Video analytics has transformed today’s security cameras. No longer are they passive systems that simply observe and record events before making footage available for playback. Video cameras can now use metadata to add sense and structure to recording, which means they have a basic understanding of what they are observing. Not only can they tell the difference between false alarms and potential threats to notify guards the moment an alarm rule is met, they can also retrieve the right footage from hours of stored video in a faster and more efficient way, making searching for evidence easier than ever before.
For many businesses, the advantages are clear. Firstly, video analytics takes the human element out of the equation. Watching live footage over multiple feeds is not a guaranteed way to ensure the highest levels of security are met. In fact, after just 20 minutes of observing live footage, 90% of screen activity is missed.
Secondly, because they can provide all sorts of statistics in terms of metadata, cameras with built-in video analytics can be used to analyse scenes for reasons that go beyond security – from people and vehicle counting to detecting backlogs in production lines and loading docks. That makes it possible for businesses to increase efficiency, reduce costs and make smarter decisions that can be measured via tangible results.
There’s no doubt that video analytics substantially improves levels of security and business intelligence. But with a standard set of alarm rules for detecting and reporting on situations, does video analytics go far enough?
Train your brain
Video analytics is often described as the ‘brains’ of a security system. So what if the ‘brains’ behind our security cameras could be trained to improve their cognitive ability to pay attention, learn, and problem-solve according to specific rules and situations that matter most to individual customers?
The result is Bosch’s new Camera Trainer technology. The next logical step in high-end video security, it harnesses machine learning to enable built-in Intelligent Video Analytics to function in ever-more accurate and application-specific ways.
Camera Trainer gives Bosch cameras the ability to identify new objects and situations within a scene that are defined by the user with an exceptional level of precision. Once an instruction has been initiated, the camera retains this information, becomes familiar with it, and refers to it when processing future scenes. New instructions can be added to the camera at any time, and these can be combined with the pre-determined alarm rules and object filters that are already embedded for even greater flexibility and accuracy.
Then there’s the question of motion. Security cameras are adept at triggering alarms based on movement detection. But what if the criteria you are interested in is a stationary object, or the absence of an object altogether? Once again, Camera Trainer takes situational awareness to new levels. Cameras can be taught to recognise and detect stationary objects, delivering data when they are present, but also when they have been removed. This provides far more informative data that can be translated into improved security, business efficiencies and revenue streams that meet users’ specific requirements.
A few examples
In parking applications, knowing how many bays are occupied would normally require a series of ground sensors or similar. But with Bosch security cameras enabled by Camera Trainer, you can teach your system to recognise occupied and free parking bays. Business intelligence that saves on sensor costs and optimises traffic flow.
Because it can identify stationary objects, Camera Trainer can also be used to detect how long a vehicle has been parked in a time-limited space. The alarm is delayed for the permitted parking period. Once this has been exceeded, an alert will inform parking operators on the ground to investigate the situation – a simple, yet effective way to enforce parking violations and boost revenue streams.
Vehicle counting is another application that is becoming increasingly important in smart cities. By training cameras to detect cars, trucks, buses or motorcycles driving in a certain direction, municipalities can benefit from a highly reliable traffic counting system that distinguishes between in-going and out-going traffic. This enables them to adjust traffic signalling systems and change vehicle priorities to prevent traffic jams and optimise the flow of traffic in and around the city.
But it’s not just traffic applications that can benefit from Camera Trainer; it can also be used to improve business efficiencies. For example, in commercial and industrial warehouses, the status of loading bays has a direct impact on how efficiently goods and materials are handled. By training the camera to recognise a physical mark on the ground, video analytics can be used to detect when there is no truck in the loading bay. The next truck can then be sent to the bay for unloading, increasing efficiency and saving on costs.
The same logic can be applied to heavily controlled transport environments, such as airports, where the awareness of traffic flow and plane status is critical to safety and security. Camera Trainer can be used to deliver detailed information on situations that can be pre-defined by the airport operator. For example, it can collect data on how long a plane is parked at a terminal, the time it takes to load passengers’ luggage, or how long it takes for a plane to re-fuel. This information can then be used to speed up the turnaround time for passengers and planes, improving efficiencies at the airport and reducing costs for airlines.
Warehouse logistics is another example of how Camera Trainer can be used to help businesses to optimise their operations. By setting an occupancy alarm on conveyor belts, it’s possible to detect when there are too many boxes on a production line and alert operators with an alarm. This avoids pile-ups that can result in costly downtime and require backlogs to be cleared before production can recommence.
Camera Trainer also comes into its own for applications that would be beyond the capabilities of standard video analytics. For example, in extreme climates, icicle formations on buildings can cause untold damage – and pose a threat to the safety of pedestrians walking by. A pan-tilt-zoom (PTZ) camera can be trained to scan multiple parts of a building by configuring alarm rules for up to 16 presets. This enables icicles to be detected and removed in the early stages of formation, eliminating costly building damage and potentially saving lives.
Free training licence
There’s no doubt that video analytics has driven innovation in security solutions. But with the introduction of Camera Trainer, it is truly proving its worth. This machine learning technology is available on all Bosch cameras with Intelligent Video Analytics and is activated via a free licence in the camera. Configuring and training the camera is quick and easy, and no calibration is necessary. Furthermore, no additional integration is needed for Bosch Video Management System (BVMS), third-party, or ONVIF support.
For system integrators, Camera Trainer offers the most flexible way to tailor security solutions to a customer’s individual requirements. For end-users, it’s a digital eye on their world with the brains to match. But for the world of security, it’s a whole new ball game. It offers the peace of mind that smart data can be used to provide valuable insights that go way beyond protecting people and property, and deliver business intelligence that ultimately improves efficiency and performance.
|Tel:||+27 11 651 9600|
|Fax:||+27 11 651 7811|
|Articles:||More information and articles about Bosch Building Technologies|
© Technews Publishing (Pty) Ltd | All Rights Reserved