In this Project we going to use U-net Neural network architecture which is the best NN structure for the purpose of Image segmentation. This NN will be fed with an Artificial Lunar Landscape dataset, to segment the lunar surface to better predict the route a lunar rover should take to avoid system failure or accidents. This objective is very crucial for a space rover as it is a sensitive and expensive device on foreign space land, there is no possibility of repairs, so the best course of action for anyone in this situation is to avoid the failure in the first place. So for a rover, “ Precaution is better than cure”, is the only option.
This is a Deep learning - computer vision project, where I have used techniques like Transfer Learning, Callbacks, U-net model building and optimization from Vgg-16 Base as a backbone, Technical Documentation and plotting are done using matplotlib.
I hope everyone gains value from this project, Happy Learning.