14 Commits

Author SHA1 Message Date
e137d4bdf3 Merge branch 'v2' of ssh://git.luc-geo.fr:8201/lumathieu/video2geoframes.py into v2 2024-06-26 01:44:44 +02:00
708c400c61 Fix release badge 2024-06-26 01:43:22 +02:00
75a167ad87 Delete french version link 2024-06-26 01:38:07 +02:00
9363029603 Spelling fix 2024-06-26 01:36:38 +02:00
bfa9f55561 Add EXTRA_LICENSES 2024-06-26 01:34:35 +02:00
466ada80ad Add global project description to README + french version 2024-06-26 01:17:35 +02:00
4a66a8a202 [v2.0-alpha9] Fixing relative paths support 2024-06-24 03:31:59 +02:00
67db172d62 [v2.0-alpha8] Multiple fixes + complete re-writing of script initial configuration + exceptions support enhancement + code cleanup 2024-06-24 03:18:27 +02:00
0040572ac8 Fixing mistakes in import. 2024-06-24 03:03:36 +02:00
9ef72f4567 [v2.0-alpha7] TUI enhancement 2024-06-24 03:00:43 +02:00
7d53803c7d [v2.0-alpha6] Adaptations on locales text 2024-06-24 02:50:23 +02:00
801ca04299 [v2.0-alpha6] Implementing resizing frame function + modifying max. default video height (16K) + code reordering 2024-06-24 02:47:03 +02:00
c64d303fe3 [v2.0-alpha5] Re-formatting displaying date in TUI + re-formatting unit in tqdm (cv2 process) + TUI enhancement + correction in locales 2024-06-24 02:32:27 +02:00
cff5d49833 [v2.0-alpha4] Adding script version in exif.software tag 2024-06-24 02:22:16 +02:00
8 changed files with 568 additions and 278 deletions

View File

@@ -1,5 +1,96 @@
# video2geoframes.py
Python script to generate a collection of geotagged images from a video and a GPS track.
![Gitea Release](https://img.shields.io/gitea/v/release/lumathieu/video2geoframes.py?gitea_url=https%3A%2F%2Fgit.luc-geo.fr&include_prereleases&sort=semver&display_name=release&style=flat&link=https%3A%2F%2Fgit.luc-geo.fr%2Flumathieu%2Fvideo2geoframes.py%2Freleases)
Designed for contribution to street-level imagery projects like Mapillary or Panoramax.
_🇬🇧 version_
Python program to generate a collection of geotagged images from a video and a GPS track.
Designed for ease contribution to street-level imagery projects like Mapillary or Panoramax.
## Quick start
Nothing simpler : collect your video, your GPS track, execute Python script and follow the guide !
In detail, the program is built around a TUI or _Textual User Interface_, permitting to launch video process easily with
step-by-step parameters input.
Input is guided by textual help indicating attempted values.
Before script starting, you need to have :
* a video file with exact timestamp (start) in local time or UTC
* a clean GPS tack file covering video duration
* a working directory.
## Documentation
_Coming soon._
## Features
_Coming soon._
### Comparison v1 / v2
| Features | v1-beta | v2-alpha9 |
|-----------------------------|------------|------------|
| Timelapse video support | ✔️ | ✔️ |
| EXIF tags writing | ✔️ | ✔️ |
| Extended tags support | ✔️ | ❌ |
| Milliseconds support | ✔️ | ✔️ |
| Progress displaying | 🟡 raw | ✔️ |
| Multilingual TUI 🇺🇳 | 🟡 limited | ✔️ |
| Configuration customization | ❌ | 🟡 partial |
| JPEG qualtiy customization | ❌ | 🔄 planned |
| TOML setting | ❌ | 🔄 planned |
## Languages
TUI is multilingual thanks to "locales" base in the form of TOML files (`locales/*.toml`) easily extensible.
| Languages | Locale | Support | Maintainer |
|--------------|---------|------------|--------------|
| 🇺🇸 English | `en_us` | ✔️ 100 % | @lumathieu |
| 🇫🇷 French | `fr_fr` | ✔️ 100 % | @lumathieu |
| 🇮🇹 Italian | `it_it` | 🔄 planned | @lumathieu ? |
## Versions
See [_Releases_](https://git.luc-geo.fr/lumathieu/video2geoframes.py/releases).
## Setup
To set up program, be enough to clone Git repository, set up software dependencies and build Python environnement.
Recommended to use a virtual environnement (venv).
### Python
Entire project is developed and tested on **Python 3.11** (Windows x86-64).
### Dépendances
Core script uses following Python libraries (see also `requirements.txt`) :
- `numpy`
- `opencv-python`
- `piexif`
- `tomlkit`
- `tqdm`.
## Compatibility
Code is designed to be platform-independent.
Official supported platforms are Windows and Linux (partially tested under Debian / Ubuntu).
## Contribution
_Coming soon._
If you are interested to project contribution, you can send a mail to campanu@luc-geo.fr.
## License
This repository, except dependencies, is licensed under **GNU AGPL v3**.
Dependencies are included in repository for development and keep their original license
(see `dependencies/EXTRA_LICENSES.md`).

