I know this forum is not the helpdesk 
But I would really appreciate some help …
rename the folder to lowercase
1 Like
nice eyes. thanks
It’s kinda working now 
Some small topics that I will need support
I managed the integration part, and create a small automation to trigger the processing scan everytime the camera is also trigered in Blue Iris via MQTT. Problems.
a) It detect my cars in the garage. I have car and persons in the config.yaml. Does it mean it will dettect both but not distinguish between the 2? I ahve no interest in detect the car in the garage.
b)The big problem is that HA is updating the video from blue iris every 10 seconds. I guess this doesn’t allow Deepstack to work properly. Does anyone know how to change this?
Thank you very much. I started with this and cannot stop. Great work on this one!!
Hey,
Have anyone tried using the snap.jpeg snapshot feature the ‘ubnt’ cameras provide?
http://xxx.xxx.xxx/snap.jpeg?
I’m facing the problem that my cameras are always 10-15 seconds behind, so when I trigger the face/object detection the actual picture used from my stream is taken 10-15 seconds later.
Help ? ? Please ?
@sender you are reporting an issue with the glance card, which is nothing to do with the deepstach integration. I think the grey image occurs when there is no up to date image from the camera
@robmarkcole thanks for replying. i was not asking about glance card. some one came up with a tip for using it. I am new to deepstack and hass… I dont know where to start can you advice me on how to count vehicles or people on my cam stream? And how to do face recognition on my cam? Dont know where to start.
I ran through this post, but I may have missed it. Is it possible to run this on a Raspberry only? I mean without the use of a usb stick.
No, the stick is required with deepstack on rpi. If you have an rpi 4 you can run tensorflow-lite instead
1 Like
Removed config and removed vm. too complex for me. thank you and goodluck.
try out the sighthound integration instead, its dead simple
Sorry, do not like cloud. Thats why I started to read this entire thread untill it was too hard. Keep up the work!
I’ve been using the Deepstack integration for a while now and it has been working great so thanks @robmarkcole!
However, ever since I updated my HA to 0.107.7 I have been receiving these error messages in my logs when trying to call the Deepstack Image processing service:
I’m calling the Deepstack Image Processing service through an automation which is triggered when my PIR sensor detects motion.
Any help / suport / guidance would be appreciated to get it back up and working!
Hi
Hope someone can shed some light? I have deepstack running on a laptop and Hassio running on a PI.
How do I configure Home assistant on the PI so the laptop can do the processing? When I add the settings to my configuration.yaml on the PI file I get error "image_processing.deepstack_object - integration ‘deepstack_object’ not found.
Any assistance will be appreciated.
I’m running a HASS-Deepstack-object container and have a Coral.
How can I get it to use it?
Do I need a Pi and coral-pi-rest-server?
@quintinsleeking you just provide the IP address to the laptop in your HA config
@surge919 yes you can use https://github.com/robmarkcole/coral-pi-rest-server
@Jaysbeekay check the HA release notes for any breaking changes affecting automations
Hi Rob
I am bit lost.I put the code below in my configurations.yaml file on the Raspberry, I get error "image_processing.deepstack_object - integration ‘deepstack_object’ not found.
Deepstack is running on the laptop 192.168.42.111 and Hassio on the raspberry 192.168.42.1
image_processing:
- platform: deepstack_object
ip_address: 192.168.42.111
port: 5000
api_key: Mysecretkey(my code)
# scan_interval: 30 # Optional, in seconds
save_file_folder: /config/snapshots/
save_timestamped_file: True
target:
- person
- car
source:
- entity_id: camera.house
name: deepstack_person_detector
What should I do on the Raspberry?
Either deepstack is not running or it is not reachable (possibly port issue).