Detection and record video from multiple camera

Cameras
,
1

Hi, is it possible to have multiple camera and when a camera detect a person and start recording and when the person disappear from the camera, but detected on another camera, the video recording continue in the same file, but add the other camera feed.

With this feature, we have one video and see all movement from a person on multiple camera.

2

Highly unlikely. That would take WAY more intelligence than I would expect from a simple system. And obviously that would only be possible on cloud based recordings since each camera knows nothing about other cameras and their recording.

1 Like
3

@sgauvin, I agree with @K6CCC, and I think you may want to take a look at the Multi-Camera Timeline View to achieve a similar effect. See this part of the video starting at 3:36:

This is from an announcement topic that posted earlier this year. That might be worth a skim to learn more.

4

Look at the monitor tab in the app.

IMG_1043
IMG_10431620×2160 626 KB

5

Seems like OP wants sequential clips stitched together or uploaded to the cloud, not 24x7 recording and having to find the spot on each cam. But that’s pretty much what you have to do. Have done it on mine many times, once you know the time on one, pretty easy to follow the person on all of them (assuming you have constant SD card recording or the motion is enough to trigger an event on each one).

1 Like
6

There’s a wishlist topic asking for that feature out here somewhere.

7

Y’know, now that I’m thinking about it, I could see a use case for Wyze’s new AI Video Search doing something like this. Maybe at a future development stage they could create a platform where…

  1. A user submits a search query for a particular target, say a dude prowling around wearing a red headband.
  2. The search platform returns thumbnails of possible matches.
  3. The user can review the thumbnails and/or video clips for genuinely relevant content (i.e., clips that actually show the suspicious dude).
  4. The user can then select the actual matches (true gnarly dude positives) and submit this back to the AI tool.
  5. The AI can then stitch the clips together in proper order (because it can determine time) to produce a seamless multi-camera single-video download that the user can then save locally and take to the authorities, showing them all the places where this nefarious character with the questionable headwear was skulking about the premises.

In step #5, I can imagine the possibility of two cameras providing coverage of the same area, so there would be some time overlap. I envision the default would be to provide both views in the final video (so your final product video continues the first viewpoint until the shady guy is out of sight, and then the transition to the next video goes back in time to where the second viewpoint initially picked up the guy and continues through until another viewpoint switch; that way there’s no loss of the double coverage), but maybe there’d be an option to select the favored camera vantage point.

It’s an infant idea with room to grow.

8

Kinda seems like trying to race a Ferrari in a Corolla.

9

:racing_car:
Whee!