sackerman-usgs / UAS_processing

2 stars 3 forks source link

UAS_processing

Creators: Seth Ackerman (@sackerman-usgs), Emily Sturdivant (@esturdivant-usgs)

Jupyter notebook to process image and GPX files.

Detailed workflow:

  1. read the GPX file, which is a telemetry log file produced by the 3DR Solo in tlog format and converted to GPX in Mission Planner. View a dataframe from the GPX file.
  2. Parse the time field in the GPX dataframe. Add fields datetime_utc and epoch_utc.
  3. Export the dataframe as a table in CSV format and make a basic map of the flight path from the GPX navigation data.
  4. Plot the flight path on an aerial photo basemap.
  5. Initialize an dataframe for the images. Include the original filename and the time in UTC, Epoch, and ISO formats.
  6. Export a CSV of the dataframe. Plot the image times and the GPX elevations by time to check that they match.
  7. Rename the photos using the survey number, the flight and camera ID, the time in ISO format, and the original filename.
  8. Geotag the photos from the GPX file using the Geosync tool in ExifTool.
  9. Update the EXIF tags to standard values.

Requirements

Python 3 with modules (in addition to defaults):

These are satisfied by using the IOOS3 environment in Anaconda.

Instructions

input variables:

created variables:

output products:

To use ipyleaflet

conda install -c conda-forge ipyleaflet
pip install ipyleaflet
jupyter nbextension enable --py --sys-prefix ipyleaflet