Skip to content

An auto-controller for Minecraft utilizing OpenCV. Features modes like, avoidance, running, harvesting, mining, and guarding using trained yolov5s models.

License

Notifications You must be signed in to change notification settings

decoles/MinecraftOpenCV-AutoController

Repository files navigation

MinecraftOpenCV-AutoController

An auto-controller for Minecraft utilizing OpenCV & OpenCV's Deep Nerual Network Module. Features modes like, avoidance, running, harvesting, mining, and guarding. Detections utilized by trained YoloV5s models converted to ONNX.

Features

The program provides some neat features such as...

Modes

Safe Walk

Fighting

Harvesting

Mining

Sentry

Self Sustainability

Auto Feeder

Auto Unstuck

Machine Learning

Four Detectable Enemies

  • Creepr
  • Zombie
  • Spider
  • Skeleton

Two Detectable Trees

  • Oak
  • Birch

Seven Detectable Ores

  • Coal
  • Diamond
  • Redstone
  • Gold
  • Iron
  • Lapis
  • Redstone

Limitations

  • Poor detection at close and at longer ranges
  • Easily looses track of enemies & other detections
  • Cannot fight players
  • Cannot fight most enemies
  • Cannot harvest most trees
  • Realistically can fight 1-2 enemies safely

Tools

  • Visaul Studio 2019
  • Cuda 11.2
  • Cudnn 8.8
  • OpenCV 4.5.5

System Specs

  • i7 10700k
  • RTX 3070
  • 32 GB DDR4 Ram @ 3200MHz
  • M.2 SSD

About

An auto-controller for Minecraft utilizing OpenCV. Features modes like, avoidance, running, harvesting, mining, and guarding using trained yolov5s models.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Languages