asimniazi63 / crop-type-narc

Paper: "Crop Type Classification using Multi-temporal Sentinel-2 Satellite Imagery: A Deep Semantic Segmentation Approach"

Crop Type Classification

This Repository contains code and trained models (will be published after paper acceptance) and benchmark results for the paper:

Crop Type Classification using Multi-temporal Sentinel-2 Satellite Imagery: A Deep Semantic Segmentation Approach

Submitted at: 5th International Conference on Robotics and Automation in Industry (ICRAI)

Segmentation Models

1. UNet
2. UNet3+
3. DeepLabv3+
4. Swim-UNet
5. TransUNet

Pre-requistes

Main packages required are:

• TensorFlow v2.9.2
• Keras v2.9.0
• rasterio
• numpy
• matplotlib
• albumentations

Datasets

We used Google Earth Engine to generate the Dataset collected from National Agriculture Research Center, Islamabad. Crop categories includes:

1. Fodder
2. Oilseeds
3. Pulses
4. Wheat
5. MSM (Millet, Sorghum, Maize)
6. Others

Visual Results

Some Visual results are following:

Quantitative Results

Performance Graphs

Loss: Different models and different band combinations

Maintainer

Asim Hameed Khan ([e-mail])