Datasets:
Tasks:
Image Classification
Modalities:
Image
Formats:
parquet
Sub-tasks:
multi-class-classification
Languages:
English
Size:
10K - 100K
ArXiv:
License:
MuafiraThasni
commited on
Commit
•
ed2431f
1
Parent(s):
104b484
Update README.md
Browse filesUpdated dataset details
README.md
CHANGED
@@ -1,3 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
dataset_info:
|
3 |
features:
|
|
|
1 |
+
---
|
2 |
+
annotations_creators:
|
3 |
+
- other
|
4 |
+
language:
|
5 |
+
- en
|
6 |
+
license:
|
7 |
+
- other
|
8 |
+
multilinguality:
|
9 |
+
- no
|
10 |
+
size_categories:
|
11 |
+
- 10K<n<100K
|
12 |
+
source_datasets:
|
13 |
+
- original
|
14 |
+
task_categories:
|
15 |
+
- image-classification
|
16 |
+
task_ids:
|
17 |
+
- multi-class-classification
|
18 |
+
---
|
19 |
+
|
20 |
+
# EuroSAT Dataset
|
21 |
+
|
22 |
+
## Overview
|
23 |
+
|
24 |
+
This dataset contains satellite images from the EuroSAT datasetThe dataset consists of RGB images with 10 different classes, each representing a distinct type of land use.
|
25 |
+
|
26 |
+
## Dataset Summary
|
27 |
+
|
28 |
+
- **Classes**: 10 (e.g., Annual Crop, Forest, Herbaceous Vegetation, Highway, Industrial, Pasture, Permanent Crop, Residential, River, Sea/Lake)
|
29 |
+
- **Number of Images**: 27,000+ images split into training and validation sets
|
30 |
+
- **Image Size**: 64x64 pixels, 3 channels (RGB)
|
31 |
+
- **Data Augmentation**: Random flipping, random rotation, and normalization
|
32 |
+
|
33 |
+
## Usage
|
34 |
+
|
35 |
+
This dataset is ideal for training and fine-tuning image classification models, specifically for applications in remote sensing, urban planning, environmental monitoring, and agricultural management.
|
36 |
+
|
37 |
+
### Key Features
|
38 |
+
|
39 |
+
- **Preprocessing**: Images were preprocessed using TensorFlow, including resizing, normalizing, and applying augmentation techniques to improve model robustness.
|
40 |
+
- **Model**: Fine-tuned with Vision Transformer (ViT) to leverage attention-based mechanisms for superior performance in image classification tasks.
|
41 |
+
|
42 |
+
## Dataset Structure
|
43 |
+
|
44 |
+
- **Training Set**: Approximately 80% of the dataset
|
45 |
+
- **Validation Set**: Approximately 20% of the dataset
|
46 |
+
- **Format**: Images are stored as PNG files with associated labels.
|
47 |
+
|
48 |
+
### Labels
|
49 |
+
- Class labels are encoded as integers and correspond to the following categories:
|
50 |
+
0. Annual Crop
|
51 |
+
1. Forest
|
52 |
+
2. Herbaceous Vegetation
|
53 |
+
3. Highway
|
54 |
+
4. Industrial
|
55 |
+
5. Pasture
|
56 |
+
6. Permanent Crop
|
57 |
+
7. Residential
|
58 |
+
8. River
|
59 |
+
9. Sea/Lake
|
60 |
+
|
61 |
+
## Performance Metrics
|
62 |
+
|
63 |
+
- **Accuracy**: Achieved X% accuracy on the validation set
|
64 |
+
- **F1-Score**: Achieved X% F1-score across all classes
|
65 |
+
|
66 |
+
## Practical Applications
|
67 |
+
|
68 |
+
- **Urban Planning**: Identifying residential and industrial areas from satellite imagery.
|
69 |
+
- **Agriculture**: Monitoring crop types and assessing agricultural land use.
|
70 |
+
- **Environmental Protection**: Tracking changes in natural land covers like forests and rivers.
|
71 |
+
|
72 |
+
## How to Use
|
73 |
+
|
74 |
+
To use this dataset, you can directly load it through the Hugging Face `datasets` library:
|
75 |
+
|
76 |
+
```python
|
77 |
+
from datasets import load_dataset
|
78 |
+
|
79 |
+
dataset = load_dataset("your_username/eurosat-land-use-classification")
|
80 |
+
```
|
81 |
+
|
82 |
---
|
83 |
dataset_info:
|
84 |
features:
|