Skip to content

This program will help you not to forget to remove cups and glasses from the table. The mug detector was written using a pre-trained model from Pytorch. If there are more than a certain number of glasses, you will receive a notification in Telegram

Notifications You must be signed in to change notification settings

LongaBonga/CupsCapture

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CupsCapture

Introduction

This program will help you not to forget to remove cups and glasses from the table. The mug detector was written using a pre-trained model from Pytorch. If there are more than a certain number of glasses, you will receive a notification in Telegram

Installation

Prerequisites

  • Linux, macOS or Windows
  • Python 3.6+

Install CupCapture

Clone repo

git clone https://github.com/LongaBonga/CupsCapture.git

Install requirements

cd CupsCapture
pip install -r requirements.txt

Run program

there are also arguments for starting this project:

  • --start_time The hour when program will start searching cups
  • --finish_time The hour when program will finish searching cups
  • --video_path The path to the video. or digit, if it is web camera
  • --time_freq The frequency with which the program will search for cups and send notifications if they are found. Default 15 minutes
  • --frame_amount The number of frames that must be processed at a time to predict the number of cups
  • --show If it is True, the image from the camera will be displayed on the screen
  • --trashhold Images with a probability below this threshold will not be considered

Note: You also need your own telegram bot more precisely, you need his token. You can get it by writing a message to @BotFather in Telegram

Run example

python3 main.py --start_time=10 --finish_time=23 --video_path=0 --frame_amount=5

About

This program will help you not to forget to remove cups and glasses from the table. The mug detector was written using a pre-trained model from Pytorch. If there are more than a certain number of glasses, you will receive a notification in Telegram

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages