That’s intended for selecting a pre-taken photo without giving an app library access. You’d have to get the user to take the picture then come back to your app.
What you really need is a system dialogue that pops up the camera and only returns the QR code to the app, the way the photo picker can see the whole library but only gives the app the one selected photo.
Pretty sure Android has this. You can make an app without camera permissions, send an intent that opens the built-in camera to take a picture and you are given access to only that picture. It means you cannot record things in the background all the time, and users don't need to make a decision about a sensitive permission.
iOS may have that as well. I think it’s part of that same photo picker interface.
But that’s not what I was imagining. I was thinking of something in the system that did the QR code scanning for you so that you could just point the camera and as soon as it recognized one the app would get the data. That way the user doesn’t have to frame it and take the picture and select that it’s OK to use in the app.
That’s how adding HomeKit devices work. You hit the add device button in the Home app and a view of the camera comes up. The instant it sees one of the HomeKit QR codes it goes away and starts doing its thing. It’s a great user experience.
But third parties can’t do that without requesting camera access first to get access to the live camera view. A system library could provide it.
What you really need is a system dialogue that pops up the camera and only returns the QR code to the app, the way the photo picker can see the whole library but only gives the app the one selected photo.