khal-it/flutter_arcore_plugin
ARCore를 통해 이미지 인식을 허용하는 안드로이드용 플러터 플러그인
- Stars
- 87
- Forks
- 26
- 최근 푸시(UTC)
- 2020. 12. 14.
- 프로젝트 상태
- 활성
언어JavaGLSLDartRubyObjective-C
사용 중인 의존성
의존성 목록 2 개
- flutter
{"sdk":"flutter"} - lint
^1.2.0
원본 README
아래는 영문 원문 스냅샷입니다. 최신 내용은 GitHub에서 확인하세요.
README 펼치기 / 접기
arcore_plugin
A Flutter plugin for Android allowing to recongize images via ARCore
- USES ARCore version 1.8.0
Features
[x] Displaying live ARCamera Feed
[x] Recognizing images via ARCore (for more information see: https://developers.google.com/ar/develop/c/augmented-images/)
[ ] Placing Objects
[ ] In-App training of the imgdb file
Installation
Android
- For Setup make sure you migrated your App to AndroidX
- in app/build.gradle set minSdkVersion to 24 and compileSdkVersion 28
defaultConfig {
...
minSdkVersion 24
...
}
and
android {
compileSdkVersion 28
...
}
- In the main function you have to you need to ensure that the binding is initialized before calling
/// ensure that the binding is initialized
WidgetsFlutterBinding.ensureInitialized();
...
- you also have to load file from asset folder and save it into cache
Future<void> _onArCoreViewCreated(ArCoreViewController controller) async {
arCoreViewController = controller;
await arCoreViewController.loadImgdbFromAssets(
tempFilePath:
'/data/user/0/com.peqas.arcorepluginexample/cache/image_database.imgdb');
await controller.getArCoreView();
}
For more information checkout the Example project
# Very important!
* AS FOR NOW THE PLUGIN DOES NOT SUPPORT IN APP TRAINING OF THE IMGDB FILE
* The IMGDB File has to be PRETRAINED and saved in a Temp Directory before the ArCoreView is created
* The IMGDB File name hast to be image_database.imgdb
* you can find a pretained imgdb file under example/assets. The images that the imgdb file is trained with are found under example/assets/tester_images
* for more info about ARCore imgdb files see : https://developers.google.com/ar/develop/c/augmented-images/arcoreimg
Note: This plugin is still under development, and some APIs might not be available yet. Feedback welcome and Pull Requests are most welcome!