Robotic Arm with Computer Vision

Overview

Robotic systems are used in consumer and commercial applications to great benefit, resulting in accurate automation and cost efficiency. Some of these require computer vision, while others may be static in their implementation.

Objectives

This project required replicating the pick and place operation with the use of computer vision and
writing three labs that teach the coding language and concepts to underclassmen, and guiding seniors
through a lab to duplicate the project.

Approach

Corresponded with sponsor to acquire design and lab expectations
• Conducted research on similar implementations to gain insight
• Attempted to utilize a Kinect camera for depth operation benefits, but was met with operating
system incompatibilities.
• Decided upon a one camera setup as opposed to two for budgetary reasons
• Created a four module design with image processing, distance calculation, coordinate
transformation and direction sending
• Began coding the four modules, later adding a calibration module
• Created model for and printed a 3-D mount system to ensure the consistency of certain
parameters
• Finished implementation of all modules
• Testing resulted in the exposure of bugs and other issues
• Bugs and errors were fixed in all but one module

Outcomes

The project was not completed due to
unknown errors in one single module,
which handled coordinate transformations.
Three of the four labs have been written,
as the fourth lab depends on finishing the
project. Despite this shortcoming, the
project currently allows for a solid
foundation for the education of incoming
freshman on the topics of computer vision,
robotics, and programming.