What are the ways to quickly sort photos by subject?

Question

In essence, the point is that I mainly upload my photos to Wikipedia, where the main sorting criterion is the object in the photo, i.e what is on the photograph. Unfortunately, the software of the Wikimedia Commons repository itself is built on uniform processing of photos. There are stand alone systems for mass processing of photos, however, it is somehow necessary to ensure that the photos are sorted according to what is on them.

As a photographer, I can take thousands of photos per week, but processing and sorting them according to the object or objects that are photographed is so difficult (I do it manually and I often have to search for information about the object) that I tend to accumulate photos on my disk rather than uploading them.

So my question is, wether there are any methods or software to ease this process?

I have recently started using the system of taking a photo of the name of the location in the maps on my mobile, so that I know from which photo the new location starts. I take a picture of my hand between individual subjects of photography (previously it was guided by the time delay) so that I know that a new one is starting. It would probably be possible to take a picture of the given object from the map before each photo shoot, but this is not applicable to all objects and even the information in the graphic display of the map is limited.

I don't have the coordinates in the EXIF of most of the photos. I am now recording them via the GPS receiver in the EXIF, but the camera does not allow recording the shooting direction, so the EXIF will only contain the position of the photographer. The objects I photograph do not necessarily have to be contained in some electronic database - often these are databases that do not have a machine-readable form.

I must add that the superficial depth of identification is insufficient. For example, tagging images as buildings, churches, or butterflies (like image recognition would do) is not enough. It is necessary to determine a certain degree of specificity to the photos, for example, that it is the Town Hall XY, or at least the house at the address XY.

Answer 1

Image recognition software is available but has several risks, as google found out, so you would need to either spend a lot of time training a machine learning image recognition system (which Google has done) or possibly initially uploading to Google Photos, letting them do it and vetting the results.

A reasonable possibility would be to use a note taking application on your phone that timestamps the entries and then use the timestamp to categorise your photos from the note. It would be nice if there was a voice to text option but there are many venues where talking is discouraged.

You can also look at extracting the GPS data (Lat/Long) e.g. with exiftool or as in this example and querying a service for what is at/near that location geopy provides the tools to do this but you may need to pay for the actual sevrice(s) queried.