113
README_fr.md Normal file
View File

@@ -0,0 +1,113 @@
# video2geoframes.py
![Gitea Release](https://img.shields.io/gitea/v/release/lumathieu/video2geoframes.py?gitea_url=https%3A%2F%2Fgit.luc-geo.fr&include_prereleases&sort=semver&display_name=release&style=flat&link=https%3A%2F%2Fgit.luc-geo.fr%2Flumathieu%2Fvideo2geoframes.py%2Freleases)
_version 🇫🇷_
Programme Python permettant de générer un ensemble d'images géotaguées depuis une vidéo et une trace GPS.
Conçu pour faciliter la contribution à des projets de photo-cartographie de rue tels que Mapillary ou Panoramax.
## Démarrage rapide
Rien de plus simple : rassemblez votre vidéo, votre trace GPS, lancez le script Python et suivez le guide !
En détail, le programme est entièrement construit autour d'une TUI ou _Textual User Interface_, qui permet de lancer
facilement le traitement de la vidéo par la saisie pas-à-pas des paramètres.
La saisie est guidée par une aide textuelle indiquant les valeurs attendues.
Avant de lancer le script, vous avez besoin d'avoir :
* un fichier vidéo avec son horodatage exact (début) en temps local ou UTC
* un fichier de trace GPS propre couvrant la durée de la vidéo
* un dossier de travail.
## Documentation
_A venir._
## Fonctionnalités
Le programme permet d'exécuter en un seul traitement les tâches suivantes :
* le séquençage de la vidéo selon un intervalle de temps
* l'horodatage incrémental de la séquence d'image
* l'export des images au format JPEG
* le géotaguage des images exportées à partir de la trace GPS.
Il inclut également :
* le support des vidéos timelapse
* le redimensionnement des images à une résolution inférieure à la vidéo d'origine tout en conservant le ratio
* l'ajout de métadonnées avec les tags EXIF `artist`, `make`, `model` et `copyright` (cf. [documentation ExifTool](https://exiftool.org/TagNames/EXIF.html))
* l'horodatage à la précision de la milliseconde
* le support du temps local décalé par rapport à UTC.
* l'ajout d'un décalage temporel pour mieux corréler la vidéo et la trace GPS.
Lors de l'export, un sous-dossier nommé selon la vidéo est créé automatiquement dans le répertoire de sortie.
### Comparaison v1 / v2
| Fonctionnalité | v1-beta | v2-alpha9 |
|--------------------------------------|------------|--------------|
| Support des vidéos timelapse | ✔️ | ✔️ |
| Écriture des tags EXIF | ✔️ | ✔️ |
| Support des tags étendus | ✔️ | ❌ |
| Support des millisecondes | ✔️ | ✔️ |
| Affichage de la progression | 🟡 brut | ✔️ |
| TUI multilingue 🇺🇳 | 🟡 limitée | ✔️ |
| Personnalisation de la configuration | ❌ | 🟡 partielle |
| Personnalisation qualité JPEG | ❌ | 🔄 planifié |
| Paramétrage via TOML | ❌ | 🔄 planifié |
## Langues
La TUI est multilingue grâce une base de "locales" sous forme de fichiers TOML (`locales/*.toml`) facilement extensible.
| Langue | Locale | Support | Mainteneur |
|---------------|---------|-------------|--------------|
| 🇺🇸 Anglais | `en_us` | ✔️ 100 % | @lumathieu |
| 🇫🇷 Français | `fr_fr` | ✔️ 100 % | @lumathieu |
| 🇮🇹 Italien | `it_it` | 🔄 planifié | @lumathieu ? |
## Versions
Voir [_Releases_](https://git.luc-geo.fr/lumathieu/video2geoframes.py/releases).
## Installation
Pour installer le programme, il suffit de cloner le dépôt Git, d'installer les dépendances logicielles et de construire
l'environnement Python. Il est recommandé d'utiliser un environnement virtuel (venv).
### Python
L'ensemble du projet est développé et testé avec **Python 3.11** (Windows x86-64).
### Dépendances
Le script principal utilise les librairies Python suivantes (voir aussi `requirements.txt`) :
- `numpy`
- `opencv-python`
- `piexif`
- `tomlkit`
- `tqdm`.
Il utilise également le programme [`ExifTool`](https://exiftool.org/) pour le géotaguage des images.
Appelée par une commande système, cette dépendance est prévue pour être supprimée dans les versions futures.
## Compatibilité
Le code est conçu pour être indépendant de la plateforme.
Les plateformes officiellement supportées sont Windows et Linux (partiellement testé sous Debian / Ubuntu).
## Contribution
_A venir._
Si vous intéressé pour contribuer au projet, vous pouvez envoyer un mail à campanu@luc-geo.fr.
## Licence
Ce dépôt, à l'exception des dépendances, est sous licence **GNU AGPL v3**.
Les dépendances sont incluses au dépôt pour le développement et restent sous leur licence d'origine
(voir `dependencies/EXTRA_LICENSES.md`).

