Video as a Sensor

Anomaly Detection | Intruder Alerts | Worker Safety | Emergency Response | Compliance & Regulatory

Video as a Sensor

Objective Utilize artificial intelligence to turn a live video feed into a smart sensor and alerting system that can help operations teams be safer and better informed.

Technologies “Qualified” video cameras and analytical data modeling, AI live video feed with smart sensor & ML alerting system, & "incident" notification

Partners Intel, Hewlett Packard Enterprise, CBT

 

                cbt

Cameras are focused on key assets and/or areas of the facility to identify when unusual activity occurs and notify authorized personnel. Such activities may include unauthorized personnel, hazardous wildlife, an accident involving workers, man down, rail cars slipping off their blocks or moving unintentionally, etc. When an anomaly occurs, plant operators are notified and provided with short video clips that capture the moments directly before, during and after the event. This enables the operator to dispatch the appropriate resolution quickly and effectively. 

Video as a Sensor enables notification of suspect activity on plant property including trespassing, theft and unwelcome/hazardous wildlife.

Video as a Sensor offers next generation digital analysis in which images are run through an AI inference engine to determine when unusual objects or actions occur. Objects and their characteristics (size and speed for example) are verified and compared to the pre-set examples or templates, trained via AI. When there is a match between object and model, then the frame and the objects are marked digitally. The operator can recall the digital marked images and accompanying video footage for further analysis. Determining if those objects and actions are ‘normal’ or ‘abnormal’ thereby enhances worker and plant safety.

 

Use Cases

Anomaly Detection | Intruder Alerts | Worker Safety | Emergency Response | Compliance & Regulatory