Home EN

CVCounter Wiki

Welcome to the documentation for the CVCounter project!

Project Description

CVCounter is a mini-website designed to track the number of objects in a video through a web interface. It serves as a simple and convenient tool for monitoring object flows (e.g., entry and exit or conveyor systems).

This solution does not require additional client-side software and works on any device with a browser. It offers three viewing modes: the main view with video, a text-only view, and a text view with dual counters. The project is intended to be lightweight and resource-efficient, especially when the video mode is not used.

Key Features:

• Web interface for displaying object counts
• Video view with recognition results
• Text-only modes for simplicity and resource-saving
• Dual counter display for monitoring entry and exit simultaneously
• Viewing reports on calculations
• Viewing system information

---

Installation

Method 1: Manual Installation

1. Clone the repository:

   git clone https://github.com/BespredeL/CVCounter.git

2. Navigate to the project directory:

   cd CVCounter

3. Install virtual environment:

   python3 -m venv venv

4. Activate virtual environment:
   • On Windows:

     .\venv\Scripts\activate

   • On Linux/Mac:

     source venv/bin/activate

5. Install dependencies:

   pip3 install -r requirements.txt

6. Rename the configuration file:

   mv config.example.json config.json

7. Change the parameters in the configuration file, add your YOLO model.
8. Run the application:

   python app.py

Method 2: Docker Installation

1. Clone the repository:

   git clone https://github.com/BespredeL/CVCounter.git

2. Navigate to the project directory:

   cd CVCounter

3. Build and run using Docker Compose:

   docker-compose up --build

---

Configuration

All primary configurations are stored in the config.json file.

General Parameters general

Parameter	Description	Default Value
debug	Enable debug mode	true
log_path	Path to the log file	errors.log
log_level	Minimum log level	INFO
log_console	Enable logging to console	false
default_language	Default language	ru
allow_unsafe_werkzeug	Allow unsafe operations in Werkzeug	false
button_change_theme	Show theme change button	true
button_fullscreen	Show fullscreen button	true
button_backward	Show back button	false
button_save_capture	Show save capture button	false
collapsed_keyboard	Show collapsed keyboard	true

Server Parameters server

Parameter	Description	Default Value
host	Server address	0.0.0.0
port	Server port	8080
threaded	Enable multithreading	true
use_reloader	Enable reload mode	false
log_output	Enable log output	true
socketio_key	SocketIO key	""
allowed_origins	Allowed address for Access-Control-Allow-Origin	*

User Parameters users

Parameter	Description	Default Value
admin	Login and password	admin:admin

Database db

Parameter	Description	Default Value
uri	Database connection URI	sqlite:///system/database.db
prefix	Table prefix	""

Form Parameters form

Parameter	Description	Default Value
defect_show	Show defect form	true
correct_show	Show correction form	true
custom_fields	Configuration for custom fields	[{"name": "field_one", "label": "Field One", "type": "text"}]

Default Detection Configuration detection_default

Parameter	Description	Default Value
model_type	Model type	yolo
weights_path	Model path	yolo_cfg/models/yolov8n.pt
video_show_scale	Video display scale on the page	50
video_show_quality	Video display quality on the page	50
video_fps	Manual FPS setting (0 - automatic)	0
confidence	Confidence threshold	0.7
iou	IOU threshold	0.7
device	Compute device	0
vid_stride	Video stream step	1
indicator_size	Indicator size	10
counting_area	Counting area (polygon)	[[0, 0], [100, 0], [100, 100], [0, 100]]
counting_area_color	Counting area color	[67, 211, 255]
classes	Detection classes	{}

Default video recording configuration detection_default.recording

Parameter	Description	Default Value
enable	Enable recording	false
path	Video saving path	yolo_cfg/saved_recordings
scale	Video size (in percent)	100
quality	Video quality	80

Detection Configurations detections.ExampleCam

ExampleCam

Parameter	Description	Default Value
label	Counter label	Label ExampleCam
start_total_count	Initial counter value	0
video_path	Path to video file or camera source	""
video_show_scale	Video display scale on the page	70
video_show_quality	Video display quality on the page	30
video_fps	Manual FPS setting (0 - automatic)	0
model_type	Model type	yolo
weights_path	Model path	yolo_cfg/models/yolov8n.pt
confidence	Confidence threshold	0.7
iou	IOU threshold	0.7
device	Compute device	0
vid_stride	Video stream step	1
indicator_size	Indicator size	10
counting_area	Counting area (polygon)	[[0, 0], [100, 0], [100, 100], [0, 100]]
counting_area_color	Counting area color	[255, 64, 0]
classes	Detection classes	{}
dataset_create.enable	Enable dataset creation	true
dataset_create.probability	Probability of image saving (0.01–1)	0.05
dataset_create.path	Path for saving dataset images	yolo_cfg/saved_images/ExampleCam
recording.enable	Turn on recording	false
recording.path	Video save path	yolo_cfg/saved_recordings
recording.scale	Video size (in percent)	100
recording.quality	Video quality	80

---

Interface

Main View:

This page displays the video feed along with counters showing the number of detected objects. It is the primary interface for real-time monitoring with recognition results.

URL:

http://127.0.0.1/counter/{location}

or

http://127.0.0.1/counter/{location}/video

Text View:

This page shows only the counter values without video. It is suitable for low-resource devices or when a simple display of information is required.

URL:

http://127.0.0.1/counter/{location}/text

Text View with Two Counters:

This view displays two counters simultaneously, which is helpful for monitoring both entry and exit flows.

URL:

http://127.0.0.1/counter_dual/text/{location_first}/{location_second}

---

Contributing to the Project

Contributions to the project are always welcome! To make changes, follow these steps:

• Fork this repository.
• Create a new branch.

  git checkout -b feature/your-feature

• Make your changes and commit them.

  git commit -m "Added a new feature"

• Push your changes.

  git push origin feature/your-feature

• Submit a PR (Pull Request) for review.

Before submitting, ensure your changes do not disrupt existing functionality.

---

Frequently Asked Questions (FAQ)

1. What are the minimum system requirements for the server?

• Processor: Modern 4-core processor (e.g., Intel Core i5 or AMD Ryzen 5).
• RAM: At least 8 GB (16 GB or more is recommended for stable video stream processing).
• Storage: SSD for storing datasets and logs.
• GPU: A GPU significantly speeds up processing and reduces system load. Minimum GPU requirements:
  • NVIDIA GTX 1050 (2 GB VRAM): Suitable for low-frame-rate image processing.
  • NVIDIA GTX 1660 (6 GB VRAM): Recommended for real-time video streams and high-resolution videos (up to 720p).
  • NVIDIA RTX 2060 или выше (6 GB+ VRAM): Stable performance for YOLO models in real-time at resolutions of 1080p and above.

Note: YOLO supports computation on NVIDIA GPUs using CUDA. GPUs from other manufacturers (e.g., AMD) may work but require additional configuration and may have lower performance.

2. What are the minimum system requirements for the client?

Any device with a web browser capable of running JavaScript.

3. How do I add a new type of counter?

The core functionality is already implemented, but you can extend the logic by modifying the data processing in the corresponding routes (views).

4. What should I do if the video does not display?

Check the camera’s functionality and ensure it is correctly connected and streaming in the web interface.

5. How do I prevent the counting interface from being closed?

You can run the browser in kiosk mode to prevent users from exiting it. For example, with Google Chrome, you can use the "--kiosk --start-fullscreen" options)

---

License

The project is distributed under the AGPL-3.0 license. Details can be found in the LICENSE file.