8
dependencies/EXTRA_LICENSES.md vendored Normal file
View File

@@ -0,0 +1,8 @@
# Extra licenses
Detail of dependencies original license.
## ExifTool (by Phil Harvey)
> This is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
> http://dev.perl.org/licenses/

View File

@@ -1,7 +1,7 @@
# Localization file for video2geoframes.py script
# Localization file for video2geoframes.py
# English (US / World)
#
# Last edition : 2024-06-22
# Last edition : 2024-06-23
[ui]
@@ -12,6 +12,7 @@ This script is designed to create geotagged frames from video and GPX track."""
end = "End of program, press Enter to quit."
paths_title = "## Paths"
parameters_title = "## Process parameters"
tags_title = "## Additional tags"
metadata = """{} ({} {}B)\n
- Duration : {} s\n
- Start time : {}.{}\n
@@ -21,16 +22,16 @@ metadata = """{} ({} {}B)\n
cv2_tqdm = 'frame(s)'
[ui.parameters]
toml_setting² = "Setting with TOML file ({}/{}) ? "
toml_setting = "Setting with TOML file ({}/{}) ? "
timelapse = "Timelapse video ({}/{}) ? "
timelapse_fps = "Timelapse framerate (frame/s) [{}-{}] : "
timelapse_fps_err = "Error... please enter a decimal between {} et {}."
frame_samp = "Enter the frame sampling in seconds [{}-{}] : "
frame_samp = "Enter frame sampling in seconds [{}-{}] : "
frame_samp_err = "Error... please enter a decimal between {} and {}."
frame_height = "Enter frame height in pixels (ratio unchanged) [{}-{}] : "
frame_height = "Enter output frame height in pixels (ratio unchanged) [{}-{}] : "
frame_height_err = "Error... please enter an integer between {} and {}."
video_start_datetime = "Enter video start datetime following ISO format (exemple : 2023-09-18T22:00:02.000) : "
@@ -38,7 +39,7 @@ video_start_datetime_err = "Error... entered datetime is not valid."
rec_timezone = "Enter time offset related to UTC (example for CEST : +02:00) : "
time_offset = "Enter time offset between video and GPX in seconds [{}-{}] : "
time_offset = "Enter time offset with GPS track in seconds [{}-{}] : "
time_offset_err = "Error... please enter a decimal between {} and {}."
[ui.paths]

