Is there a list of what camera's have been used on Mapillary?

Contributing and equipment
1

I’m imagening something like taginfo but for Mapillary.
Where you can see how many images have been made using which camera.

But tbh, just seeing a complete lists of camera’s that have been used on Mapillary, with a way to filter by them would be interesting.

Now for example, There is a possibility that someone made imagery using that new Kandao Qoocam 8K.
But there is no way to tell because we can’t search for it.

Camera’s a wish to know if they have been used on Mapillary:

  • Insta360 ONE X
  • GoPro Max
  • Kandao Qoocam 8K
1 Like
2

your best bet might be asking in the thread about these particular cameras. I don’t see dev resource being spent on this any time soon

3

I have never used it but the API image/camera data seems readable

I assume not a keyed field though and a large search.

The camera make/model is also visible on the web gui, but I’d suspect that since that data is optional (eg in mapillary_tools) there is going to be a high percentage of image data without it.

4

The best way to check would be using the API as @bob3bob3 suggested. You can also see the make and model by selecting ‘Image options’ on the right-hand panel.

Here is an example of a GoPro Max sequence. The workflow for this camera is not straight forward and we are doing some internal testing. I’d hold off purchasing until we have more information.

I’ll post here if I find example sequences from the other two cameras.

1 Like
5

As far as I know, camera makes and models cannot be queried directly from API. So in order to determine the cameras used, all the metadata has to be downloaded. I have some image data in store from last year and the top 15 cameras there (specific regions and limited timeframe) are:

make model count
samsung SM-G950F 289919
Apple iPhone 160138
samsung none 73308
RaspberryPi RP_ov5647 48306
samsung SM-G903F 44976
Apple iPhone7,2 39130
Sony D5503 37799
GoPro HERO7 Black 31608
Apple iPhone6S 30643
LG Electronics LG-E430 29817
Blackvue DR900S-1CH 26780
Apple iPhone8,1 26101
NOKIA RM-914_eu_euro2_367 25512
HUAWEI WAS-LX1 21729
Garmin VIRB 21174

There was one Insta360 sequence:

Update:
I downloaded all image metadata for 2020 january and compiled a list of all cameras used:

Every model has a link to image sample. I kept the data as is, so there are duplicates caused by different capitalization. Also, several smartphone manufacturers use their internal model numbers so there are lot of Samsungs and Huaweis which are not really that different.

2nd update

Same for december 2019:

This should give a pretty good idea about what cameras are used and in some occasions, what shouldn’t be.

3 Likes
6

Impressive work, but the numbers seem to be way off.
I looked at my cameras and did not see what I expected.
For instance:
Sony HDR-AS200V 1401 pictures in december 2019.
I beat this with my first four sequences in december 2019.
Mapillary 613 pictures
Mapillary 418 pictures
Mapillary 544 pictures
Mapillary 818 pictures

Even more easy with only two sequences:
RaspberryPi RP_OV5647 898 pictures in december 2019
Mapillary 741 pictures
Mapillary 760 pictures

I’am pretty sure I had at least 4-6000 pictures with the Raspberry and over 10000 with the Sony.
Would like to see the “real” numbers

7

I guess my queries hit some limit on either side after first 1000 pages or so. But I think it serves its purpose by giving rough proportions. Probably someone within Mapillary can get the real numbers, but it might not be a trivial effort considering the amount of data.

Thought about getting sequences instead of images, but sequences only have camera make, not model.