Datasets:

Modalities:
Image
Languages:
English
Tags:
DOI:
License:
CEPhillips Muthukumaran commited on
Commit
f29f6c9
1 Parent(s): 6274300

impact-comms-update (#5)

Browse files

- impact-comms-update (a27e0935b3cd13bdc71dc5b6febb4a2f142e1a22)


Co-authored-by: Ramasubramanian <Muthukumaran@users.noreply.huggingface.co>

Files changed (1) hide show
  1. README.md +5 -5
README.md CHANGED
@@ -15,13 +15,13 @@ language:
15
 
16
  ### Dataset Summary
17
 
18
- This dataset contains Harmonized Landsat-Sentinel imagery of burn scars and the associated masks for the years 2018-2021 over the contiguous United States. There are 804 512x512 scenes. It's primary purpose is for training geospatial machine learning models.
19
 
20
  ## Dataset Structure
21
 
22
 
23
  ## TIFF Metadata
24
- Each tiff file contains a 512x512 pixel tiff file. Scenes have 6 bands, and masks have 1 band. For satellite scenes, each band has already been converted to reflectance.
25
 
26
  ## Band Order
27
  For scenes:
@@ -35,12 +35,12 @@ Channel, Name, HLS S30 Band number
35
 
36
  Masks are a single band with values:
37
  1 = Burn scar
38
- 0 = Not-burned
39
  -1 = Missing data
40
 
41
  ## Class Distribution
42
  Burn Scar - 11%
43
- Not-burned - 88%
44
  No Data - 1%
45
 
46
  ## Data Splits
@@ -48,7 +48,7 @@ The 804 files have been randomly split into training (2/3) and validation (1/3)
48
 
49
  ## Dataset Creation
50
 
51
- After co-locating the shapefile and HLS scene, the 512x512 chip was formed by taking a window with the burn scar in the center. Burn scars near the edges of HLS tiles are offset from center.
52
  Images were manually filtered for cloud cover and missing data to provide as clean a scene as possible, and burn scar presence was also manually verified.
53
 
54
  ## Source Data
 
15
 
16
  ### Dataset Summary
17
 
18
+ This dataset contains Harmonized Landsat and Sentinel-2 imagery of burn scars and the associated masks for the years 2018-2021 over the contiguous United States. There are 804 512x512 scenes. Its primary purpose is for training geospatial machine learning models.
19
 
20
  ## Dataset Structure
21
 
22
 
23
  ## TIFF Metadata
24
+ Each tiff file contains a 512x512 pixel tiff file. Scenes contain six bands, and masks have one band. For satellite scenes, each band has already been converted to reflectance.
25
 
26
  ## Band Order
27
  For scenes:
 
35
 
36
  Masks are a single band with values:
37
  1 = Burn scar
38
+ 0 = Not burned
39
  -1 = Missing data
40
 
41
  ## Class Distribution
42
  Burn Scar - 11%
43
+ Not burned - 88%
44
  No Data - 1%
45
 
46
  ## Data Splits
 
48
 
49
  ## Dataset Creation
50
 
51
+ After co-locating the shapefile and HLS scene, the 512x512 chip was formed by taking a window with the burn scar in the center. Burn scars near the edges of HLS tiles are offset from the center.
52
  Images were manually filtered for cloud cover and missing data to provide as clean a scene as possible, and burn scar presence was also manually verified.
53
 
54
  ## Source Data