Superior Automatic Incident Detection (AID) with Multi-Sensor Technology

OPAX’s real-time automatic incident detection (AID) solution provides significantly improved traffic tunnel safety through an innovative combination of video and radar sensors in a single platform.

The solution offers weather and light-independent radar detection in tunnel entrances and exits combined with stateof-the-art video analytics covering the tunnel’s inside. Working independently, video and radar sensors consistently classify objects, offering improved reliability through sensor redundancy, higher precision and lower false-alarm rates. The flexible and cost-effective client-server architecture uses standardsbased, interoperable components that are easily integrated with existing infrastructure.

 

WHY CHOOSE OPAX DIGITAL SOLUTIONS

Currently, AID systems primarily use analogue infrastructure and video. The alternative is a digital, network-based infrastructure with digital cameras and video. They entail high-resolution (megapixel) cameras, consistent image quality regardless of distance between cameras and a receiving unit (e. g. screen), Power over Ethernet (PoE) and wireless functionality, improved scalability and redundancy. Today's network cameras, computers and network technology, network camera-based network video systems are not only a viable alternative to analogue systems, but can vastly improve the cost-efficiency and quality of AID.

By using both commercially available, off-the-shelf (COTS) hardware and standards-based, interoperable, networking components, the cost of installation and maintenance is lowered while the risks associated with investments in proprietary technology are eliminated. A network-based infrastructure also enables centrally placed, physically protected and redundant, scalable processing units, supporting single-point software updates and maintenance.

The varying lighting conditions (backlighting from the sun, contrasting shadows and fog) pose significant issues for camera-based AID sensors at tunnel entrances and exits. A possible solution is to supplement camera sensors with radar technology. Radars have successfully been used in the military for decades and the main hurdle for civilian has been the high costs associated with the technology. With improved manufacturing efficiency, affordable radar sensors are now commercially available.

OPAX A/S, a leading vendor of high-end video content analytics (VCA) software, has merged the best of VCA and radars in a coherent, modern and fully digital AID system that use both camera and radar sensors to improve system reliability and performance.

 

SOLUTION OVERVIEW


1. Tunnel Sensors
Video sensors continuously transmit video feeds and radar units transmit tracking data.

2. Analytic Server(s)
Process data from the sensors and generate alarms if incidents occur (e.g. stopped vehicle, a wrongway driver, or smoke). Employs a number of filters to extract information and generate alarms:
• Traffic analysis – uses video data to extract traffic information such as average speed and occupancy levels.
• Smoke detection – via video.
• Radar analysis – uses radar data to extract traffic information such as average speed and occupancy levels.

3. AID Module
Aggregates the information and alarms from the different filters and work as a communication link between the OPAX’s solution and the overlying traffic monitoring and control system.

4. Control room
Alarms are presented to operators in an existing overlying traffic monitoring and control system.

NB. OPAX also works well with analogue infrastructure but preferably with digital infrastructure.

 

Architecture Benefits

The system architecture creates a scalable solution with natural redundancy, improving system reliability. The centralized analytics engine enables single-point updates and makes it possible to perform several analytics algorithms on a single sensor. Being software-based also enables that service and updates can be performed remotely without disturbing traffic.

 

Benefits – Easy to extend functionality

• High-resolution (megapixel) cameras

• Consistent image quality regardless of distance between cameras and a receiving unit (e. g. screen) - Power over Ethernet (PoE)

• Wireless functionality

• Improved scalability and redundancy

• Several algorithms can be performed on the same camera, in parallel

 

SOLUTION FEATURES

The AID and traffic monitoring features are presented in the tables below. Features regarding camera and radar sensors are also outlined.

 

Table 1 - AID features

AID feature

Camera Sensor

Radar Sensor

Stopped vehicle

X

X

Wrong-way driver

X

X

Pedestrians in tunnel

X

Lost cargo

X

Smoke

X

Queue

X

 

Table 2 - Traffic monitoring features

Traffic monitoring feature

Camera Sensor

Radar Sensor

Average speed

X

X

Occupancy levels

X

X

Vehicle counting

X

X

Licence Plate Recognition

X

 

One of the benefits of the solution is that several algorithms can be performed on the same camera, in parallel. This modular, software-centric, approach makes it possible to add features as needs arise and technologies evolve. For example, using Licence Plate Recognition in combination with AID features such as stopped vehicle and wrong-way driver.

 

Below are a number of images, exemplifying some AID features.

Wrong-way driver (service vehicle)

 

 

Queue alarm per lane. The number in the upper left corner shows the average speed and coverage per lane.

 

 

Stopped vehicle. A car waiting to change lane in rush hour traffic

 

 

Camera Sensors and Analytics

Cameras are the primary sensors used inside the traffic tunnel. Traditional AID systems rely on some sort of background subtraction and segmentation of the video stream from the cameras. OPAX circumvents problems inherent in such approaches by modern image processing methods implemented in the traffic filters. The filters extract traffic data and generate alarms from each video stream from the tunnel.

Instead of relying on background subtraction, the core of OPAXs traffic filter employs a state-of-art method, usually deemed as too expensive to compute. OPAX has managed to craft and optimize the methods to perform in real-time on COTS hardware. It is a noble image processing method, also used in robotics and SLAM (Simultaneous Localization and Mapping) applications.

Image frames from a traffic camera

 

 

Radar Sensors

Radar sensors provide a weather and light independent solution that breaks free from the limits of camera sensors in tunnel entrances and exits, including backlighting from the sun, contrasting shadows and fog.

The SIRS 100 and SIRS-200 are 24 GHz Pulse Doppler radars designed for use in several ITS applications in several market domains, including highway monitoring, urban traffic control and tunnel supervision.

The SIRS -radars continuously detect and track vehicles (moving and stationary) and other objects up to 120 meters distance while measuring their range and radial speed. Based on the detection and tracking data, the SIRS also include in-sensor processing capabilities that produce the relevant application information. Typical output is traffic flow information and alerts for anomalies in the flow, which is used for monitoring and control purposes in traffic systems of various kinds.

The SIRS sensor has built-in diagnostic functions to be able to verify the integrity of the measurements. Most radar parameters in the sensor are software controlled making the sensor highly configurable. All communication to and from the sensor are done via a CAN-bus interface. In addition a digital open collector output is available that can be configured to activate on pre-defined events, e.g., to trig external systems. SIRS-200 fulfils requirements according to ITUs definition of a narrow band transmitter in the 24 GHz band.

 

SOLUTION BENEFITS

Ours is the first AID solution in the world to combine multiple sensors to improve AID reliability. Tunnel entrances and exits, notoriously challenging, are reliably monitored with radar sensors. As the analytics is performed in the same platform, a single management and monitoring system can be used. The solution architecture also improves reliability through redundancy and the ability to physically protect analytics units.

 

COST SAVINGS

• The overall cost is lower than existing AID solutions

• The use of COTS hardware and standards-based, interop erable networking components lower the initial cost of acquisition

• Cost of system ownership is reduced by simplified, centralized management which can also be performed remotely