ADIPEC · 2 Nov 2026
Edge AI Vision Camera Intelligence and Autonomous Mobile Robots (AMRs) for Real-Time Industrial Safety Monitoring in Oil & Gas Facilities
A. B. V. Dilip
Citations by database
Each database indexes a different set of journals, preprints and books, so they are shown side by side rather than blended.
- OpenCitations0
- Crossref0
Citations per year
The current year is still counting.
- 2026 so far0
Abstract
Abstract Objectives/Scope Oil & gas operations require continuous monitoring of safety compliance and high risk anomalies; however, conventional supervision and manual CCTV monitoring are constrained by limited coverage, delayed detection, and inconsistent enforcement. This paper presents an AI-based industrial safety monitoring framework for edge computer vision analytics and Autonomous Mobile Robots (AMRs) to enable real-time detection of safety violations and hazardous conditions. Methods, Procedures, Process The proposed solution integrates deep learning-based video analytics with existing CCTV infrastructure and AMR-mounted cameras to extend monitoring into blind spots and dynamic work zones. AI inference is performed on edge GPU servers to ensure low latency, reduce cloud dependency, and support uninterrupted site operations. The platform detects over 100 industrial safety anomaly types, including PPE non-compliance, restricted zone intrusion, work-at-height violations, fire/smoke indicators, spill/leak anomalies, housekeeping deviations. Temporal tracking and Multi-frame validation is applied to improve event confidence and reduce transient false triggers. Rule-based escalation workflows generate real-time alerts with timestamped evidence frames, which are stored in an auditable event database for compliance reporting, trend analytics, and continuous improvement. Results, Observations, Conclusions Performance evaluation across operational deployments indicates a detection accuracy of 90-95% depending on use case complexity and environmental conditions. End-to-end alert latency from detection to notification was measured at approximately 3 seconds, enabling near real-time intervention by EHS teams. The combined CCTV + AMR monitoring approach improved coverage in areas where fixed cameras had occlusion limitations, particularly in large fabrication yards, laydown zones, and high-traffic operational corridors. Observed deployment constraints included low-light sensitivity, PPE partial visibility and false triggers caused by reflections and steam. These were mitigated through site-specific confidence threshold tuning, improved zoning logic, and periodic model refinement using operational datasets. The system also enabled automated reporting outputs such as violation heatmaps, repeat violation identification, shift-wise compliance scoring, and incident categorization. The paper concludes that integrating edge AI vision with existing CCTV infrastructure and AMR-based mobile sensing provides a scalable and measurable approach to proactive industrial safety monitoring, improving response time, compliance governance, and data-driven EHS decision-making. Novel/Additive Information This paper presents additive field evidence on deploying a low-latency (~3 seconds) edge AI vision system integrated with existing CCTV networks and Autonomous Mobile Robots (AMRs) to monitor 100+ industrial safety anomalies. It contributes practical deployment methodology, performance benchmarking (90-95% accuracy), and mitigation strategies for real operating constraints such as occlusion, low-light, reflections, and steam. The work advances scalable, audit-ready AI safety intelligence for petroleum operations.
What the data shows
- No database records a citation yet, as is usual for a paper this new.
Graph
This paper in the middle, its authors and their most-cited papers below, the papers it cites to the left and the papers citing it to the right. The timeline lays every paper out by year.
Select a node for its details. Drag to pan, scroll to zoom, drag a node to move it.
Start with
Cited by 0 papers
Who cites this paper, from open citation data, with Crossref's record of each; the 100 newest when there are more.
References (4)
What it cites, from open citation data, with Crossref's record of each.