I’ve worked through Chapters 1-6 of Fastbook. I’m all set to do a computer vision project now…
Whilst I’m working through the rest of the book, I’m trying to formulate a real research project and I was hoping for some guidance.
If I have aerial photographs of the UK and I want to identify a certain agricultural feature from them to see how many there are in each region, I would first create a database of the relevant images by cropping them, saving them in a folder and labelling them, right?
How do would you go about quantifying this and avoiding the same feature from being calculated multiple times? i.e. if the aerial photographs are a set of repeated images, surely the model would keep finding the same feature multiple times?
Grateful for any thoughts.