ZoneMinder Forums

Hi all,

I have been trying to debug why the ZM motion detector quite often ignores obvious motion ( me walking in front of the camera for > 5 secs ) while at the same time reports minor movement of trees. This means a lot of tweaking of the Buffer + Zone sensitivy settings as well as dis/enabling diagnostic images, etc.

Every time I try a new setting, I have to go out and walk in front of the camera for a while and go back inside to see if the new settings are any better. This is a very cumbersome process.

I have multiple events that I recorded with 'Mocord' settings that are not detected by the motion detector. I wish I could just refeed these events with new settings to see what the results of new settings are. This would greatly improve debugging why the motion detector does *not* detect motion as opposed to debugging false positives.

Would it be very difficult to do this? Maybe it is already possible via commandline tools? Please let me know!

I too would really like this feature.

During the long summer days I get lots and lots of "motion" which turn out to be the clouds moving in front of the sun altering the overrall brightness level. I'm currently trying some preclusive zones which do help a little but doesn't solve the problem completely, and of course I can't try the preclusive zones in new positions when an overcast rainy day is forecast for tomorrow.

mhens wrote:Every time I try a new setting, I have to go out and walk in front of the camera for a while and go back inside to see if the new settings are any better. This is a very cumbersome process.

...and the walks in front of the camera also may not be an exact replication, since if the daylight or lighting has changed since the missed event then that can skew the results.

I agree, a valid request but in the interm maybe this will help: http://www.zoneminder.com/forums/viewto ... ore+frames

I have used a technique of recording a stream to a mjpeg file and the just set that as the camera source. I then have ZM spin up that camera and its does the analysis. I recommend having a decent tail after the event so that the camera doesnt fail right away (reach end of file). In addition you may want to look at my contribs, one of them dumps diagnostic images and additional logging data to FIFO which can make zone setup and debugging a lot easier.