Computer Vision · YOLO
Telegram bot that counts free parking spots from photos
Accepts a parking photo, runs YOLOv8 on CPU, returns an image with detected cars and a free spot estimate.
Timeline~18 days
TypeTelegram bot + Computer Vision
StackPythonYOLOv8OpenCVAiogram 3
Context
A request for a full parking occupancy tracking service with camera connection, occupancy charts and analytics.
Task
There was no automated way to count occupancy. Manual counting was irregular and inaccurate. Needed to work on limited CPU resources.
Solution
- Telegram bot accepts a photo or camera frame
- Analyzes via YOLOv8, optimized to work on CPU without GPU
- Returns an image with detected vehicle markup
- Tracks an occupancy chart over time
Result
The project focused on a practical business task: removing manual routine, reducing errors and giving the team a clear tool for everyday work. The main effect is reflected in the metric: ~95% — vehicle detection accuracy on CPU without GPU.
Project gallery
Click an image to open a larger view
Result
KPI~95%
vehicle detection accuracy on CPU without GPU
Free audit
Want something similar? Free audit
Tell us about your task — we'll figure out what and how to automate in one call.
No advance payment until contract
Estimate in 24 hours
Working since 2019