xView Detection Challenge is a satellite imagery project that teaches a computer to identify and locate objects like vehicles, buildings, and structures in photos taken from space. The project provides tools to download satellite images, train a detection model, and then use that trained model to automatically find objects in new overhead photos. The author is experimenting with different approaches to find the best balance between accuracy and how quickly the computer can learn.
How It Works
You learn that computers can look at photos taken from space and find objects like cars, buildings, and boats automatically.
You get the xView dataset containing over 1,000 satellite photos with different locations and objects to study.
You adjust a simple settings file to tell the program where your photos are stored on your computer.
The program converts your satellite photos into a format that the computer can use to learn what to look for.
Your computer studies hundreds of examples until it learns to recognize objects on its own, which takes some time.
Your trained model draws boxes around the objects it finds in new satellite photos, and you get a score showing how well it worked.
Star Growth
Repurpose is a Pro feature
Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.
Unlock RepurposeSimilar repos coming soon.