I don’t think they report the person detection to ifttt or anything, so you’re kinda stuck. If you have an android phone, you may be able to get creative with Tasker doing something in SharpTools based off the notification.
Note: I haven’t specifically tested any of this yet, but I have many of these, and install them for clients as well. So, I will be getting to this shortly.
In the meantime, as I understand it…
Notifications on Person Detection (NOT on motion-only)
How it works is that it begins recording on motion, and once that starts, it examines the recording to figure out if there is a person present in the motion event that triggered the recording.
It is possible to configure the system to only send notifications when it determines a person has been detected (and to specifically NOT send notifications when it simply detects motion).
At first, this may seem to be a good solution, but…
Unfortunately, due to a built-in delay between initial motion being detected, recording being triggered, and when it times out, and is open to further detection, if we have it configured to only send notifications when a person is detected, that means that (and this is an example directly out of their user forum), if a car drives up in the driveway (which would be detected as motion, but NOT get sent as a notification), and then a person gets out, even though they would normally be detected as a person, in this case, there will be no notification sent at all, unless they are still there, in the driveway, or walking up the sidewalk after the timeout.
Wyze knows about this issue, and I assume they will make some sorts of modifications to their system to ‘fix’ it at some point, but no idea on timeline for anything on this.
Here are links to a couple relevant threads in their forum…
I don’t think it’s possible for them to fix it. Processing power on the camera is already maxed out, so they can’t analyze every frame of the clip. So there will always be some sections of video that aren’t checked for a person.
Well, I’m not talking about anything to do with processing power, or the accuracy or efficacy of the detection algorithms, etc. I’m only talking about the detection and notification logic, timing, etc. My points assume the tech is up to the task. If it’s not, then yes, that’s definitely another issue. lol
i.e. If the system could be configured to keep paying attention beyond the initial ‘motion detected’ event, and actually send a notification when it detects a person (during the period that would normally be within the timeout/cool-down window), then it seems as though the desired behavior should be possible.
Again though, if the processing power in the cams wouldn’t be up to the task, then they will need to ‘fix’ that as well, and just charge us more for these things (I REALLY like the current price points).
Ya, the delay issue is definitely why ifttt has been relegated to ‘last resort’ and ‘just for fun’ when it comes to which HA ecosystems get employed in my world.
Ooo, I just thought of a work-around for the security drawback with person detection in the Wyze universe… separate, distinct, Z-Wave/Zigbee motion sensor mounted directly on or nearby each Wyze Cam (or at least somehow monitoring the same area as each cam).
That way, you can leave the Wyze system configured for Person Detection Only, and still get ‘regular’ motion detection as well.
OK…so then, the way Wyze can solve this problem is to build a distinct motion sensor into the body of the camera. It wouldn’t have to interfere with, or in any way change any of the configuration, or logic already in place on their end. There’s plenty of physical space on these things to add it into the mix, and they’d just need to find space in the onboard chip that holds the firmware, to hold whatever extra bit of code would be needed to track and react to the motion sensor.
Until then though, a hot glue gun or some 3M Command strips (or a nearby perch) should work just fine.