Skip to content

roncanciovl/objectclassification

Repository files navigation

Object Classification on Robotic Application

Garbage Classification and selection using robotic arm

Table of Contents

Quickstart

In order for this code to work you will need to install a few packages

pip install -r requirements.txt

The Non Quickstart

Recomended instructions for a person new to Python and git hub on how to install the packages, clone the repository, and run the code: deployment in (Spanish)

For more information about how to deploy software, you may also check out tutorials

Source code

Modules

  • ai.py This is a module that classifies the content of the image as blottle, can, other, and no_object. It returns a category.
  • lookforObjects.py This is a module that looks for the center of any object in the image and their orientations. It returns a list of orientations, a list of centers and an image with visual info. If no objects, it returns empty lists.

Test files (to use as Scripts)

  • kinematics.py kinematics calculations and robot model
  • serialcomTest.py test on serial communications with pyserial
  • lookObjectsVideo.py lookforObjects module working on on frames from webcam's video
  • ai_video.py test on frames from webcam's video
  • test_camera.py Code to check out whether your camera is working

Files to be imported as additional source code

  • robotClasses.py This is a file with the Class Diagram for the robot use cases. Edited in www.genmymodel.com
  • robotStateMach.py This is a file with the Robot State Machine.
  • gui.py This is file with a Simple graphical user interface for the robot
  • training.py This is a file with the AI training code using Convolutional Neural Networks

For any other questions, solicitations email Henry Roncancio (initial author)