srtgeotag

srtgeotag extracts still frames from drone video and embeds the matching DJI SRT telemetry in each image. It works as both a Python library and a command-line tool.

Features

• Single videos, multiple videos, or every video in a directory
• Automatic same-name SRT discovery, including .SRT on case-sensitive systems
• Sampling by frames per second or by time interval
• JPEG, PNG, and WebP output
• Standard EXIF GPS, altitude, capture time, ISO, exposure, aperture, and focal length
• Complete telemetry preserved as compact JSON in EXIF ImageDescription and UserComment
• Fast, single-pass extraction with FFmpeg; no ExifTool or source-FPS argument required
• Live progress bars for frame extraction and EXIF telemetry tagging

Requirements and installation

Install FFmpeg and make sure ffmpeg is on PATH, then:

pip install srtgeotag

For local development:

python -m pip install -e ".[dev]"

To install directly from a downloaded or cloned project folder:

cd srtgeotag
python -m pip install .
srtgeotag --version

CLI

The SRT is found beside each video by matching its filename stem.

# One frame per second
srtgeotag DJI_0001.MP4 -o frames --fps 1 --format jpg

# One frame every two seconds
srtgeotag DJI_0001.MP4 -o frames --interval 2 --format png

# Several files (all frames go into the same directory)
srtgeotag DJI_0001.MP4 DJI_0002.MP4 -o frames --fps 0.5

# Optionally use one named subdirectory per video
srtgeotag DJI_0001.MP4 DJI_0002.MP4 -o frames --fps 0.5 --separate-directories

# Every supported video in a directory
srtgeotag ./flight -o frames --interval 5

Use --overwrite to replace existing numbered images. Progress is displayed by default; use --no-progress when running in scripts or CI. Run srtgeotag --help for all options.

Python API

from srtgeotag import extract_video, extract_videos, parse_srt

result = extract_video(
    "DJI_0001.MP4",
    "frames",
    fps=0.5,                 # one image every two seconds
    image_format="jpg",
)
print(result.frames)

telemetry = parse_srt("DJI_0001.SRT")
print(telemetry[0].latitude, telemetry[0].longitude)

extract_video() accepts srt=... for an explicitly named subtitle and ffmpeg=... for a custom executable path. By default, extract_videos() writes the entire batch into one directory and prefixes filenames with each video's stem. Pass separate_directories=True to create one named subdirectory per video.

Telemetry mapping

Each output image is matched to the nearest SRT record using elapsed video time. This avoids accumulated timing errors for fractional rates such as 29.97 fps. Latitude, longitude, absolute altitude, capture time, ISO, shutter/exposure time, aperture, exposure bias, and focal length use standard EXIF fields. DJI-only values such as relative altitude, color mode, and color temperature remain available in the embedded JSON payload.

License

MIT

About AutoWaypoints.com

This project is open-sourced by AutoWaypoints.com, a DJI drone mission-planning platform for automated mapping, photogrammetry, 3D capture, and Gaussian splatting workflows.