fine tune a pre-trained CNN model (VGG16) to work on the land cover classification task.
can be considered as endorcment to work done in the paper :
- Training Deep Convolutional Neural Networks for Land–Cover Classification of High-Resolution Imagery Grant J. Scott, Member, IEEE, Matthew R. England, William A. Starms, Student Member, IEEE, Richard A. Marcum, and Curt H. Davis, Fellow, IEEE
References:
- inspired by and used some code snippets from : https://github.com/udacity/deep-learning-v2-pytorch/
- dataset used is UC Merced(UCM) Land Use at : http://weegee.vision.ucmerced.edu/datasets/landuse.html