View File

@@ -1,7 +1,7 @@
# Localization file for video2geoframes.py script
# Localization file for video2geoframes.py
# French (France)
#
# Last edition : 2024-06-22
# Last edition : 2024-06-23
[ui]
@@ -11,6 +11,7 @@ intro = """Bienvenue dans le script video2geoframes.py !
Ce script permet, à partir d'une vidéo et d'une trace GPS, de créer un ensemble de photos géotaguées."""
end = "Fin du programme, appuyez sur Entrée pour fermer."
paths_title = "## Chemins"
tags_title = "## Tags additionnels"
parameters_title = "## Paramètres du traitement"
metadatas = """{} ({} {}B)
- Durée : {} s
@@ -27,10 +28,10 @@ timelapse = "Vidéo timelapse ({}/{}) ? "
timelapse_fps = "Débit d'image du timelapse (image/s) [{}-{}] : "
timelapse_fps_err = "Erreur... entrez un entier entre {} et {}."
frame_samp = "Entrez l'espacement temporel des images en secondes [{}-{}] : "
frame_samp = "Entrez l'espacement temporel en secondes entre les images [{}-{}] : "
frame_samp_err = "'Erreur... veuillez entrer un nombre décimal entre {} et {}."
frame_height = "Entrez la hauteur des images en pixels (ratio inchangé) [{}-{}] : "
frame_height = "Entrez la hauteur en pixels des images en sortie (ratio inchangé) [{}-{}] : "
frame_height_err = "Erreur... veuillez entrer un nombre entier entre {} et {}."
video_start_datetime = "Entrez l'horodatage du début de la vidéo au format ISO (exemple : 2023-09-18T22:00:02.000) : "
@@ -38,7 +39,7 @@ video_start_datetime_err = "Erreur... l'horodatage entré est invalide."
rec_timezone = "Entrez le décalage horaire par rapport à UTC (exemple pour CEST : +02:00) : "
time_offset = "Entrez le décalage temporel entre la vidéo et le GPX en secondes [{}-{}] : "
time_offset = "Entrez le décalage temporel avec la trace GPS en secondes [{}-{}] : "
time_offset_err = "Erreur... veuillez entrer un nombre décimal entre {} et {}."
[ui.paths]

6
requirements.txt Normal file
View File

@@ -0,0 +1,6 @@
colorama==0.4.6
numpy==2.0.0
opencv-python==4.10.0.84
piexif==1.1.3
tomlkit==0.12.5
tqdm==4.66.4

