It claims one-button StreetView uploads, so there’s always hope for seamless Mapillary capture and uploads.
I don’t think it will stitch images in camera, you’ll still need the app or insta360 studio to do this.
It looks like they have solved a lot of issues that the old One has, esp. the android compatibilty and not able to access a lot of features without physically connect the camera to an Iphone really sucks.
I also wonder if they improved the in build-in audio sound, because the old One has an awful sound which is too loud when shooting timelapse photos, the sound can only turned off (only with an Iphone of course) but then you are not sure if the camera is still working. Sometimes the timelapse automatically turns off after 5 minutes. So I’d like to know if they have solved those issues with the new X. If that is the case, anyone here interested in an old One?
Being their top-end consumer camera, does it at least retain gps coords?
UPD: some are reporting they strip gps exif in post processing. maybe to sell the gps remote. ridiculous if so
Looks like they now add gps info from the phone to overlays, failing to ship the remote and make this a paid feature. Still likely useless for Mapillary
Got One X for quick testing, result: not suitable for Mapillary use.
- Shortest interval in photo mode is 3 seconds, which is suitable for walking, but nothing else.
- GPS location is only updated once when interval photo shooting mode is started (with Android connected through wifi), therefore all images get just one location facepalm However, this could be easily fixed in firmware… Processed .insp-images with Insta360 Studio in Windows. Phone was also saving gps track in OsmAND at the same time.
- Automatic levelling doesn’t work when camera mounted to car roof, horizon whoops here and there every few images.
- RAW DNG filesize is whopping 36Mb, non-RAW proprietary .insp-format is 4-9Mb.
- GPS location drift about 9 seconds compared gps track saved with phone that was also connected through wifi to Insta.
sad, considering the price. No wonder people mostly go for Fusions.
Let’s see if the new Pro is any better, I’m somewhat wary of this startup that made it now
I have to take back parts of my earlier review “not suitable for Mapillary use”. After all, I decided to keep the camera and it seems to work out ok on an experimental backpack rig.
3 second interval is good for walking, levelling seems to work better when camera really moves around rather than is fixed on car roof.
still sad to see the level of detail on top consumer 360 cameras.
i guess you have to merge with an external gpx track?
Yes, OsmAnd app is always tracking in my android phone while mapping with any camera. As different cameras drift and do what ever they wish (like reset date to 1970 when changing battery…) - simple procedure of taking a picture of accurate clock is never out of fashion
well, looking at the thread stats on this forum, it seems most people go for Xiaomi Sphere (110 posts / 9.4k views), comparing to 16/783 for Fusion
Personally I have chosen Fusion just because it seems to be the most ‘pnp’ camera available for use with Mapillary. You do not have to care of any external gps logging and sync, manual post-processing, etc.
Just removing XMP tags (if it’s still causing an issue with mapillary_tools).
And maybe finding a proper way to adjust default angle pointing to the front during stitching (when using a side-view mount)
Maybe reality is different as I have not got a chance to test it myself yet.
I was referring to general sales. Here on the forum, Xiaomi is indeed more popular, primarily because it is a more affordable way to try out 360 I guess.
The default angle could be potentially adjusted with a batch offset command. Do you mount yours facing to the side for aerodynamics?
at the moment I’m only studying the theory while waiting for delivery. Grabbed a good price tag (400eur) in Germany.
Yes, I am thinking of mounting it facing side-ward because of:
- primarily - keep the lens clean from occasional impact with insects
- aerodynamics - not sure if mounting side-ward will have a big improvement. But more stable on higher speed - that’s for sure.
Re the offset - I guess the issue reported earlier with side-mount is not with the heading/compass angle in Exif metadata - this could be easily fixed with the tools.
Just guessing - it could be related with a geometrical center (front) of the stitched panorama.
hello, guys, does the Insta360 one R will be more suitable to mapillary thant the one X? does any of you tried already?
i doubt it will be any better in terms of image quality, but it might be a bit more sturdy
There is a 1-inch sensor version available for one R: https://store.insta360.com/product/one_r However it seems that GPS is possibly only available in the GPS-remote.
I successfully upload a sequence of 360 photos with my Insta360 ONE R and I would like to share my workflow in the forum to help those that are feeling troubles with it.
I still need to figure out:
- How to make sure all the 360 are oriented to the right direction
- How to make sure the sequence of 360 are stitched smoothly in between each other as a 3D reconstruction. (as we see in this example: https://www.mapillary.com/app/?lat=22.121499665035955&lng=113.56619301620958&z=14.448423330379175&x=0.6343862701932591&y=0.5355096161240673&zoom=0&panos=true&focus=map&pKey=w2ziPab4aXNd4wX7213l4A )
(in this next example i have both of the issues i pointed out above.https://www.mapillary.com/app/?lat=22.17766419537648&lng=113.5444945902318&z=14.360406870855378&x=0.24175850244356178&y=0.4972382179144055&zoom=0&panos=true&focus=map&pKey=KYZQOyZLbT0TBpsPBLzgrQ )
can someone explain me what might be the problem?
I have the same problem with orientation with the Insta360 ONE X. I hope someone can find a solution. I decided not to upload sequences until I find a proper workflow. https://www.mapillary.com/app/user/mikelo?lat=43.3665730203941&lng=-1.8180341901281736&z=17&pKey=-_6cdyXnHl910vm-ne8lug&x=0.054177753297547926&y=0.5429869218106146&zoom=0
Exactly, this is the problem i am facing, how to understand the software that process our intput. Because we don’t have much settings to handle when uploading the 360. But still, some of your 360 are correctly processed, facing the right orientation and with the transition between photos correctly mapped in terms of 3D. (https://www.mapillary.com/app/user/mikelo?lat=43.36696085999752&lng=-1.8198768599978277&z=17&pKey=7Rpqd_gV7oxgYTs5W8RnnQ&x=0.47422164091868657&y=0.4822254648873291&zoom=0 )
I think the question is what makes te software recognize the 3D space and join together all the pictures as a whole? If we re-orient the 360 does it help? Does it help to take photos closer to each other in order to capture more common visual points across pictures?
how do you capture Pictures with GPS values ?
Is this automaticly merges into the picuters during caputring ?
i need a new Cam for taking pictures and have no to choose a gopro hero 9 or an insta 360 one X
So any help would be great
This is a late reply. hope can help.
My process is:
-take sequence of pictures with insta360 one R every 3 secounds;
-at same time, track my GPS location with “JAS Tracker” (app on iPhone);
-Batch export all 360 with the desktop software of insta360;
-export the GPS log to the same folder as the sequence of images
-use the desktop software “GeoSetter” to merge the location into every image;
-export to mapillary with the dedicated desktop software.