aelevine99/me500_vision

ME500_VISION

Welcome to our READ ME for our FFF Machine Vision Error Detection Project! This is the place where you can find all of the information on our files that we've developed over the past couple of months to get up to speed. Overall, we hope this application can detect, pause or stop a failed print, and notify the user of the failure by whatever means necessary.

Documentation

Roboflow Models

Roboflow “FDM Failures” Model

Deployment Instructions

File Organization and Helpful Links

Four folders show our journey through this project: References, Roboflow_Integration, Raspberry Pi Integration, and Email_Integration. The References folder gives the scripts from an OpenCV Bootcamp that is linked here. Going through this was pretty helpful when it came to relearning Python and getting up to speed with OpenCV's image-processing capabilities even though it is used pretty sparsely in our scripts. The Roboflow Integration folder contains the scripts that we used to familiarize ourselves with the Roboflow API and how we integrated it into our application. You should be provided links and files to the models we have trained along with the annotated images along with them. Here are some links that we've found helpful and are a good introduction to the platform as well:

Deploy your Roboflow Train models to Raspberry Pi.

Roboflow Documentation

Next, we used a Raspberry Pi 4 to run our application. To use your time efficiently, ensure that you are familiar with how code is executed on a Raspberry Pi (i.e. using virtual environments, how to install necessary libraries, etc.)

Create a Python virtual environment.

Using a virtual environment within a folder in your directory makes sure that all of the libraries you install for a project stay organized. This is good practice across the board if you're new to writing software.

To transfer files between our computers and the Raspberry Pi remotely, we used FileZilla over an SSH connection. You can download that here. We also used Putty to run scripts remotely on the Raspberry Pi over an SSH connection. Download that here.

Also, here are some YouTube videos that helped us immensely in understanding how to work with computer vision models on the RasPi:

Beginner Tutorial: How to Stream Video from Raspberry Pi Camera to Computer using Python (P2)

How to Deploy Computer Vision Models to Raspberry Pi with Docker

Lastly, we have Email Integration. In this folder, we used MailJet API (documentation linked here) to send and receive emails. We attempted to use the GMail API but ran into several issues when it came to authentication (most notably in IntegrationTest.py) However if you want to learn more about or use the GMail API authentication process, see the link here. Ideally, the user would receive an email with an attachment including the annotated image of their failed print. However, what we have currently sends an email with the necessary objects (To, From, Body).

If any questions arise or you require access to anything, feel free to reach out. Our emails are [email protected] and [email protected].