View File

@@ -2,13 +2,13 @@
# -*- coding: utf-8 -*-
"""video2geoframes.py
Python script to generate a collection of geotagged images from a video and a GPS track.
Python program to generate a collection of geotagged images from a video and a GPS track.
Designed for contribution to street-level imagery projects like Mapillary or Panoramax.
"""
__author__ = "Lucas MATHIEU (@campanu)"
__license__ = "AGPL-3.0-or-later"
__version__ = "2.0-alpha3"
__version__ = "2.0-alpha9"
__maintainer__ = "Lucas MATHIEU (@campanu)"
__email__ = "campanu@luc-geo.fr"
@@ -19,9 +19,9 @@ from datetime import datetime, timedelta
import cv2
import piexif
from tomlkit import dumps, loads
from tomlkit import loads
from tqdm import tqdm
from exif import Image, GpsAltitudeRef
# Functions
def unix_path(path):
@@ -45,103 +45,155 @@ def byte_multiple(size):
size = size / 1024
multiple = multiples[i]
return size, multiple
def existing_items(expected_items: list, items: list):
presents_items = []
duplicated_items = []
missing_items = []
for eit in expected_items:
i = 0
for it in items:
if it == eit:
i += 1
if i == 0:
missing_items.append(eit)
elif i == 1:
presents_items.append(eit)
else:
duplicated_items.append(eit)
return {'presents': presents_items, 'duplicated': duplicated_items, 'missing': missing_items}
def list_enumerator(item_list: list, intermediate_separator: str, last_separator: str):
i = 1
for it in item_list:
if i == 1:
enumerated_list = it
elif 1 < i < len(item_list):
enumerated_list = '{}{}{}'.format(enumerated_list, intermediate_separator, it)
else:
enumerated_list = '{}{}{}'.format(enumerated_list, last_separator, it)
i += 1
return enumerated_list
# Start
print('# video2geoframes.py')
print("# video2geoframes.py (v{})\n".format(__version__))
# Configuration settings
base_path = unix_path(os.path.dirname(__file__))
ini_file_path = '{}/video2geoframes.ini'.format(base_path)
ini_file_err = False
conf_file_path = '{}/video2geoframes_conf.toml'.format(base_path)
conf_file_err = False
## Default values
# Default values
mandatory_parameters = ['locale', 'exiftool_path']
locale = 'en_us'
min_frame_samp = 0.5
max_frame_samp = float(60)
max_frame_samp = 60.0
min_timelapse_fps = 1
max_timelapse_fps = 15
min_frame_height = 480
max_frame_height = 9000
min_time_offset = -10.0
max_time_offset = 10.0
## Platform-dependent commands
# Platform-dependent default paths
if platform.system() == 'Windows':
ffmpeg_path = '{}/dependencies/ffmpeg-essentials/bin/ffmpeg.exe'.format(base_path)
exiftool_path = '{}/dependencies/exiftool.exe'.format(base_path)
else:
ffmpeg_path = 'ffmpeg'
exiftool_path = 'exiftool'
## ini file reading
if os.path.exists(ini_file_path):
configuration = {}
try:
# Configuration file reading
try:
with open(ini_file_path, 'r') as file:
for line in file:
if line[0] == '#':
continue
if os.path.exists(conf_file_path):
with codecs.open(conf_file_path, mode='r', encoding='utf-8') as f:
conf_toml = loads(f.read())
f.close()
else:
(key, value) = line.split()
configuration[key] = value.replace('"', '')
raise FileNotFoundError
locale = configuration.get('ui_language')
max_frame_samp = float(configuration.get('max_frame_sample'))
ffmpeg_path = configuration.get('ffmpeg_path').replace('./', '{}/'.format(base_path))
exiftool_path = configuration.get('exiftool_path').replace('./', '{}/'.format(base_path))
except:
print('\nError... not readable or incomplete ini file. Default configuration will be used.')
# Configuration check
reading_parameters = conf_toml['system'].keys()
check_result = existing_items(mandatory_parameters, reading_parameters)
# Localization
locale_file_path = '{}/locales/{}.toml'.format(base_path, locale)
if len(check_result['missing']) != 0:
missing_parameters_list = list_enumerator(check_result['missing'], ', ', ' and ')
if os.path.exists(locale_file_path):
if len(missing_parameters_list) > 1:
verb = 'is'
else:
verb = 'are'
print("(!) {} {} missing in configuration file.".format(missing_parameters_list, verb))
raise ValueError
# Configuration assignment
locale = conf_toml['system']['locale']
exiftool_path = unix_path(conf_toml['system']['exiftool_path']).replace('./', '{}/'.format(base_path))
except (FileNotFoundError, ValueError):
print("\nError... configuration file doesn't exists or invalid.")
default_conf = str(input("Use default configuration instead (Y/N) ? ").upper())
if default_conf != 'Y':
raise InterruptedError
