Modern manufacturing processes have a high degree of automation and at the same time high quality requirements. Our machine vision solution helps ensuring these quality requirements are met by providing means for automatic recognition of defects. In addition to the recognition, it is essential to store the data about defects. This is mandatory for us to constantly improve our models. Also, our customers need be able to analyze for example common defect types.
The service runs on a docker container and set up was implemented using docker-compose
- git
- docker
- docker-compose
- Clone this repo and go to repo folder
git clone https://github.com/fedjo/deevio-project.git && cd deevio-project
- Build docker image of the developed web service
$ ./build.sh <TAG>
- Set docker containers running
docker-compose up -d
- Examine app logs
docker-compose logs -f app
The application is running on http://127.0.0.1:5000
Request predictions for specific image
/api/v1/predictions/<imageId>
Request weak classifications
/api/v1/classifications/weak
You can publish classification results to local mosquitto broker on localhost:1883.
For more info please read the docs