/Arknights-Drop-Recognition

Image recognition module of Arknights‘ drop screenshot

Primary LanguageJupyter Notebook

Arknights-Drop-Recognition

This project is a prototype module for image recognition of the game Arknights. It is designed to correctly identify the type and amount of loot items, stage name and other useful information from the game screenshots. Simple OCR (for digits show in this game only) and image classification functions are implemented.

Requirements

Python >= 3.5 is required. Required libraries are listed in requirements.txt.

Usage

>>> from fullStream import FullStream
>>> data = FullStream('test.png')
>>> data.print()
STAGE: 1-7
STAR: 3
STATUS: ['正常']
LV: 83
EXP: 2522
ITEMS:
{'id': '5001', 'count': '72', 'name': '声望', 'success': True}
{'id': '2001', 'count': '1', 'name': '基础作战记录', 'success': True}
{'id': '30012', 'count': '2', 'name': '固源岩', 'success': True}
{'id': '4001', 'count': '72', 'name': '龙门币', 'success': True}

Notebook

TODO

  • Access to api of Penguin-Stats
  • User interface or front-end page