ONVIF events/motion

I stumbled across bluecherry today and gave it a whirl. Honestly I think it has a lot of potential and am really impressed with it so far.

I was able to get it set up and going fairly quickly, as a note, if you don’t have apache running by default, you shoudl call out that you need to make sure apache starts and can bind to ports. I saw that bc-server process was bound to 7002 even though all the documentation claims 7001, until I realized that it uses apache as the web host and that apache couldn’t bind to ports :slight_smile:

Anyway, I did some more digging around and was unable to find any ONVIF triggered events/motion.
My server is running on a simple i3 NAS server, so I want to offload all the loading to the cameras (as I’ve done in my previous NVR softwares). They simply subscribed to the camera’s ONVIF’s motion events and then it provided easy viewing for me when I reviewed footage. I don’t think I see this in the software (and for that matter, I don’t see a way to disable software motion control either).

ONVIF events is something we very much want to add. We are working on a v3 release now and while better ONVIF support is coming later in v3 it will not be in the initial v3 release. Right now you have options for software motion detection (cpu intensive) or continuous recording.

An i3 should be able to handle 8 cameras. If you build from the master branch there is some GPU enhancements that offload motion detection to the GPU, so give that a try.


Glad to hear ONVIF events are coming.

As a note, a hook for being able to push external events from something other than the camera would be useful too.

I pipe my camera’s video to a OpenCV instance and do object detection in that. Would like to be able to send events to the server to tag events appropriately. A lot of my intruder/alarm logic is handed off to wrappers I have around OpenCV.

1 Like

I’m interested in this feature as well. Currently testing BC as a solution for my home CCTV platform. The motion detection works fine, but too many false positives. Haven’t played too much with the detection settings yet, but I suspect it’ll never be entirely accurate without object detection in the mix.