[Solved] Why might analysis images not be available?

[vitus]

I'm working on tuning the zone for a camera, per the wiki. I've found the analysis images to be extremely helpful in identifying patches of pixels that are causing false positives. However, not every event has analysis images. I also noticed that the option in the monitor's Storage tab is labeled: "Analysis Images Only (if available)". Why might analysis images NOT be available? After all, the camera is recording because the analysis determined that it should be recording, so logically one should always be able to also record which pixels were responsible.

[vitus]

Nevermind. I had the monitor set to "mocord" (record continuously with motion highlighted) instead of "modect" (record only when motion is detected).

[vitus]

What would be really useful is if you could take a manually triggered event and then re-run the analysis on it at various settings, to see which would yield the most precise detection. For example, if I manually started an event, walked through the field of view, and then manually stopped the event. Then I could re-run analysis at various pixel color thresholds and blob sizes.

[knight-of-ni]

You can export the event as an mp4 then play it back with ffserver. Zoneminder would see it as a continuously looping rtsp camera. This is how the "cameras" at demo.zoneminder.com are set up. See this for details:

https://zoneminder.blogspot.com/p/zonem ... a.html?m=1

[Raam401]

But would the replay help detecting the ideal settings? I mean, you can do it manually by replaying mp4s taken when doing the same thing (like walking accross the room) with different camera settings and then choosing the best one, but it's a little bit low tech.

[knight-of-ni]

Your original post sates that doing it the way you described would be useful to you. I am not agreeing nor disagreeing with that. I am merely supplying the technical details to accomplish what you set out to do. It's up to you to decide if this is worth it or not.

If you are asking this question because you are struggling to set up motion detection, then I would recommend you follow our documentation and the countless posts in this forum that deal with the topic. The basic idea is to turn on blob detection, set the sensitivity very high, then compare the motion statistics between events you want to see versus events you don't want to see... then adjust accordingly. For example, if squirrels tend to generate 500 pixels of motion but humans generate 2000, then adjust the camera's motion sensitivity to be require >500 pixels to filter out squirrels.

[Raam401]

Oh, sorry to create confusion, I'm not the original poster who created the thread, I'm a newbie who's trying to learn by reading the manual and the forum

[knight-of-ni]

Ah fair enough. I didn't notice that.