BigEarthNetDataLoaderOptions
Options & Configuration · AiDotNet.Data.Vision.Benchmarks
Configuration options for the BigEarthNet data loader.
How It Works
BigEarthNet is a large-scale multi-label remote sensing dataset consisting of 590,326 Sentinel-2 image patches from 10 European countries. Each patch is 120x120 pixels with 12 spectral bands. The multi-label classification task uses 19 CORINE Land Cover classes (BigEarthNet-19) or 43 original CLC classes.
Constructors
| Constructor | Summary |
|---|---|
BigEarthNetDataLoaderOptions() |
Properties
| Property | Summary |
|---|---|
DatasetSplit Split { get; set; } |
Dataset split to load. |
bool AutoDownload { get; set; } |
Automatically download if not present. |
bool Normalize { get; set; } |
Normalize pixel values to [0, 1]. |
bool Use19ClassScheme { get; set; } |
Use the simplified 19-class label scheme (BigEarthNet-19). |
int NumBands { get; set; } |
Number of spectral bands to load. |
int? MaxSamples { get; set; } |
Optional maximum number of samples to load. |
string DataPath { get; set; } |
Root data path. |