Table of Contents
This project is a flutter application that allows you to perform object detection over IoT devices figures. To use this application is first necessary to deploy this project into a Docker container, then you can build and install the application in your device.
Follow the steps below to build the app.
This section list any frameworks and libraries used to bootstrap this project.
Follow this instructions to build the entire project.
- Install docker
- Install open JDK 21
- Install android studio
- Install and configure flutter (it is recommended to follow this guide)
This application sends the photo taken to an external server which will process it and returns the picture with rectangles and labels above the recognised objects.
To deploy the following services I recomend to use the dockerfile in this repository.
-
Clone this repo
$ git clone https://github.com/davidemonnati/TBDM-AndroidObjectDetection $ cd ./TBDM-AndroidObjectDetection
-
Clone TBDM-VGLS-2023 repo
$ cd docker $ git clone https://github.com/PROSLab/TBDM-VGLS-2023
This operation must be done manually because the TBDM-VGLS-2023 repository is private.
-
At this point it is necessary to create the detectron model and the .env file. To do this, please consult this guide.
-
Copy the newly created files inside the TBDM-VGLS-2023 directory
-
Run docker compose file
$ docker compose up
If you have flutter installed, you can check the status of flutter installation using the following command: $ flutter doctor
.
- Flutter
- Android toolchain
are correctly installed.
If you have problems check this link.
Finally, to build the apk open a new terminal on the root of this project and use this commands:
$ cd app
$ docker build apk
When the build is finished you can find the apk here: build\app\outputs\flutter-apk\app-release.apk
Davide Monnati - @linkedin - [email protected]
Christopher Pinesi
Project Link: https://github.com/davidemonnati/TBDM-AndroidObjectDetection