-
23
Projects submitted
-
10
Published projects
-
25
days until applications close
AI-based OTDR Testing Automation
-
Project manager on customer side
Sanjay Kumar Behera
HFCL
Head IT
-
Category
-
Nomination
-
Goals
To design and implement an AI-powered solution that integrates with OTDR testing systems.
The system will automatically identify and classify fiber, tube, and cable colors across diverse Optical Fiber Cable (OFC) types, improving accuracy, speed, and reliability in OTDR testing of various stages of operations.
-
Timeline
october, 2025 — august, 2026 -
Project completion year
2026
-
Project scale
64 automated workplaces -
Results
The AI-based OTDR testing system is installed and running at our Hyderabad plant on a dedicated NVIDIA Jetson AI unit. It was commissioned and tested on site on 10–11 August 2026. Multi-Tube and IBR cables were tested end to end; Flat Ribbon testing is to be completed by our QA team.
Ribbon ID exact match 96.5%.
96.2% fiber color matching.
95.4% tube color matching.
-
Project uniqueness
· AI-powered color recognition model.
· Integrated OTDR + AI software module.
· Technician dashboard with visualization and reporting features.
· Training dataset and documentation.
· Pilot deployment and field validation results.
-
Software used
High resolution camera - alvium c2460 with 25mm lens
NVIDIA Jetson Orin NX
Standalone AI server (A VM with Ubuntu 24.04.5, 8 vCPU, 1TB storage, 32 GB RAM)
New workstations with >90 CRI LED lighting
Vimba X Module
Docker, Node.js, SQL server
-
Solution from the Global CIO catalog
The project does not use solutions from the Global CIO catalog
-
Project implementation complexity
Ring marking identification on the fibers.
Color and marking on OFC having 6912 fibers.
-
Description
AI Model Development
· Training computer vision models to recognize fiber/tube/cable colors under varied conditions.
· Incorporating global standards (e.g., TIA-598 color codes) and HFCL-specific variations.
Data Acquisition
· Collecting high-resolution images of fibers/tubes/cables for all types of OFCs manufactured at HFCL Plants across locations.
· Building a labeled dataset for supervised learning.
OTDR Integration
· Synchronizing color identification with OTDR event data (faults, splices, reflections).
· Enabling color-coded mapping of fiber paths in OTDR reports.
User Interface
· Developing a technician-friendly dashboard.
· Features: OTDR trace visualization, color-coded fiber mapping, automated reporting.
Reporting & Documentation
· Exportable reports with visual color identification.
-
Project geography
Our plants in 3 locations Hyderabad, Chennai and Goa.
-
Customer
HFCL
Only authorized users can comment.
We suggest you log in or register.