Dedensify images?

Imagery, data, and integrations
1

So, I’m brand new to all of this. It looks like a lot of fun.

I deliver food here in Blacksburg, VA, which doesn’t have a lot of coverage yet. I deliver food all over town all day. I could probably get images of every street in town if I just started up my Mapillary app at the beginning of each day.

However, that would create a huge surplus of images for most of the town. We really don’t need 200 images for each five meter segment of Main Street, or 1,000 pictures of the Mcdonald’s parking lot where I wait for calls. But I also can’t be distracted with turning the app on and off all the time. I need to focus on my driving.

Is there a strategy for paring down the number of images for an area to a manageable density?

2

The original system (that I still use) has is a minimum distance moved between each frame. I think it is set to 2m but it can be set manually. Your sequence will get pruned on processing, so Maccas parking lot is only covered on the way in and out. The same kind of thing happens at traffic lights.

High density of the main streets just happens. A new set of frames every week I wouldnt say is excessive. I’d expect if it got huge in future Mapillary would create a dupe removal engine or you could even request it . The more data that is present the higher accuracy there is of object detections.

(Note that I don’t use the mobile app but I understand that delta distance based frame grabs are adjustable)