I have developed a new integration to Shinobi which allows both control of the monitors (cameras) in Shinobi and the use of Shinobi motion detection to trigger motion events in Hubitat. This allows a single interface to Shinobi to manage a variety of cameras and video recording as supported in that platform.
Installation is via the Hubitat Package Manager
The code is in the Hubitat-Shinobi Github Repository with more details in the ReadMe
Feedback and additional ideas welcome!
ooo, i'll have to try this as i recently moved from zoneminder to shinobi and definitely enjoy it
just played a little with it, and i like it. this should help me from having to buy a motion sensor for my garage now to keep lights on
@brad2 is there any way to remove a device? I initially imported 5 cameras, but one of them gave out on me (ok, i was playing with it and bricked it). I went back through the app and only the 4 existing ones show on the list, but in my devices the 5 original ones are there. The option to "Remove Device" is grayed out as well
EDIT: disregard. I removed the app all together, then added back my NVR and it only has the 4 cameras now
A gap in any case - I added some code so the App will now remove child devices when you uncheck them in the App Setup screen.
Thanks for this! I have motion events for some of my cameras working, no errors logged so far. Appreciate the work!
I'm a little late to this game. I've got a working Shinobi with motion and tensorflow object detections and mqtt publishing appearing to work (Mint 19.1, Wyze RTSP camera). When I install the HE app with the API Key and GroupKey I get this error.
From a quick glance at the code, I'm guessing Shinobi did not return what was expected. I'm pretty sure the Keys are correct. Any ideas what to do next other that print out what was received.
Eventually I tracked the problem down to authentication errors on my Shiboni server. The Hubitat App did toss a large error message or two to the log but the child devices were created and they work.
I've played with this a bit, and it's awesome, thanks! My goal is to get the notifications to include tensorcore/object information in the API call as well.
The only issue I run into is when I use my Jetson Xavier NX as my Shinobi server, all my alerts just say they're from 'multipleRegions.' I have set webhook to 1 minute, etc, and nothing seems to help.
When I run it from an Ubuntu VM on my other server, I can get very quick zone-specific triggers, with webhook timeout set to 0. Any thoughts? Almost as though the API call is not formatted correctly. Or perhaps it's sending multiple events despite the webook setting.
I do not have any real insights on that problem between two different devices. On the Hubitat side it is just receiving the webhook calls from Shinobi and processing them. If you set your SHinobi devices to "Trace" for the logging level you will get more detail on what is happening.
Thanks, I'll give it a try. On another note, has anyone figured out how to get 'smarter' motion detection without object recognition? Somehow my Arlos and even cheap chinesium cameras are able to ignore 'gradient' changes like headlights fading across the driveway, or a palm tree swaying repeatedly.
Maybe Shinobi isn't the best answer for that? I've been fighting the region threshold for months.