# Localization
locale_file_path = '{}/locales/{}.toml'.format(base_path, locale)
if os.path.exists(locale_file_path):
with codecs.open(locale_file_path, mode='r', encoding='utf-8') as f:
locale_toml = loads(f.read())
f.close()
else:
print("Error.... file for locale \"{}\" doesn't exists or invalid.".format(locale))
ValueError
else:
print("\nError.... file for locale \"{}\" doesn't exists or invalid.".format(locale))
raise InterruptedError
user_agree = locale_toml['user']['agree'][0].upper()
user_disagree = locale_toml['user']['disagree'][0].upper()
path_error = locale_toml['ui']['paths']['path_err']
user_agree = locale_toml['user']['agree'][0].upper()
user_disagree = locale_toml['user']['disagree'][0].upper()
path_error = locale_toml['ui']['paths']['path_err']
# Introduction text
print(locale_toml['ui']['info']['intro'])
# Introduction text
print(locale_toml['ui']['info']['intro'])
# User input
## TOML setting file
toml_setting = input('\n{}'.format(locale_toml['ui']['parameters']['toml_setting'].format(user_agree, user_disagree)))
i = 0
# User input
# TOML setting file
toml_setting = input('\n{}'.format(locale_toml['ui']['parameters']['toml_setting'].format(user_agree, user_disagree)))
i = 0
if toml_setting.upper() == 'O':
if toml_setting.upper() == 'O':
while True:
try:
i += 1
toml_file_path = unix_path(input('{}'.format(locale_toml['ui']['paths']['toml_file']))).strip()
if os.path.exists(toml_file_path):
break
else:
print('{}\n'.format(locale_toml['ui']['paths']['path_err']))
True
except:
print('{}\n'.format(locale_toml['ui']['paths']['path_err']))
with codecs.open(toml_file_path, mode='r', encoding='utf-8') as f:
setting_toml = loads(f.read())
f.close()
break
else:
raise FileNotFoundError
except (FileNotFoundError, ValueError):
print('{}\n'.format(locale_toml['ui']['paths']['path_err']))
True
# <--coding in progress-->
video_path = ''
gps_track_path = ''
raise NotImplementedError
## Paths
else:
# Paths
else:
print('\n{}'.format(locale_toml['ui']['info']['paths_title']))
### Video file
# Video file
while True:
try:
video_path = unix_path(input('{}'.format(locale_toml['ui']['paths']['video_file']))).strip()
@@ -154,7 +206,14 @@ else:
print('{}\n'.format(locale_toml['ui']['paths']['path_err']))
True
### GPS track file
# Video metadatas extraction
video = cv2.VideoCapture(video_path)
video_fps = video.get(cv2.CAP_PROP_FPS)
video_width = video.get(cv2.CAP_PROP_FRAME_WIDTH)
video_height = video.get(cv2.CAP_PROP_FRAME_HEIGHT)
video_total_frames = int(video.get(cv2.CAP_PROP_FRAME_COUNT))
# GPS track file
while True:
try:
gps_track_path = unix_path(input('{}'.format(locale_toml['ui']['paths']['gps_track']))).strip()
@@ -167,35 +226,34 @@ else:
print('{}\n'.format(locale_toml['ui']['paths']['path_err']))
True
### Output folder
# Output folder
output_folder = unix_path(input(locale_toml['ui']['paths']['output_folder']))
## Parameters
# Parameters
print('\n{}'.format(locale_toml['ui']['info']['parameters_title']))
### Timelapse video
timelapse = input(locale_toml['ui']['parameters']['timelapse'].format(user_agree, user_disagree))
# Timelapse video
timelapse = input(locale_toml['ui']['parameters']['timelapse'].format(user_agree, user_disagree)).upper()
if timelapse.upper() == user_agree:
### Timelapse framerate parameter
if timelapse == user_agree:
# Timelapse framerate parameter
while True:
try:
timelapse_fps = int(input(locale_toml['ui']['parameters']['timelapse_fps'].format(min_timelapse_fps,
max_timelapse_fps)))
if max_timelapse_fps >= timelapse_fps >= min_timelapse_fps:
frame_sampling = 1 / timelapse_fps
frame_sampling = float(1 / timelapse_fps)
break
else:
print(locale_toml['ui']['parameters']['timelapse_fps_err'].format(min_timelapse_fps,
max_timelapse_fps))
print('\n{}'.format(locale_toml['ui']['parameters']['timelapse_fps_err'].format(min_timelapse_fps,
max_timelapse_fps)))
True
except ValueError:
print(locale_toml['ui']['parameters']['timelapse_fps_err'].format(min_timelapse_fps, max_timelapse_fps))
print('\n{}'.format(locale_toml['ui']['parameters']['timelapse_fps_err'].format(min_timelapse_fps, max_timelapse_fps)))
True
else:
### Frame sampling parameter
# Frame sampling parameter
while True:
try:
frame_sampling = float(input(locale_toml['ui']['parameters']['frame_samp'].format(min_frame_samp,
@@ -204,15 +262,15 @@ else:
if max_frame_samp >= frame_sampling >= min_frame_samp:
break
else:
print(locale_toml['ui']['parameters']['frame_samp_err'].format(min_frame_samp, max_frame_samp))
