Deleting my own images

I got a note from Karin.
Mapillary (Mapillary)
Jul 24, 15:19 CEST

Hi JB,

Thanks for chiming in! What you can do to help is to use our editor and remove those photos/sequences yourself. That way it will go fastest. You can find an overview of the process here:

All the best,

IF I happen to come upon one of my images that I feel needs to be removed, and if I am sitting at my laptop and if I can figure how to delete an image, might give it a go.
The chances of that happening are far slimmer than my winning the lottery (even though I’ve never bought a ticket)
I don’t go looking through my pictures. The last 3 weeks I’ve posted over a quarter million pictures per week.
Some of those are of food, and indeed have no good reason to be on Mapillary.
I’m just not willing to look through over 250 thousand pictures, boring pictures of roads I’ve all ready seen, to find the offenders. (That would be a week’s worth)
It is great there are others who are willing to plow through these and find the .001% that need deleting.
I’ll leave that job for them.
I have my hands full just doing what I do.
But like I said, if I happen onto one, and if I am at my laptop, I’ll give a stab at trying to delete the offending image.
I post here because I’m not linked back to the image with the comment.


a) I’d love to see some AI that could better flag questionable pictures. That’s an insane problem to crack right now though.

b) I pop open my pictures in windows explorer, set it to large icons and cruise through the folder ( NOT windows image viewer ). I can get through a lot very quickly. I get most of them that way. Usually the bad ones I miss are a one off in the middle of a sequence or a folder I just failed to scan through.

In your case, I believe you’re using the phone app for a lot of them. That’s not so easy to mass edit.

One possibility would be to dump them all onto a pc / mac, run through the photos and then upload them via the browser, the python script or just dumping them back onto the phone into the mappilary image folder. It’s an extra step but it could help catch the obvious stuff.

How did Google’s Panimora [sic] handle bad images?


I made a browser at which lets you put in a sequence key and see the images tightly packed in the smallest 320 width format. In most browsers you can just press space to scroll down. When you click an image it will open in Mapillary and you could delete it from there. You must use a shortcut in the browser to open in a new window. Using that, I just found this image:

My tool should work nicely on a mobile, but Mapillarys site is mostly useable on a desktop. My site still uses the old v2 api and I don’t know it it will stop working a some time. The site is open source - it has a link to the sources at GitHub.

1 Like

@tryl you should migrate to the new API v3 since we are indeed planning to deprecate v2 sometime soon (in the middle of migrating all of our own stuff just now). Some more information:

1 Like

I have transfer all onto my laptop, run through the photos and then try to upload them via the browser, Unfotuntaly my system get crashed. It took a lot of time re configure it.

Hi @tryl, I tried to use your browser tool, where I entered the dates where I know I’ve captured some sequences and uploaded, along with my username and nothing happened when I clicked Show Images From Map.

Any ideas?


Hi @laye
It is probably because I wrote it for the v2 of the API while Mapillary have transitioned to a version 3. I got started rewriting it for version 3, but never got it to work properly and have run out of time. The proper way to do it is a complete rewrite of both client and server and I does not have the time for that at the moment.

Nice. When you have a moment, please share the github project name / link. Gracias

1 Like

Here you go:
If you can help fix the issues it would be awesome :slight_smile:

1 Like

Hello - Bumping this thread. I need to delete my images in bulk.

Today I captured images on a new road not yet in the Mapillary maps. Mapillary assigned those images to a different road, so now the 2nd road’s images are wrong. The link at the top of this message no longer works. Any advice?

The new road is properly reflected in OpenStreetMap, but not in Mapillary’s maps.


Hi @gabrielm9

The ability to delete images/sequences has been removed from the interface since the ‘upgrade’ by Facebook. Send us a URL of the area where the error occurred and we can possibly check what has gone wrong. You could also report the issue to

Cheers - Phil

1 Like

Hi @tastrax

Thanks for the reply. The problem begins here. The old US-206 continues to the west, but the new US-206 Bypass continues south. Switch to the OpenStreetMaps layer and you’ll see the new roadway. I captured images while driving on the new US-206 Bypass and Mapillary placed those images on the old roadway. To make matters worse, it started raining just at the moment I started driving on the new roadway so you’ll see several out of focus images.

To avoid this problem in the future, do I need to confirm the new roadway is available on the Mapillary map before I use the iOS app for capture? Or should have have switched to the OSM layer before initiating the capture?

If it’s easier, it would be fine to just delete all my images from 26-Jun-2021 between 11-12 EDT. I can repeat the capture.


I dont think Mapillary does any ‘snapping to roads’. It will simply place the images at the GPS locations within the images. It makes no difference which map layer is showing in the interface.

I dont actually see your images in the system. The ones you link to are from 2017

Looks like your sequences have not yet been fully processed. They will turn up eventually in the correct location

Ah, that makes sense. Thanks, @tastrax. The iOS app indicated images were processing, then uploaded, so I assumed all the processing was complete.


Having said all that @gabrielm9 , there are processes whereby images will be rejected if they are low quality. This may be the case where the rain has made some of your images ‘unusable’. These sequences may get rejected automatically. This is a new process and I think Mapillary are still working on the finer aspects of this. If you find that some sequences are missing, this may be the reason. Again, may be able to provide some guidance in this regard.