Security & Safety Things on AI and IoT tech in traffic surveillance

Security cameras are not new in traffic surveillance with video technology already deployed to catch speeding, red-light runners and monitor traffic density, in order to avoid traffic jams. With the rise of AI, IoT and smart cities, it's time to think about taking traffic video surveillance to the next level. This article highlights how security cameras equipped with AI can help make roads and highways safer.

Smart traffic surveillance

Traffic sensors are specialised in detecting ‘simple’ traffic incidents, such as cars driving through red lights or exceeding speed limits. They also help detect street conditions quantitatively, e.g. counting passing cars, enabling human operators to recognise traffic jams, before they occur.

Qualitative traffic analysis and control still requires humans. However, this will change soon with a new generation of smart video technology.

Combing IoT cameras and AI video analytics

Smart traffic surveillance combining IoT security cameras and AI video analytics helps authorities

Smart traffic surveillance combining IoT security cameras and AI video analytics helps authorities, traffic planners and highway operators to systematically monitor and understand road users' behaviour, thereby gaining valuable insights from video data to optimise traffic flow and minimise accident risks, automatically and in real time.

Imagine video surveillance systems at busy intersections in downtown areas not only detecting who violates traffic regulations, but also automatically analyses and controls traffic flow. Recognising actual road occupancy, smart camera systems could help manage traffic, for example, by controlling traffic lights that would no longer be time controlled but smart.

Smart security camera systems

IoT Cameras can do more than just video. Security & Safety Things aims to make security camera systems smarter so that they can perform complex tasks independently. With the open IoT platform, including a Camera OS and an Application Store, Security & Safety Things enable camera users and system integrators to equip their camera systems with the latest AI video analytics.

“It couldn't be easier. You simply install the apps you need from our Application Store on your cameras with a few clicks, just as you know it from your smartphone,” said Ajay Kattige, Solutions Consultant at Security & Safety Things.

Smart video analytics in traffic monitoring

Security cameras are designed to take videos and make it available for evaluation. To get more out of video footage, cameras need video analytics capabilities, for example, to distinguish people from cars, analyse traffic flow or ‘read’ licence plates.

In the Application Store, users can find a wide range of apps, easy to install on security cameras and ready to perform a variety of business tasks.

Licence plate capture in bad light and weather

NeuroCar LPR EU is a camera app designed for detection and identification of car number plates in motion

Recognising the license plate of a fast passing car is very difficult, especially in poor light conditions. Where the human eye reaches its limits, high-resolution cameras empowered with smart video analytics can help identify speeding or control access systems in streets or parking lots. Traffic surveillance becomes both more reliable and more efficient.

For instance, NeuroCar LPR EU is a camera application designed for detection and identification of vehicles and car number plates in motion. The app is not only trained to recognise sequences of letters on plates, but also to country or region of origin.

AI video analytics to analyse street layouts

Sustainable urban planning needs long-term traffic data, in order to make the right decisions for necessary changes or adjustments in road layout and control. If more than random samples or snapshots are needed, smart surveillance cameras equipped with AI video analytics can provide valid data over a long period of time. Additionally, they turn data into valuable insights.

For example, GoodVision Connect captures and digitises traffic insights from activity on roads and junctions using proprietary AI and provides powerful tools to obtain tedious traffic behavioural parameters in a fully interactive way. This helps transport surveyors and modellers to achieve more productivity and quality in traffic data analytics.

Real-time insights to detect traffic anomalies

Traffic on busy roads is complex and often difficult to understand because of its dynamics. To avoid traffic jams and reduce the number of accidents, experts around the world are studying the laws of the road. As described above, traffic data is crucial for identifying systematic problems and taking appropriate action to solve them. Smart surveillance cameras can provide this data.

Isarsoft Perception offers crowd counting, occupancy monitoring and cross line detection for traffic analysis

For instance, Isarsoft Perception offers crowd counting, occupancy monitoring and cross line detection for traffic analysis in a single application. The camera screenshot illustrates how cross lines, drawn directly into video, are used to detect, for example, if vehicles stay in their lanes or tend to change lanes frequently (which can cause traffic jams).

FLOW Traffic is another app that enables security cameras to detect traffic anomalies and find the reasons why traffic is not moving as it should. Traffic can be continuously analysed according to a variety of criteria, such as object types, properties, types of movement, speeds, etc. Controlling and configuring the app is amazingly simple.

Monitoring unauthorised traffic

Traffic isn’t just about flow of vehicles and pedestrian motion. There are also hidden aspects that cannot be observed easily or even automated using conventional video cameras. A vehicle that drives over a red traffic light is easy to recognise, but a vehicle that is parked in a parking lot longer than allowed often remains undetected. The same applies to parking spaces reserved for disabled people, with parking space surveillance a complex and costly affair.

ParkingProtect uses computer vision methods to identify when a parking violation has occurred. Once a violation is detected, a notification with a snapshot of the event is sent to parking space operators or authorities. By using ParkingProtect, surveillance cameras determine whether parking time restrictions are met or whether people leaving their car in a store's customer parking lot actually shop there. This helps retailers keep customer parking under control.

There are many ways to take traffic surveillance to the next level and make Smart City traffic safer & free flowing, using smart video technology.