- Restores confidence in traffic video analytics,
- Deep learning algorithms – Up to 3x better detection rate, 4x less false alarms, 2x better classification then traditional systems,
- Achieves both high accuracy rate and low false alarm rate - No tradeoffs between high accuracy rate and low false alarm rate as traditional systems,
- Adapts to environmental conditions through project specific fine tuning,
- Works in tunnels, freeways and arterials,
- Easily expandible with Complete CCTV package: surveillance, real-time automatic incident detection and traffic statistics - incident management and decision support system,
- Proven in detailed tests by respectable authorities worldwide.
X-AID™ Traffic Video Analysis - Automatic Video Incident Detection (AVID) system implements Artificial Intelligence and Machine Learning algorithms to achieve less false alarms ratio and better detection accuracy than traditional non-AI video incident detection systems.
CCTV systems are the most widely used traffic monitoring systems in the world. And yet, the analysis of hundreds of simultaneous streams is still being done by operators. The systems for video analytics that have been used in the past are considered as not reliable due to their high false alarm rate and rather low detection accuracy.
X-AID™ uses deep learning technologies to perform incident detection and classification in real traffic environment. Verified in comparative tests, in realistic conditions (all weather conditions, bad illumination, poor video streams ...), it meets needs and expectations of Traffic Agencies on systems for video analytics.
Area of Use:
- Traffic Irregularities Detection,
- Traffic Flow Monitoring.
- Wrong-way Driving,
- Stopped Vehicle,
- Traffic Slowdown/Congestion,
- Slow vehicle,
- Reduced or Loss of Visibility (smoke, fog, etc.),
- Debris on the Road,
- Hard shoulder driving,
- Truck /Slow Vehicle in High Speed Lane,
- Trip wire detection.
Not affected by common problems:
- Camera Vibrations,
- Weather Conditions,
- Variations in Illumination,
- Low Video Quality.
Traffic Video Analytics:
- Average Vehicle Speed,
- Traffic Volume,
- Vehicle Classification.
- Digital Video Management System and Digital Video Recorder,
- System Configuration Tool – calibration, detection zones, alarm thresholds and configuration.
1. Solved Problems of Traditional AVID
Working in direct light on the exit of the tunnel
Automatic Video Incident Detection is the fastest way to detect a traffic incident and, as such, exponentially increases awareness about traffic conditions, and traffic safety. Video detection systems require ideal working conditions (camera placement, constant illumination, environmental impacts, moving traffic), to reach ideal performance. In real life operation, ideal working conditions are not met. Traditional AVID systems are typically far from performance rates stated in their brochures (detection rate is typically <50% and false alarm rate is typically >5/camera/day). The main advantage of XAID system, comparing to traditional systems, is ability of learning (fine tuning) which is based on Artificial Intelligence and Machine Learning algorithms. Learning ability of XAID enables it to learn how to perform in real-life conditions and tune its performance as close as possible to theoretical. Traditional systems do not have learning capability. Technical specifications of video analytics systems describe operation and performance of the system in theoretical (ideal) conditions. In real applications, cameras are usually not ideally positioned, illumination is not ideal and environmental influences are expected (shadows, obstacles, etc). Because of these reasons, ideal performance of the system most likely cannot not be reached. Yet, X-AID™ will reach much better performance than traditional systems, due to its fine tuning ability.
2. Connection Between AVID and Safety Management
With hundreds of cameras deployed throughout highways and tunnels delivering the immense amount of information simultaneously, it is impossible, even for a group of human operators, to monitor and evaluate the information effectively or efficiently. Reliable Video Automatic Incident Detection system is a must in such environments because it signiﬁcantly improves the efﬁciency of trafﬁc management systems. It automatically reports dangerous situations and irregular trafﬁc conditions, such as stopped vehicle, slow vehicle, driving in opposite direction, pedestrians in trafﬁc lanes, dropped cargo, ﬁre/smoke in a tunnel, or trafﬁc congestion.
Main constrains in Video Automatic Incident Detection industry is never ending trade-off between the need to have the least possible number of false alarms and still not to miss some of alarms that did occur on the road. Today’s traditional AID systems have come to dead end with technology currently being used.
X-AID™, the new generation of AID has done a big step further and solves critical issues traditional AID systems can’t solve - unique and the most advanced algorithms based on cognitive recognition open path to the new reliability standard in video detection industry. Machine learning setup process enables fine tuning for the best fit to the particular scene and environmental condition, and experience-based improvement in reliability and accuracy.
3. Complete Platform for Video Surveillance and Video Analytics
XAID combined with particular topXview™ modules provides complete platform for video surveillance and video analytics
- Complete CCTV package: surveillance, real-time automatic incident detection and traffic statistics,
- Out-of-the-box integral solution eliminates integration challenges between separate video Automatic Incident Detection (AID) systems and video management systems,
- State of the art Video Management System (video layouts management, video players, video wall management, video recording and reviewing, PTZ controls, support for latest video standards),
- Industry’s proven and most precise Video Analytics System currently available on the market:
- Incident detection and vehicle classification / traffic statistics,
- Artificial Intelligence algorithms and Deep machine learning – provides highest detection rate and lowest false alarm rate without trade-offs between them (as in traditional systems),
- Expertise and transparency from the beginning (realistic execution schedule, realistic expected performance),
- Customizable alarm management,
- Automatic incident recording,
- Detection metadata natively overlayed over live and recorded video,
- Intelligent filters based on detection metadata for video summarization,
- Easy to interface with Traffic Management Systems, SCADA systems or other incident management systems - Restful API for exchange of incident data, NTCIP C2C, Advanced proprietary TCP interface,
- Optional incident management and decision support system.