0x365's picture
Upload README.md
2e180a9 verified
metadata
dataset_info:
  features:
    - name: id
      dtype: string
    - name: B02
      dtype: image
    - name: B03
      dtype: image
    - name: B04
      dtype: image
    - name: B05
      dtype: image
    - name: B06
      dtype: image
    - name: B07
      dtype: image
    - name: B08
      dtype: image
    - name: B11
      dtype: image
    - name: B12
      dtype: image
    - name: label
      dtype: image
    - name: lon_min
      dtype: float32
    - name: lon_max
      dtype: float32
    - name: lat_min
      dtype: float32
    - name: lat_max
      dtype: float32
    - name: unclassified_space
      dtype: float32
configs:
  - config_name: default
license: cc-by-4.0
task_categories:
  - image-segmentation
language:
  - en
tags:
  - image
  - satellite
  - eo
  - crop
  - crop-type
  - sentinel
  - sentinel-2
  - belgium
pretty_name: Satellite Earth Observation Crop Type Segmentation
size_categories:
  - 1K<n<10K

Crop Type Segmentation Earth Observation Dataset of Belgium

This dataset is a ML ready dataset with Earth Observation data from ESA Sentinel-2 with a GSD of 10m. The segmentation labels are given for the crop-type in each 10m pixel. The spectral bands provided are those suggest in ESA WorldCereal documentation (Bands: 2,3,4,5,6,7,8,11,12). The dataset covers the majority of Belgium. Each image is 256x256 pixels.

  • Author of this dataset: Robert Cowlishaw (0x365)

Input data - Sentinel-2 Satellite Imagery from ESA Copernicus

  • Bands: 2, 3, 4, 5, 6, 7, 8, 11, 12
  • GSD: 10m
  • Image Dimensions: 256x256 pixels
  • Date of acquisition: 29/06/2018 to 01/07/2018
  • Location: Belgium

Label data - ESA WorldCereal dataset originally created by Department of Agriculture and Fisheries of the Flemish government (Belgium)

  • GSD: 10m
  • Image Dimensions: 256x256 pixelsB02
  • Date of observation: 01/06/2018
  • Location: Belgium

Licensing Information

The dataset is available under the Creative Commons License (CC BY 4.0). Please read further information about the datasets used to build this one to correctly credit the data creators here.

  • Contains modified Copernicus Sentinel data 2018
  • Contains modified Copernicus Service information 2018

Load dataset

from datasets import load_dataset

dataset = load_dataset("0x365/eo-crop-type-belgium", split="train")

Labels

Class ID Crop Type Feature Count
255 Background/Unclassified 208288074
0 pasture 68096739
1 maize 50068391
2 winter wheat 16664081
3 potatoes 14379282
4 fodder beets 6565216
5 leguminous cover crop 4502781
6 winter barley 3933897
7 pears 2582060
8 appels 1488258
9 beans 1417765
10 # engelwortel 1255222
11 hemp 1067147
12 cauliflower 1043146
13 fodder carrots 1009874
14 other crops 1001480
15 brussels sprouts 912454
16 triticale 831751
17 sjalots 827927
18 leek 780744
19 chicory 748404
20 peas 705120
21 fodder mixture 625563
22 spinach 528456
23 parsnip 301429
24 strawberries 291863
25 cherry 289896
26 spring barley 244321
27 vegetables - seeds 243480
28 cucumber herb 216599
29 zucchini 181351
30 winter rapeseed 157622
31 asparagus 146538
32 spring wheat 132405
33 fodder cabbages 128322
34 butternut 113114
35 sword herd 88687
36 haricot 87384
37 winter rye 85854
38 grape seedlings 59326
39 ornamental plants 56964
40 other annual fruit 53807
41 buckwheat 45499
42 fodder turnips 42663
43 tagetes 37126
44 blueberries 27031
45 fennel 26752
46 red berries 21415
47 raspberry 17584
48 other fodder crops 15525
49 aromatic herbs 13340
50 sorghum 13042
51 soybeans 12271
52 yellow mustard 11877
53 tobacco 11793
54 walnuts 9260
55 barley 8821
56 spring rapeseed 7944
57 kiwi berries 7369
58 spring rye 6328
59 plums 6090
60 chervil 5653
61 phacelia 4977
62 gooseberries 2909
63 hazelnuts 1564
64 sunflower 1013
65 mixture winter wheat and leguminous crops 0
66 sweet potatoes 0
67 nut-sweet rapeseed 0
68 nyger seed 0
69 lupine 0
70 peach 0
71 oil seeds 0
72 tomatoes 0
73 eggplant 0
74 cucumber (greenhouses) 0
75 oats 0
76 artichokes 0
77 sudan grass 0