Motion algorithm to detect camera vibration caused by wind

My camera is mounted securely with 3 screws, but the wind still shakes it enough to cause it to record movement. A detection algorithm could be created to allow compensation for slight movements caused by the vibration. This “anti-shake” algorithm is common in handheld cameras, as well as some security cameras. Here is an open source link to an example. GitHub - daien/camocomp: Camera Motion Compensation

That is a really cool project (I went and read about it). Wyze has used a lot of opensource stuff before.

I don’t know if this will prevent many motion events though. Usually if the wind is strong enough to move the camera, that means it is outside, and the wind will also be moving other things, even just causing plants to move or their shadows, or any number of other things outside that will also shake or move. So in this instance, while it could compensate and differentiate between foreground motion and motion caused by camera [de]stabilization, any time there is camera destabilization from wind, 99% of the time there would also be foreground motion caused by the wind as well, negating the point of the stabilization correction. Wyze’s current solution for wind false alerts is using Cam Plus to just tell us when there is a person and ignore the wind, etc.

Having said that, better stabilization for everyone (even those without the AI) would still allow for better recovery of clear video (instead of blurry video) and improved AI detection in general, so it is still definitely a great thing for Wyze to look into and play around with this because there are many potential benefits, especially if it is added as an option we can toggle for those of us who would like it. And it might work better in windy conditions than I am thinking, especially since people can always modify their motion sensitivity levels to compensate for any minor debris or shadows or whatever else might be detected as motion.

Anyway, thanks for sharing, that’s great opensource suggestion.

Thank you for the thoughtful response. Some further information in my case. I do use detection zones to block out areas where wind causes irrelevant movement. This does not impact my ability to detect humans and animals - both are of interest to me.

Wyze has a pretty effective choice of tools that allow me to hone in on what is of interest. The camera vibration is one for which I have not found a work-around.

Thanks for putting my request on the wish list, and for taking the time to research the link I supplied.


As @wjlapaire indicated I suspect this estimate is overstated. Plenty of outdoor camera views would have minimal motion even in high winds. As long as plant life and flags are out of view I would think it could be a majority.

On the other hand I haven’t noticed this as a problem since I only care about person detection.