Motion Detection Algorithms...
-
I own several Reolink cameras including the 410WS 411WS and Keen. I love these cameras, but on all three, I feel like the motion detection algorithms could be significantly improved. While the sensitivity and area features are nice, the sensitivity does not take into consideration a bug flying past the camera at night and often triggers false alarms. With some basic feature recognition and contiguous feature settings, I think these algorithms could be improved significantly. Here are some ideas:
- Provide a setting for how long motion must be detected before a recording is triggered. Again, this would eliminate a recording for a bug or bird that flies by the screen in 1/4 second. This would probably be my biggest suggestion.
- Provide the user with the ability to adjust 'contiguousness' - how many pixels of motion must be adjacent to each other. I think this would help eliminate some of the array sensitivity false alarms I'm seeing.
- Provide sound as another way to trigger recordings. Easy to do, just have a dB and time setting for audio triggers.
- Use some feature recognition algorithms to detect the difference between a bug flying past the screen (super high pixel values in a straight line) vs. a human walking
- Provide the user the ability to show moving pixels/areas on the live video (maybe as red pixels)... this may help the user define or eliminate areas, and determine what is triggering motion (as it is often not obvious).
Thanks for your consideration! -
I totally agreed with all the points there!
Hope to see an upcoming firmware with those features, if not all, at least some of them. -
Thank you so much for sharing your brilliant ideas. They seems reasonable.
They will be forwarded to our R&D team for analysis. It takes time to make them all happen. But we will be thrilled to see some of them become reality. With that, the cameras will certainly be more intelligent and more usable.
Thanks again.
Wish you a nice day. -
Sounds great thanks! Any updates?
