Modular Multi-Probe Videoscope Platform with AI-Assisted Defect Recognition

Product Overview

The Sugia OmniScope F4 is the capstone of the OmniScope product line — a modular, AI-assisted videoscope platform that unifies the complete Sugia probe ecosystem under a single console architecture and adds automated defect recognition, intelligent report generation, and enterprise asset-management integration layers on top of the best-in-class imaging and measurement hardware shared with the F1, F2, and F3.

The F4’s AI-Assisted Defect Recognition (AIDR) engine runs on a dedicated neural processing unit (NPU) within the console and has been trained on a curated dataset of over 250,000 NDT-inspection images annotated by Level III inspectors across aerospace, power generation, and oil & gas domains. The engine operates in “assist” mode rather than “replace” mode: it highlights potential anomalies with bounding boxes and confidence scores, queues them for operator review, and records the operator’s confirmation or rejection of each detection as feedback for future model updates (under an optional fleet-learning program). The operator retains full authority over all inspection accept/reject decisions.

The automated report generator reduces post-inspection documentation time from hours to minutes. It compiles inspection metadata (asset ID, operator, date, probe used), all operator-confirmed findings with images and measurements, and a configurable severity-classification table into an ASME/API/EN-compliant PDF report with embedded images at native resolution. The report engine supports custom templates configured per asset type, operator, or regulatory jurisdiction, and can automatically email or upload completed reports to configured destinations.

Key Engineering Features

• Universal Probe Compatibility: Accepts all OmniScope probe modules: 4.0 mm (A1 type), 6.0 mm (M2/A2 type), 6.1 mm HD (A2 HD type), 6.2 mm 3D (F1 type), and 8.4 mm (A4 type). Probe recognition and configuration in < 30 seconds.
• AI-Assisted Defect Recognition (AIDR): Dedicated NPU (24 TOPS INT8); trained on 250,000+ Level III-annotated NDT images; detects corrosion, cracking, coating loss, FOD, erosion, and pitting with < 5-second inference latency per frame; operator-in-the-loop with confirmation/rejection and feedback collection.
• Automated Report Generation: Configurable templates per asset type and regulatory jurisdiction; auto-populates all inspection metadata; embeds confirmed-finding images at native resolution with measurement overlays; outputs ASME/API/EN-compliant PDF reports; auto-email/upload to configured destinations.
• Enterprise Integration Suite: REST API with OpenAPI 3.0 specification; native connectors for SAP PM, IBM Maximo, and Infor EAM; inspection-data export in SQLite, CSV, JSON, and DICONDE (ASTM E2339) formats for CMMS/digital-twin ingestion.
• Fleet-Level Program Management: Console manages up to 5000 asset records with inspection history, next-due-date tracking, and configurable inspection-plan templates per asset type; trending dashboard for fleet-condition visualization across multiple assets, sites, and inspection campaigns.

Technical Specifications

Probe Compatibility Matrix:

Probe Type	Diameter	Lengths	Articulation	Resolution
Micro	4.0 mm	2.0 / 3.5 m	4-way 130°	1.0 MP
Standard	6.0 mm	2.0–7.5 m	4-way 180°	1.0 MP
HD	6.1 mm	3.5 / 5.0 m	4-way 160°	1.2 MP
3D Phase	6.2 mm	5.0 m	4-way 180°	1.2 MP + 3D
Large-Diameter	8.4 mm	3.5–7.5 m	4-way 160°	2.0 MP

Console Specifications:

Parameter	Value
Display	15.6-inch 4K UHD IPS, 3840 × 2160, 500 cd/m², touchscreen
CPU	Intel Core i7 (14-core)
GPU	NVIDIA RTX A2000 (8 GB GDDR6)
NPU (AIDR)	24 TOPS INT8 dedicated inference accelerator
RAM	32 GB DDR5
Storage	2 TB NVMe SSD (user-accessible) + 512 GB system SSD
AIDR Training Dataset	250,000+ annotated NDT images
AIDR Inference Latency	< 5 seconds per frame
AIDR Detection Classes	Corrosion, cracking, coating loss, FOD, erosion, pitting
Asset Database Capacity	5,000 assets with full inspection history
Report Engine	Configurable templates, ASME/API/EN compliant
DICONDE Export	ASTM E2339 compliant
Enterprise Connectors	SAP PM, IBM Maximo, Infor EAM (native)
Network	10 Gigabit Ethernet (SFP+) + Wi-Fi 6E + Bluetooth 5.3
I/O	4× USB 3.2 Gen2, 2× Thunderbolt 4, HDMI 2.1, DP 2.0
Battery	2× 7.4 V Li-ion, 95 Wh each, hot-swappable
Runtime	~2 h (full AI + 3D measurement active)
Ingress Protection	Console IP65 (closed), probe IP67
Operating Temperature	-20 °C to +50 °C
System Weight	8.5 kg (console + dual batteries, without probe)

Target Applications

• Fleet-Level Aerospace Engine Management: Centralized borescope-inspection program management for airline/operator fleets of 50–500+ engines; automated trend analysis across serial-numbered components for predictive-removal planning; AI-assisted anomaly detection for consistent inspection quality across multiple inspectors and sites.
• Refinery Turnaround Inspection Management: Multi-asset, multi-inspector inspection campaign coordination during plant turnarounds; automated report generation for hundreds of pressure vessels, piping circuits, and fired heaters in a single inspection window; direct CMMS integration for finding-to-work-order workflow.
• Power-Generation Fleet Management: Centralized condition monitoring across multiple generating units and sites; trending of repeat degradation modes for fleet-wide corrective-action planning; automated compliance-documentation generation for statutory inspection requirements.
• Defense & Government Fleet Support: Standardized inspection-data format across multiple platform types (fixed-wing, rotary-wing, UAV, marine, ground vehicles); DICONDE-compliant data archive for long-term technical-record preservation; AI model training on platform-specific defect populations for fleet-unique anomaly detection.