Crop Mapping Using Multispectral Sentinel-2 Dataset Assignment

Assignment Task

Project Topic

Crop Mapping With Sentinel-2 Multispectral Data Using Cnn-Transformer Model  

The problem is more like a segmentation problem. where a multispectral land image taken by sentinel-2 is given as input and the output gives what all the crops types are mapped to the given land (with segmentation)

ARCHITECTURE

The following architecture is the classification scheme for multitemporal multisensor images based on CNN-transformer sequence correlation extraction. The lower part of the figure indicates multitemporal feature extraction and position feature embeddings; and the upper part expresses the sequence feature extraction with multilayer encoder modules.

This method utilizes the final output sequence of the multilayer encoder module to extract the category feature. The outputs of different encoder layers can express the sequence correlation of different levels.

This Engineering has been solved by our PHD Experts at My Uni Paper.