ISSAP Module

Purpose

Chapman University MLAT

Location

Orange County, USA

Year

2021

We are still in the research portion of this project meaning that our GitHub repository is currently private. However, here are some snippets of the code and the way we are approaching the project.

For this project, it was our intent to train a convolutional neural network to identify the module location of pictures taken within the International Space Station.

Screen Shot 2021-12-06 at 10.25.08 AM.png

First, we prepare the training data and then prepare the data pipeline.

Next, we define the architecture. Here is a tiny snippit.

Screen Shot 2021-12-06 at 10.27.53 AM.png

At this stage, we had noticed that our CNN's accuracy was not where we wanted it. To combat this I created a python program to iterate through our training data, assess if the size of the image could be even sliced, and if so, slice the image into up to 9 more images that were then added to the training data.

Screen Shot 2021-12-06 at 10.40.56 AM.png

Here is a piece of that code. After this process, we had significantly more training data to work with.

We are currently in the research phase of this project as of December 2021. Once published, this page will be updated with the link to the manuscript.