Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/th-nuernberg/meteor-scatter
Detect and classify meteors using ML methods
https://github.com/th-nuernberg/meteor-scatter
machine-learning meteor scatter
Last synced: 9 days ago
JSON representation
Detect and classify meteors using ML methods
- Host: GitHub
- URL: https://github.com/th-nuernberg/meteor-scatter
- Owner: th-nuernberg
- License: mit
- Created: 2025-01-14T13:32:06.000Z (10 days ago)
- Default Branch: main
- Last Pushed: 2025-01-14T14:44:44.000Z (10 days ago)
- Last Synced: 2025-01-14T15:43:36.932Z (10 days ago)
- Topics: machine-learning, meteor, scatter
- Language: Python
- Homepage:
- Size: 4.76 MB
- Stars: 0
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Meteor Scatter
This project was developed as part of a student project. The aim was to detect and classify meteors using ML methods.
The project consists of two parts: the **Meteor Detection & Classification** and the **Flask Server**. The meteor detection & classification part is responsible for saving the detections in a CSV file. The Flask Server application reads the CSV file and displays the data in a web interface.
## Authors
- [Robert Heinrich](mailto:[email protected])
- [Sebastian Kolb](mailto:[email protected])
- [Maximilian Bundscherer](mailto:[email protected])
- [Prof. Dr. Thomas Lauterbach](mailto:[email protected])
## Meteor Detection & Classification
See other README file and subset in seperate folder [meteor_detect_class/](meteor_detect_class/)
## Flask Server
### Features
The Flask Server will process always the last full 30 days of csv-files and create multiple graphics. Build in Slideshow with
- daily display with hourly resolution of all Meteors and over critical ones
- 7-days-overfew with the sum of each day of all Meteors and over critical ones
- 30-days-overfew with the sum of each day of all Meteors and over critical ones
- a separate gauge-chart displaying the average hourly rate of the last day
With every startup of `app.py` the work-data-csv-file `final_dataframe.csv` will be verified and checked to its current status. Should there be missing data sets, it will be displayed at the bottom of the website Missing hours in single days will be skipped and not displayed in the charts.
Furthermore the `final_dataframe.csv` will be validated periotically (setup in `config.ini`: `#Scheduler Aktualisierungszeit CSV-Datensatz in minutes - schedule_interval = 2`) the validation process will search for the last entry and complare to the date of yesterday, no action if it matches, update file if there is a missmatch on the frontend the charts will be reloaded with a certain interval to be sure all displays are always up to date. (setup in `config.ini`: `#Aktualisierungszeit Website reload - reload_interval = 150000 - #60000 1min / 300000 5min`)
### Data Requirements
All data must be in the following structure to ensure 100% functionality:
`name: "YYYYMMDD.csv" in folder "csv_files"`
```
content format:
Timestamp;Anzahl;Kritisch
2024-11-30 00:05:00;101;0
2024-11-30 01:05:00;8;3
2024-11-30 02:05:00;99;21
...
```
### Structure
```
-csv_files
-20241210.csv
-20241211.csv
-20241212.csv
...
-static
...
-templates
...
-app.py
-config.py
-config.ini
-database.py
-initapp.py
-plot.py
-(final_dataframe.csv)
```
### Description
`app.py`: (run for main App)
The main App with all Flask routes. The IP adress and port can be set in the following line: `Line 174: app.run(host="0.0.0.0", port=5000, debug=debug_value)`
`config.py`:
Settings and fallback values as well as `config_read` and `config_get` functions.
`config.ini`:
Manipulate settings here if you are not a programmer and just the user to make it more like you need it. Rerun `app.py` and reload Website after changes.
`database.py`:
Processes all data provided by the Grabber and save in `final_dataframe.csv`.
`initapp.py`:
Initialise Flask App before start.
`plot.py`:
Creates all charts. If needed resolution of the charts can be manipulated here. Needs to be changed for each charts seperatly. (300dpi and figsize 10,6 will create a figure with resolution of 3000x1800px, can be reduced if less performace is needed).
### Install
Create a virtual environment, activate it, and install the requirements (`requirements.txt`).
### Run
Activate the virtual environment and run `app.py`.
### Debug
In `config.py` change following line, to get more Debug/Logging printout:
```
logging.XXXXXXXXXX ERROR/WARNING/INFO options
console = logging.StreamHandler()
console.setLevel(logging.ERROR) # Konsole für nur `INFO`-Level oder höher aktivieren
formatter = logging.Formatter("%(asctime)s - %(levelname)s - %(message)s")
console.setFormatter(formatter)
logging.getLogger("").addHandler(console)
```