HomeAssistant and Deepstack in Docker on Jetson Nano

Third party integrations
1

Hi all.

I am using the Docker Installation of HomeAssistant on the Jetson Nano.
I used HACS to install the custom component for deepstack object detection by @robmarkcole

Problem:
Deepstack does not create files in my deepstack folder in homeassistant.

Configuration.yaml

homeassistant:
  whitelist_external_dirs:
    - /tmp
    - /config/deepstack
    - /config/camera_motion
  media_dirs:
    local: /config/deepstack/

media_source:

image_processing:
  - platform: deepstack_object
    ip_address: 10.1.1.200
    port: 80
    save_file_folder: /config/deepstack/
    save_file_format: png
    save_timestamped_file: True
    always_save_latest_file: True
    scale: 0.75
    roi_x_max: 0.8
    roi_y_max: 0.8
    targets:
      - target: person
      - target: vehicle
        confidence: 60
      - target: car
        confidence: 40
    source:
      - entity_id: camera.cube_hikvision

I run Deepstack with:

run --runtime nvidia -e VISION-DETECTION=True -v localstorage:/datastore -p 80:5000 deepquestai/deepstack:jetpack

Now I use the image_procession service and the Deepstack terminal shows:

toto@toto-desktop:~$ docker run --runtime nvidia -e VISION-DETECTION=True -v localstorage:/datastore -p 80:5000 deepquestai/deepstack:jetpack
DeepStack: Version 2021.02.1
/v1/vision/detection
---------------------------------------
---------------------------------------
v1/backup
---------------------------------------
v1/restore
[GIN] 2021/04/30 - 17:57:53 | 200 |   19.2577713s |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 17:57:53 | 200 |   2.32863163s |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:33:26 | 200 |   343.13171ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:34:04 | 200 |  291.541356ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:34:06 | 200 |  276.001673ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:35:02 | 200 |  275.417861ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:35:05 | 200 |  284.229022ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:35:11 | 200 |  244.277822ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:42:21 | 200 |  307.781014ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:42:24 | 200 |  336.591342ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:42:27 | 200 |  271.352372ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:43:17 | 200 |  284.660314ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:43:20 | 200 |  315.020571ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:43:24 | 200 |   242.76217ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:43:34 | 200 |  284.635823ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:47:26 | 200 |  260.039553ms |      10.1.1.200 | POST     /v1/vision/detection
[GIN] 2021/04/30 - 18:53:28 | 200 |  256.655056ms |      10.1.1.200 | POST     /v1/vision/detection
2

Hi Mate,

Exactly the same issue for me, I note that there appears to be some permission issues… so maybe I need to clean this up… I really don’t know what else to do.

Anyone got any ideas?

3

It could solve it but i forgot to post and i am using frigate now. It was an issue with the path/volumes

4

hello, how is it going frigate on the jetson nano?

5

Hello,

Would you mind sharing your experience running HA on Jetson Nano with object recognition/face detection/etc?

I have one unused Nano 4GB and would love to put into action.

Thanks