How I Trained a YOLO Model to Read Bus Stickers in Real Time
A deep dive into the dataset collection, manual annotation process, and deployment challenges of our campus transport monitoring.
Computer Vision
Robotics & Automation Enthusiast
First-year EEE student at SRM TRP. Building autonomous systems to make humans supervisors, not workers.
I'm Hariharan V, a first-year Bachelor of Engineering student in Electrical and Electronics Engineering at SRM TRP Engineering College, Trichy, Tamil Nadu. My obsession lies at the intersection of robotics, automation, AI, and computer vision.
My mission is simple: automate the repetitive so humans can focus on what truly matters — thinking, creating, and leading. Whether it's software pipelines or physical hardware, I'm always looking for what can be made smarter and autonomous.
I've already shipped two real-world projects used on my campus, trained custom YOLO models, and built AI-powered institutional systems. This is only the beginning.
→ Currently exploring advanced perception systems for autonomous robots
A smart campus transport system connected to college gate CCTVs that continuously monitors every vehicle crossing the gate. The system detects college bus stickers using a custom-trained YOLOv8/v10 model trained on 1,000+ manually annotated images by our team. It reads bus route numbers via OCR, checks against a live database, and logs entry and exit timestamps. If a bus arrives late, an automated SMS is instantly sent to the manager with the bus route details. Tested and deployed live on campus — built to make our campus smarter.
An AI-powered secure institutional digital approval workflow system designed to replace traditional paper-based approval processes in colleges. Faculty register using official email, employee ID, department, designation, and approval roles. The system dynamically shows only authorized roles — Class Coordinator, Mentor, HOD, or Principal. After password-verified signing, the digital signature and date are auto-inserted into the document. TRP Digital Sign then detects the next approver, sends smart professional emails using Gemini AI, generates QR-based document verification, exports clean PDFs, and securely stores final documents in structured Google Drive folders. Published live on the Google Workspace Marketplace.
Designing autonomous systems and hardware-software integration for real-world automation challenges.
YOLOv8/v10 training, OpenCV pipelines, EasyOCR, object detection on live CCTV streams.
Core programming for ML models, data processing, hardware interfacing, and backend logic.
HTML, CSS, JS, Google Apps Script — building dashboards, workflow tools, and campus-grade applications.
Integrating Gemini AI for smart automation, email generation, document processing, and intelligent workflows.
Google Drive API, Sheets, Gmail API, Pandas, CSV pipelines, and structured cloud storage systems.
Electronics fundamentals, circuit design, sensors, and microcontroller integration for automation.
Google Workspace Marketplace publishing, QR-based verification, live campus deployments.
SRM TRP Engineering College, Tamil Nadu
First-year undergraduate with a focus on robotics, automation, and AI integration in hardware systems. Actively building and deploying real campus projects while exploring computer vision and intelligent systems.
Campus Innovation Project
Designed, trained, and deployed a full AI-powered vehicle monitoring system live at the college gate — handling real-time CCTV feeds, custom YOLO models, and automated SMS alerts.
Independent Project
Built and published an AI-powered institutional document signing platform used by faculty and administrators, integrating Gemini AI, Google APIs, and role-based authentication.
IBM SkillsBuild Certificate — IBM
Published TRP Digital Sign — Google
A deep dive into the dataset collection, manual annotation process, and deployment challenges of our campus transport monitoring.
Computer VisionMy philosophy on robotics and automation — and why making humans supervisors instead of workers is the most empowering goal.
OpinionThe story of automating our college's approval workflow — the problems, the tech decisions, and what it taught me about real-world deployment.
Project StoryWhat I learned in my first year of engineering, why I pivoted hard into AI and robotics, and what I'm building next.
PersonalHave a project idea, collaboration, or just want to talk robots and automation? I'm always open to conversations that push things forward.