Can I batch geotag JPEGs from a CSV by matching each photo’s keyword or ID?

Question

I have many JPEG photos that need GPS coordinates added. Separately, I have a CSV lookup table with a unique ID for each photographed object and its X/Y coordinates. My goal is to manually add the matching ID as a keyword/tag to each photo, then batch-apply the correct GPS data from the CSV so the images can later be shown as accurately placed points in QGIS.

Example CSV structure: ID, X, Y.

Is there a relatively simple workflow to match a photo’s tag/keyword to the CSV entry, convert those coordinates if needed, and write proper EXIF GPS data to each image without renaming the files?

user9161 · Answer

It seems your main problem is:

How can I show photos on a map in QGIS?

You have the following at your disposal:

A set of photos
A lookup table with IPTC keywords and their corresponding coordinates (in the EPSG:25832 coordinate system)

This can be divided into the following sub-problems:

Identify which keyword(s) a photo has
If one of the keywords of the photo matches one in the lookup table:
- Lookup the coordinates that corresponds with that tag
- Reproject the EPSG:25832 coordinates to WGS84 (EPSG:4326) latitude and longitude coordinates
- Convert the lat/lon coordinates to a format suitable for the EXIF GPS tag
- Write the EXIF GPS tag of the photo
Read in the photo into a QGIS map

The stackexchange network is not a coding service, so I cannot provide you with ready to go scripts, but I have suggestions on how to solve the sub-problems.

1. Identify which keyword(s) a photo has

Using ExifTool, no need to install, it's a self-contained executable.
Using a Python library such as:
- exif
- PyExifTool
- Pillow using the _getexif() method

2a. Lookup the coordinates that corresponds with that tag

This can be done in a myriad of ways, but I am familiar with Python where you can e.g. use a dictionary to store your lookup table:

lookup = {
    'A4166': [546043.0552, 6189132.4628],
    'A4165': [546031.2343, 6189128.3107],
    etc...
}

Alternatively you could use a pandas DataFrame:

import pandas as pd
lookup = pd.read_csv('/path/to/file.csv')
coords = lookup.loc[lookup.ID == 'A4166']

2b. Reproject the EPSG:25832 coordinates to WSG84

Again, many ways, but in Python I would use geopandas to do any kind of reprojecting of coordinates. Assuming you would have used a Dataframe in the previous step:

import geopandas as gpd
gdf = gpd.GeoDataFrame(lookup, geometry=gpd.points_from_xy(lookup.X, lookup.Y)
# Let geopandas know that the data is in EPSG:25832
gdf.set_crs(epsg=25832)
# Reproject it to 
gdf.to_crs(epsg=4326, inplace=True)

Now you have a lookup table in the variable gdf with the coordinates in the WGS84 coordinate system.

2c. Convert the lat/lon coordinates to a format suitable for the EXIF GPS tag

The ExifTool documentation describes how the GPS tags look. You will have to create some conditions based on the coordinate sign on when to write N or S in the GPSLatitudeRef and E or W in the GPSLongitudeRef fields.

2d. Write the EXIF GPS tag of the photo

The same libraries from step 2a can be used to write to the appropriate fields.

3. Read in the photo into a QGIS map

This tutorial was the first hit on Google when I looked for "place photos in QGIS". Roughly the steps are:

Import Geotagged Photos via Processing Toolbar > Vector Creation > Import Geotagged Photos

View photos in popup: set the HTML Map Tip in the Display property of the layer to:

If you run into any issues with solving one of the sub-problems you are very welcome to edit your question (or post a new one) showing what you've tried and where you get stuck.