The fish eye effect makes it impossible to set a detection zone looking through window of front lawn that is rectangular. Or sidewalk as a boundary or street. All these curve but the detection selector is a rectangle - one can not align the zone to the image.
Either multi-dot trapezoid or multiple rectangles would improve this.
Just a request for more detailed detection zones. Would be great if there could be exclusion zones (boxes or circles maybe). Or just allow detection zones to be made up of multiple boxes potentially overlapping so a person can make up a detection area that’s not just a single box. For example a camera can have a bush in that gets blown by the wind; allowing exclusion of that sort of thing would be super helpful.
@WyzeGwendolyn, How do we get this feature? I’ve uprgaded the app to 2.15.51 on Android, 2.15.41 on iOS, firmware is 126.96.36.199, and I still don’t see any changes. (I don’t see a beta version in the App Store or Google Play.) In fact, now both the Android and the iOS versions are limiting the minimum size of the detection zone area to a relatively large box. I’d be interested in helping to test it if it’s still beta.
I have a row of bushes which appear diagonally in my cam. In the wind, all the leaves flutter causing motion tagging (and event triggering). However, instead of being seen as a small diagonal region of motion, the motion-tagging shows as the bounding box for the entire row of bushes, which is a much larger region than is actually moving. See attached grab.
This problem forces me to define a detection zone which excludes a MUCH larger than I would like to – I would like to exclude just the bushes and not the much larger bounding box for the bushes. Its very nice we now have “tiled” detection zone, but the motion tagging is still old school.
Currently, detection zones are static and user-defined. I propose the notion of “Dynamic, Computed Detection Zones” where the camera keeps track of the recent motion-tagged regions and automatically excludes those regions which have frequently triggered an event. The more frequent the region appears in the list, the longer it should be an excluded region.
This would allow trees, leaves, shadows, traffic to be automatically excluded if they are triggering events often. Also, as the sun moves shadows around, the excluded regions could adapt as the older active regions decay out of the recent-event-list. This scheme requires no AI to ascertain what is triggering the event, it just blindly masks off regions which are triggering many events in a short period of time.
With this feature, Wyze could allow more than one 12-second event per 5-minute window as long as the trigger came from a region which is NOT recently triggered.
As for insects flashing across at night, this computed-zone scheme would not help, but another heuristic could be applied: very short duration events should be ignored.