Subject: Spacecraft image analysis using computer vision

​

Hi guys,

Im looking to develop a system that uses computer vision algorithms to analyze images captured by spacecraft cameras and identify potential safety hazards or security threats. For example, the system could detect debris or other objects in orbit that could pose a risk to spacecraft.

I am looking to do this using all AWS tools. I am pretty new to this and am developing a technology architecture project around this topic to present for a program I'm doing.

How would I go about approaching/doing this? I am looking to find/create my own mock datasets as well as present the alogrithm/code I used to train my model. More specifically, I am focusing on these aspects for my project:

Preprocess the images: Preprocess the images to improve their quality and prepare them for analysis. This could include cropping, resizing, and adjusting the brightness and contrast of the images.

Train the computer vision algorithms: Train the computer vision algorithms using the dataset of images. There are various computer vision techniques that could be used, such as object detection, segmentation, or classification. The specific technique will depend on the requirements of the system.

​

In addition, it would be cool to have some sort of hardware/interactive portion that actually utilizes a camera to detect things in space. That can be implemented into the system. Once the computer vision algorithms have been trained and evaluated, implement the system. This could involve integrating the algorithms into a larger software system that can process images captured by spacecraft cameras in real-time.

Thank you