Love them or hate them surveillance cameras are all around us and are here to stay. Patrick Pirim, CEO of Brain Vision Systems at BVS explores how bio-vision smart cameras are providing an extra visionary sense to the market
The global video surveillance market is booming with an accumulated growth of 13 percent in 2011 and over a 9 billion dollar turnover according to the market research firm iSuppli Corp.
This dynamism is due to a wealth of new technologies including, data decentralisation, high-resolution cameras and increased power of calculating processors. This is also explained by the extension of video security applications to new sectors such as hospitals, museums or industrial firms.
Ensuring efficient surveillance while managing an important fleet of cameras is a challenge for all parties involved in video surveillance.
The human brain is the most able system to distinguish a potentially dangerous intrusion from a negligible event. To help the operator, various decision-making tools have been developed in order to gain time searching for events and preventing false alarms.
Smart cameras bring high value-added skills to meet these needs including:
• Analysis of images – detection of movements, colours or shapes
• The distance control and setting of objects owing to an IP or digital network
• The distant dialogues between objects
These new abilities in cameras create the issues of data weight; the higher the image quality, the less data can be stored in a unique and small processor unit.
The fact that information flows are becoming increasingly vital is spurring current trends including the transmitted image size, reaching now 1 or 2 megapixels, with this set to grow.
“The use of Brain Vision Systems sensors is less expensive and more efficient. It also brings services in the field of the Internet of things, which needs a strong perception of the environment so that objects can interact”
Most of the current installations use centralised data storage in a smart central server, which for most of the time is slow and expensive. What is at stake by a firm using video surveillance is better management of data storage.
The decentralisation of meta data inside the smart camera, which incorporates its own processor and autonomous data analysis system, is an interesting solution. The camera then has an embedded perception and understanding capability.
Given the complexity of needs on the market, Brain Vision Systems has invested much time and development in creating their bio-inspired perception systems, the BIPS.
These smart cameras and OEM products include special electronic components featuring similar skills to a human brain such as, perception, recognition and action command.
It detects and analyses in real-time various flows: colour, movement direction and speed, oriented edges, curves and texture of objects.
The firm commercialises standard products, each of them meeting the requirements of specific applications.
The company’s BIPcam smart camera is linked by USB port for studies and prototyping whilst the BIPeye smart camera communicates by CAN for robotics motors. Additionally the smart camera option includes an embedded LINUX operating system with an external Ethernet link for autonomous applications.
Application case example
As a case example, monitoring the flow of passengers on a bus with a smart camera (figures 1& 2) needs to take into account obstacles for the reliability of data while maintaining a competitive price: a close location, with electric pollution or lightening variations. The BIPS process includes algorithms able to compensate the low light density of the environment.
Analysing an over-taking scene on a motorway requires the characterisation in real-time of the type of vehicle and to represent its geographic position (Figure 7). In general, the image is directly sent to a centralised external processing unit and a heavy data transfer may be difficult if many sensors must be managed.
The company’s Brain Vision Systems cameras can process the whole data in situ due to the perception processor BIPS. As a result, they only send important information, reducing the amount of information sent. Their low energy consumption allows the use of a small solar panel associated to a Wi-Fi module so that the system works autonomously.
It is then possible in a city or town location to manage a large fleet of cameras with low performance processing.
In addition to this, the company is currently testing a process of individual characterisation, which means objects can be recognised by their silhouette. This can be adapted to many applications for example, it is possible to activate traffic lights when a bus approaches, or to keep them red if required to stop traffic.
If in a tunnel, characterisation allows the mapping of road traffic and statistics on the type of vehicles at any time. It can facilitate the intervention of police and rescue teams in emergency situations such as accidents.
In museums or public access buildings, the interest of visitors can be measured by monitoring the flow and paths taken whilst at the same time providing video surveillance.
The use of Brain Vision Systems sensors is less expensive and more efficient. It also brings services in the field of the Internet of things, which needs a strong perception of the environment so that objects can interact.
The user of a video surveillance system can then benefit from a rapid return on investment.
BVS
