Skip to content

giving Baxter-Robot Face/Object Detection capabilities and basic object grasping

License

Notifications You must be signed in to change notification settings

igor-lirussi/Baxter-Robot-ObjDet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

56 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Baxter-Robot-ObjDet

Repository for some experiments with Baxter robot and Object Detection.

The code allows the robot:

  • to detect different objects,
  • approach them with his arm, independently from their position on the table in front of it,
  • pick them up safely and move them.

The models of large size in the folder models can be downloaded from the releases

Result

Baxter Video

Run

Please clone this repo with "recurse-submodules" option, or the submodules will be missing:

git clone --recurse-submodules git@github.com:igor-lirussi/Baxter-Robot-ObjDet.git

run python reach_obj_det_YoloV4.py --help for info on executing it.

Jupyter Notebooks

The notebooks contain some examples of image processing with OpenCV and Pyhton.

The notebooks study both face detection and object-detection.

If it's not in the right path, open the notebook with Anaconda Propt passing the correct disk. Example:

jupyter notebook --notebook-dir=D:

Colors Toys Object Detection Model

A Yolo v4 model has been trained to detect 9 different classes of food toys present in CoLoRs Lab.

In the releases you can download:

  • colors.names
  • yolov4-colors.cfg
  • yolov4-colors.weights

dataset

Built With

  • Python 3.7.18

Changelog

Version 1.0 - Initial release, 31-01-2022

Version 1.1 - Added Colors yolov4 trained net, 28-09-2023

Authors

  • Igor Lirussi @ BOUN Boğaziçi University - CoLoRs Lab

License

This project is licensed under License - see the LICENSE file for details