Harshit Makwana

Computer vision / Case study

VehiSense Pro

I developed a vehicle-monitoring system with fine-tuned YOLOv8, tracking and event-detection logic, connected to a FastAPI web application through WebSockets.

Harshit Makwana

Concept illustration of vehicle tracking and parking analysis, with detection outlines over a miniature roadway.
AI-generated concept illustration · not a project screenshot
My role
Computer vision models, event logic & web application
Timeline
2024–ongoing
Focus
Computer vision

01 / Context

From vehicle detections to events a monitoring interface can use.

A vehicle detector identifies objects, but monitoring also requires understanding movements, incidents and parking conditions. I built VehiSense Pro to connect real-time vehicle classification and tracking with event logic and a web interface.

02 / Ownership

What I built

  • Fine-tuned YOLOv8 for vehicle classification and real-time tracking.
  • Implemented computer-vision logic for vehicle-speed calculation and U-turn detection.
  • Added accident detection and hit-and-run scene capture.
  • Developed restricted-area parking detection and parking-occupancy analysis.
  • Gathered outputs from the processing workflow and delivered them to a web application through WebSocket APIs.
  • Built the HTML, CSS and JavaScript frontend and FastAPI backend endpoints.

03 / Engineering

Technical approach

Build monitoring on top of detections

I fine-tuned the detection model and worked on the temporal logic needed for vehicle tracking, speed calculation and U-turn detection. I extended the workflow with incident and parking features so the application could present more than isolated detections.

Connect events to the web experience

I brought the outputs together through FastAPI and WebSocket APIs, with a frontend written in HTML, CSS and JavaScript. This connected the vision-processing work to a web application that could display the resulting monitoring information.

04 / System view

Workflow at a glance

  1. Visual input

    Vehicle observations

  2. Detection & tracking

    Fine-tuned YOLOv8 and tracking logic

  3. Event analysis

    Movement, incidents and parking

  4. Web application

    FastAPI, WebSockets and browser UI

05 / Delivery

Outcomes

The project brings vehicle detection, movement analysis, incident capture and parking analysis into one web-connected workflow. Development is ongoing, with my work spanning model adaptation, event logic and browser delivery.

Technologies & methods

  • YOLOv8
  • Computer vision
  • FastAPI
  • WebSockets
  • HTML
  • CSS
  • JavaScript

Roadway Solutions

I developed road-inspection workflows using fine-tuned YOLO, ByteTrack, MiDaS and SAM 2, with OpenVINO experiments for inference performance.

Image Inpainting Using GANs

I developed a GAN-based image-inpainting model, trained it on artificially masked images and built a Streamlit interface for uploading images and viewing restorations.

Satellite Image Change Detection

Developed a system to detect spatial changes over time using satellite imagery and advanced image processing techniques. The solution combines deep learning models, clustering, and dimensionality reduction for environmental and urban insights.