning algorithms for improved detection; and warning light functionality to provide remote alerting and warning to intruders.
Dahua has developed TiOC, also known as three-in-one camera series, to meet multiple end-user challenges on one device. TiOC integrates 24/7 full-colour monitoring, active deterrence and AI into one solution.
Full-colour technology means 24/7 colour monitoring which increases the probability of collecting valid human, vehicle, and event evidence. Adopting a F1.0 super large aperture, the camera collects 2.5 times the amount of light compared with a F1.6 aperture. More light means brighter images. Furthermore, the integrated Dahua ISP 4.0 and high-performance sensor (30% increase in photosensitive capacity using a larger sensor part) improves the Signal to Noise Ratio (SNR), provides high colour reproduction and low light performance for a more vivid and brighter image.
TiOC supports SMD 2.0 and Perimeter Protection (both AI algorithms) which help filter out false alarms triggered by non-concerned targets. They can also recognize humans and vehicles more effectively. Both functions help keep false alarm rates below 2%. SMD 2.0 also increases the detection distance by more than 185%. These advantages reduce the customer’s costs.
Dahua’s siren and light active deterrence technology supports light alarm and voice alarm when an incident occurs to provide effective intervention. Both SMD 2.0 and Perimeter Protection support accurate classification and detection of humans and vehicles and can intelligently analyse multiple types of intrusions.
Different types of sounds and warning lights can be set according to defined, recognized intruder classifications in the TiOC. This can be used to better deter targets and provide effective prevention of incidents. Moreover, the siren and light alarm technology also support the configuration of different voice formats set by the user as well as warnings for different alarm rules. The sound volume can be adjusted so that the device can be flexibly applied in different scenarios.
In the same installation, alarms configured for different time periods can provide different effects. For example, during business hours of a shop, the voice can be mild and welcoming. During off-hours the warning audio can be used. Consequently, one camera can be used for multiple purposes giving a more economical solution.
Villa case study
TiOC is suited for public locations as well as private buildings such as shops and warehouses where the cameras are installed to maintain order and warn off intruders. In a villa project in South Africa, TiOC has been deployed to address the following issues:
- Identify suspects in low light conditions. Reliable information is key during evidence retrieval. The Dahua TiOC can deliver colour images in day and night, improving the ability to identify suspects, while increasing the quality of video evidence. According to the end-user, there had been incidents of theft in the villa before the system was deployed. With traditional infrared cameras, there was limited feature information captured, such as general appearance or a description of the intruder’s clothes. This made it difficult to provide a positive result in the investigation. Using TiOC, the night imaging ability was able to capture the clothing colour of the people being monitored which allowed the police to better trace the target, improving the chance of solving the case.
- Quick response when an incident occurs within a large geographic area. Instead of relying on the recorded video as evidence after an event, the TiOC allows a response during the event with its Active Deterrence and Real-time Alarm functions. These features help users to feel safer as they can directly respond to an intruder. If the intrusion occurs anywhere around the villa, TiOC will trigger a sound and light alarm and push real-time alarm messages to the end-user’s smart phone so they are able to view the events. For intruders who still do not leave, a two-way talk function can be used to call the intruder and address them directly. The red and blue flashing warning lights in the TiOC are clearer than white light, even in dense fog and heavy rain.
- Filter out false alarms and quickly search targets. In traditional security alarm systems, time can be lost on false alarms. Users often receive lots of alarm messages within a day however only a small number are actual events. According to feedback from the villa site owner, previously, small animals crossing the fence at night would trigger an alarm. Each event would require to be checked, which was often late at night and tiring to do. False alarms have been reduced by TiOC which is saving the end-user both time and energy. If necessary, they can also quickly locate a target in a mass of recorded video footage after an event.
Advances in network camera technology have increased adoption of video surveillance cameras. They have also helped security end-users and systems integrators better manage the various physical security threats that are ever-present. These threats can be met by technology that exist in the market now. However, they can be more effectively met by integrating multiple technologies and functionality on the same device so that one product can support the technological, networking, regulatory, image quality, environmental and operational challenges found in network camera deployments.
- End-users have different threats depending on the industry and risk level. However, security cameras are often required to deal with similar challenges whether they are deployed by transportation, utilities, commercial or retail customers.
- Building multiple technologies into one camera represents an opportunity for end-users to protect their facilities and respond to threats in a more integrated and cohesive way.
- It also represents an opportunity for equipment vendors to meet different end-user challenges without the customer having to buy different devices. This increases the efficiency of the installation and can reduce overall costs.