Indoor facility pathways at an Oil & Gas plant are high-traffic zones used continuously across all shifts. Any foreign object left on a pathway - tools, equipment, debris, or misplaced items - creates an immediate trip and fall hazard for workers moving through the facility. In an Oil & Gas environment, a trip incident near active machinery or pipelines carries consequences far beyond a standard workplace injury.
Monitoring pathways manually was not viable at scale. Periodic checks left significant windows where an object could remain on a pathway undetected across an entire shift. A proactive, continuous detection system was needed that could identify any foreign object the moment it appeared and prompt immediate removal without requiring manual supervision.
Problem Statement
THE CHALLENGE
Challenges
-
No automated object detection on pathways
Any foreign object placed or dropped on a pathway remained undetected until a manual check or worker encounter.
-
Periodic checks insufficient for 24/7 operations
High-traffic pathways required continuous monitoring - not interval-based supervision.
-
Standard object detection not suitable
Generic object detection models could not reliably distinguish foreign objects from normal pathway fixtures and equipment backgrounds.
-
Immediate removal required
Object presence on a live pathway needed instant response - not a logged report for later action.
Solution
THE SOLUTION
We deployed a background differencing-based object detection system across indoor facility pathway cameras. Rather than using standard object classification, the system establishes a clean baseline of each pathway and flags any object that appears as a deviation from that baseline - making it effective for detecting any foreign object regardless of type, size, or shape. On detection, an automated speaker announcement is triggered prompting immediate removal by nearby personnel.
Flow: Live pathway camera feed -> Background differencing detects foreign object -> Object presence confirmed -> Speaker announcement triggered -> Personnel alerted to remove object -> Pathway cleared -> Incident logged on dashboard
-
Background differencing approach used instead of standard object detection - any deviation from pathway baseline flagged immediately regardless of object type.
-
Existing indoor pathway cameras integrated with detection layers across all monitored corridors.
-
24/7 continuous monitoring across all pathways - no shift gaps or manual check dependency.
-
Automated speaker announcement triggered on detection - prompts immediate removal without requiring supervisor intervention.
-
Every detection is logged with image, location, and timestamp on the dashboard for safety audit records.