print('\n{}'.format(locale_toml['ui']['parameters']['frame_samp_err'].format(min_frame_samp, max_frame_samp)))
True
except ValueError:
print(locale_toml['ui']['parameters']['frame_samp_err'].format(min_frame_samp, max_frame_samp))
print('\n{}'.format(locale_toml['ui']['parameters']['frame_samp_err'].format(min_frame_samp, max_frame_samp)))
True
## Frame height parameter
min_frame_height = 480
max_frame_height = 6000
# Frame height parameter
if video_height <= max_frame_height:
max_frame_height = int(round(video_height, 0))
while True:
try:
@@ -221,123 +279,111 @@ else:
if max_frame_height >= frame_height >= min_frame_height:
break
elif frame_height == 0:
break
else:
print(locale_toml['ui']['parameters']['frame_height_err'].format(min_frame_height, max_frame_height))
print('\n{}'.format(locale_toml['ui']['parameters']['frame_height_err'].format(min_frame_height, max_frame_height)))
True
except ValueError:
print(locale_toml['ui']['parameters']['frame_height_err'].format(min_frame_height, max_frame_height))
print('\n{}'.format(locale_toml['ui']['parameters']['frame_height_err'].format(min_frame_height, max_frame_height)))
True
### Video start datetime parameter
# Video start datetime parameter
while True:
try:
video_start_datetime = input(locale_toml['ui']['parameters']['video_start_datetime'])
video_start_datetime_obj = datetime.strptime(video_start_datetime, '%Y-%m-%dT%H:%M:%S.%f')
break
except ValueError:
print(locale_toml['ui']['parameters']['video_start_datetime_err'])
print('\n{}'.format(locale_toml['ui']['parameters']['video_start_datetime_err']))
True
### Video recording timezone
# Video recording timezone
video_rec_timezone = input(locale_toml['ui']['parameters']['rec_timezone'])
### Time offset parameter
min_time_offset = -10.0
max_time_offset = 10.0
# Time offset parameter
while True:
try:
time_offset = float(input(locale_toml['ui']['parameters']['time_offset'].format(min_time_offset, max_time_offset)))
time_offset = float(input(locale_toml['ui']['parameters']['time_offset'].format(min_time_offset,
max_time_offset)))
if max_time_offset >= frame_sampling >= min_time_offset:
break
else:
print(locale_toml['ui']['parameters']['time_offset_err'].format(min_time_offset, max_time_offset))
print('\n{}'.format(locale_toml['ui']['parameters']['time_offset_err'].format(min_time_offset, max_time_offset)))
True
except ValueError:
print(locale_toml['ui']['parameters']['time_offset_err'].format(min_time_offset, max_time_offset))
print('\n{}'.format(locale_toml['ui']['parameters']['time_offset_err'].format(min_time_offset, max_time_offset)))
True
### User-defined metadata
# User-defined metadata
print('\n{}'.format(locale_toml['ui']['info']['tags_title']))
make = input(locale_toml['ui']['metadatas']['make'])
model = input(locale_toml['ui']['metadatas']['model'])
author = input(locale_toml['ui']['metadatas']['author'])
# Video metadatas extraction
print('\n{}'.format(locale_toml['processing']['reading_metadatas']))
# Video metadatas formatting
print('\n{}'.format(locale_toml['processing']['reading_metadatas']))
video = cv2.VideoCapture(video_path)
video_fps = video.get(cv2.CAP_PROP_FPS)
video_width = video.get(cv2.CAP_PROP_FRAME_WIDTH)
video_height = video.get(cv2.CAP_PROP_FRAME_HEIGHT)
video_total_frames = int(video.get(cv2.CAP_PROP_FRAME_COUNT))
video_file_name = os.path.basename(video_path)
video_file_size = byte_multiple(os.stat(video_path).st_size)
video_duration = video_total_frames / video_fps
video_file_name = os.path.basename(video_path)
video_file_size = byte_multiple(os.stat(video_path).st_size)
video_duration = video_total_frames / video_fps
video_start_datetime_obj = video_start_datetime_obj + timedelta(seconds=time_offset)
video_start_datetime = video_start_datetime_obj.strftime('%Y-%m-%d %H:%M:%S')
video_start_subsectime = int(int(video_start_datetime_obj.strftime('%f')) / 1000)
video_start_datetime_obj = video_start_datetime_obj + timedelta(seconds=time_offset)
video_start_datetime = video_start_datetime_obj.strftime('%Y:%m:%d %H:%M:%S')
video_start_subsectime = video_start_datetime_obj.strftime('%f')
# Metadata recap
print('\n{}'.format(locale_toml['ui']['info']['metadatas'].format(video_file_name,
# Metadatas recap
print('\n{}'.format(locale_toml['ui']['info']['metadatas'].format(video_file_name,
round(video_file_size[0], 3), video_file_size[1],
video_duration, video_start_datetime,
