cvat_integration

run ellipse_detector

-N image Name -D DataSet Name -S The threshold of ellipse score -R The threshold of Reliability -C The threshold of CenterDis -M The method id -P Working Directory

$ serverless/opencv/ellipse_detection/bin/ellipse_detector -N 027_0003.jpg  -S 0.85 -P . -M 9

ellipse 236.69  155.871 24      17.52   2.94961 0.971709
ellipse 311.22  246.163 27      20.25   3.00197 0.931649
ellipse 250.711 276.921 28      20.72   2.77507 0.908392
ellipse 123.281 208.007 26      18.46   2.6529  0.908055
ellipse 212.087 225.196 26      20.54   2.70526 0.858707
ellipse 97.3799 165.463 24      16.56   2.58309 0.853357

image files should be inside images dir

python module

Python running executable as subprocess

$ python tests/test_subprocess_popen.py
Execute command: serverless/opencv/ellipse_detection/bin/ellipse_detector -N 027_0003.jpg -S 0.85 -P . -M 9
{'_xc': 236, '_yc': 155, '_a': 24, '_b': 17, '_rad': 2.94961, '_score': 0.971709}
{'_xc': 311, '_yc': 246, '_a': 27, '_b': 20, '_rad': 3.00197, '_score': 0.931649}
{'_xc': 250, '_yc': 276, '_a': 28, '_b': 20, '_rad': 2.77507, '_score': 0.908392}
{'_xc': 123, '_yc': 208, '_a': 26, '_b': 18, '_rad': 2.6529, '_score': 0.908055}
{'_xc': 212, '_yc': 225, '_a': 26, '_b': 20, '_rad': 2.70526, '_score': 0.858707}
{'_xc': 97, '_yc': 165, '_a': 24, '_b': 16, '_rad': 2.58309, '_score': 0.853357}

test

before test copy serverless/opencv/ellipse_detection/bin/ellipse_detector into ~/bin dir OR /usr/local/bin

$ pytest tests/test_subprocess_popen.py 

Deploy

Docker build

$ docker build -f docker/Dockerfile --tag jjanzic/docker-python3-opencv:contrib-opencv-3.4.11 .

nucleo deploy

$ cd serverless/
$ ./deploy.sh

Test inside docker

$ docker exec -it nuclio-nuclio-opencv.ellipse_detector ellipse_detector -N 027_0003.jpg  -S 0.85 -P . -M 9

Up CVAT

start CVAT with serverless function

nuctl deploy --project-name cvat \
    --path `pwd`/openvino/omz/public/yolo-v3-tf/nuclio \
    --volume `pwd`/openvino/common:/opt/nuclio/common \
    --platform local

Check function

$ nuctl invoke opencv.ellipse_detector --method POST --body '{}' --content-type "application/json" --platform local

OR

$ nuctl get functions opencv.ellipse_detector --platform local
  NAMESPACE |          NAME           | PROJECT | STATE | NODE PORT | REPLICAS  
  nuclio    | opencv.ellipse_detector | cvat    | ready |     32769 | 1/1 
$ curl -X POST -d "{}" -H "Content-Type: application/json" http://localhost:32769

OR run container with /bin.bash

$ docker run -it -v /home/vplatonov/workspace/globallogic/sport_radar/Pan-tilt-zoom-SLAM/:/home/PTZ-SLAM jjanzic/docker-python3-opencv:contrib-opencv-3.4.11 /bin/bash

Appendix

500 server error

run nucleo container separately

docker run -d -p 172.20.15.32:8080:8070 -v /var/run/docker.sock:/var/run/docker.sock -v /tmp:/tmp --name nuclio quay.io/nuclio/dashboard:stable-amd64

connect nucleo container to cvat_default network https://docs.docker.com/engine/reference/commandline/network_connect/

docker network connect cvat_default nuclio

if linux server was rebooted - start stop function inside nuclio dashboard (qui)