GithubHelp home page GithubHelp logo

Comments (5)

NickM-27 avatar NickM-27 commented on May 31, 2024

I have a couple of cameras that are constantly detecting motion due to shadows/leaves/trees blowing in wind but only on one side of the frame

sounds like a good place to use motion masks

motion detection isn't run on the full size frame, so zones likely wouldn't scale very well and also zones are meant for a different purpose

from frigate.

kirsch33 avatar kirsch33 commented on May 31, 2024

I have a couple of cameras that are constantly detecting motion due to shadows/leaves/trees blowing in wind but only on one side of the frame

sounds like a good place to use motion masks

motion detection isn't run on the full size frame, so zones likely wouldn't scale very well and also zones are meant for a different purpose

in my specific case, these are areas where I don't want to mask because "real" motion is definitely possible in these areas and its a significant portion of the frame. It just needs a different contour area/threshold versus the side of the frame that looks 40ft down the sidewalk with no trees.

would there be any way you could see a functionality like this fitting into the frigate framework? if not zones, then different motion settings by grid region or something? I'm curious if this is something worth my time trying to experiment with or if it is just fundamentally not a practical idea.

from frigate.

NickM-27 avatar NickM-27 commented on May 31, 2024

in general motion being detected during windy times is not a big issue, I don't think aiming to only ever have interesting motion detected as motion is a good goal because that is why object detection exists - to verify the motion that is detected.

If your concerns are around resource usage, we already have ideas on how to reduce that considerably in a future version of frigate.

It is possible that far in the future frigate could have some sort of "dynamic" motion detection by knowing what motion was due to a real object historically and do something like this automatically but I am not sure if that effort is worth it due to the reasons outlined above

from frigate.

kirsch33 avatar kirsch33 commented on May 31, 2024

just to provide some context (not that it changes your response any) I don't think its solely a windy day type of issue for me. a frame from my front camera is attached below. its eave mounted, pointed down quite a bit and has a wide FOV. the result of this is that a valid person on the extreme left side of the frame versus at my porch versus at the far right side all have vastly different sizes. so its difficult to settle into a contour area/threshold that gets me good motion detection in the grassy area below the timestamp without also making motion too sensitive on the left side.

also, as to not masking the left side as you mentioned above, 5 months out of the year the big tree to the left has no leaves and people are much more visible over there and I would prefer to retain the ability to trigger motion in that area.

272652945-f153a907-233d-452d-b2d3-92fd93e9cc46

from frigate.

NickM-27 avatar NickM-27 commented on May 31, 2024

right, I think even if you consider just the part behind the branches, different motion settings would be needed for a person walking vs the branches moving on a windy day. In general I would recommend just using a motion setting that ensures all true objects are detected. Some uninteresting motion being detected is not a problem in of itself

from frigate.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.