int(int(video_start_subsectime) / 1000),
'{:03d}'.format(video_start_subsectime),
video_rec_timezone)))
# Output folder creation
output_folder = '{}/{}'.format(output_folder, video_file_name)
existing_path(output_folder)
# Output folder creation
output_folder = '{}/{}'.format(output_folder, video_file_name)
existing_path(output_folder)
# Processes
## Frame sampling + tagging (OpenCV + piexif)
print('\n{}'.format(locale_toml['processing']['sampling']))
# Processes
# Frame sampling + tagging (OpenCV + piexif)
print('\n{}'.format(locale_toml['processing']['sampling']))
i = 0
i = 0
if timelapse == user_agree:
if timelapse == user_agree:
frame_interval = frame_sampling / video_fps
else:
else:
frame_interval = frame_sampling
cv2_tqdm_unit = " {}".format(locale_toml['ui']['units']['cv2_tqdm'])
cv2_tqdm_range = int(video_duration / frame_interval)
cv2_tqdm_unit = locale_toml['ui']['units']['cv2_tqdm']
cv2_tqdm_range = int(video_duration / frame_interval)
for i in tqdm(range(cv2_tqdm_range), unit=cv2_tqdm_unit):
for i in tqdm(range(cv2_tqdm_range), unit=cv2_tqdm_unit):
t = frame_interval * i * 1000
video.set(cv2.CAP_PROP_POS_MSEC, t)
ret, frame = video.read()
# Image resizing
if frame_height != 0:
resize_factor = video_height / frame_height
image_height = frame_height
image_width = int(round(video_height * resize_factor, 0))
frame = cv2.resize(frame, (image_width, image_height), interpolation=cv2.INTER_LANCZOS4)
frame_name = '{:05d}'.format(i)
image_name = "{}_f{}.jpg".format(video_file_name.split('.')[0], frame_name)
image_path = "{}/{}".format(output_folder, image_name)
cv2.imwrite(image_path, frame, [cv2.IMWRITE_JPEG_QUALITY, 88, cv2.IMWRITE_JPEG_PROGRESSIVE, 1, cv2.IMWRITE_JPEG_SAMPLING_FACTOR, 0x411111])
## Time tags preparation
# Time tags formatting
time_shift = i * frame_sampling
current_datetime_obj = video_start_datetime_obj + timedelta(seconds=time_shift)
current_datetime = current_datetime_obj.strftime('%Y:%m:%d %H:%M:%S')
current_subsec_time = int(int(current_datetime_obj.strftime('%f')) / 1000)
# exif code
# with open(image_path, 'rb') as image_file:
# image = Image(image_file)
# image.make = make
# image.model = model
# image.author = author
# image.copyright = "{}, {}".format(author, video_start_datetime_obj.strftime('%Y'))
# image.datetime_original = current_datetime
# #image.offset_time_original = video_rec_timezone
#
# if current_subsec_time > 0 :
# image.subsec_time_original = str(current_subsec_time)
#
# with open(image_path, 'wb') as tagged_image_file:
# tagged_image_file.write(image.get_file())
# piexif code
image_exif = piexif.load(image_path)
@@ -346,7 +392,7 @@ for i in tqdm(range(cv2_tqdm_range), unit=cv2_tqdm_unit):
piexif.ImageIFD.Model: model,
piexif.ImageIFD.Artist: author,
piexif.ImageIFD.Copyright: "{}, {}".format(author, video_start_datetime_obj.strftime('%Y')),
piexif.ImageIFD.Software : 'video2geoframes.py'
piexif.ImageIFD.Software: 'video2geoframes.py (v{})'.format(__version__)
}
exif_tags = {
@@ -365,12 +411,17 @@ for i in tqdm(range(cv2_tqdm_range), unit=cv2_tqdm_unit):
i += 1
# Geo-tagging (ExifTool)
print('\n{}'.format(locale_toml['processing']['geotagging']))
# Geo-tagging (ExifTool)
print('\n{}'.format(locale_toml['processing']['geotagging']))
geotagging_cmd = '{} -P -geotag "{}" "-geotime<SubSecDateTimeOriginal" -overwrite_original "{}/{}_f*.jpg"'\
geotagging_cmd = '{} -P -geotag "{}" "-geotime<SubSecDateTimeOriginal" -overwrite_original "{}/{}_f*.jpg"'\
.format(exiftool_path, gps_track_path, output_folder, video_file_name.split('.')[0])
geotagging = os.system(geotagging_cmd)
geotagging = os.system(geotagging_cmd)
# End
input('\n{}'.format(locale_toml['ui']['info']['end']))
except NotImplementedError:
print("\nSorry, this function is not implemented, work in progress ;)")
except InterruptedError:
input("\nEnd of program, press Enter to quit.")
# End
input('\n{}'.format(locale_toml['ui']['info']['end']))

19
video2geoframes_conf.toml Normal file
View File

@@ -0,0 +1,19 @@
# Default configuration file for video2geoframes.py
# See documentation for more information
#
# Last edition : 2024-06-23
# Mandatory section
[system]
# See documentation for supported locales
locale = "en_us"
exiftool_path = "./dependencies/exiftool.exe"
[default]
# Optional section to avoid manual input
# Tags can be overwrited if presents in TOML setting file
[default.tags]
author = "Campanu"
make = "Camera maker"
camera = "Camera model"