Skip to content

Commit

Permalink
add validation of MLM output using classification:classes (fixes #48) (
Browse files Browse the repository at this point in the history
  • Loading branch information
fmigneault authored May 14, 2024
1 parent c5fbed3 commit bbe04d7
Show file tree
Hide file tree
Showing 4 changed files with 417 additions and 3 deletions.
4 changes: 4 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0

### Added

- Validation by the JSON schema of `classification:classes` definition within a
[Machine Learning Model (MLM)](https://github.com/crim-ca/mlm-extension) output.
- Example STAC `item-model-classes.json` demonstrating the use of `classification:classes` within a `mlm:output`.
- `nodata` to `classification:classes`
- `nodata` to Class Object
- `percentage` to Class Object
- `count` to Class Object
Expand Down
8 changes: 6 additions & 2 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -12,8 +12,10 @@ This document explains the Classification Extension to the

- Examples:
- Asset level:
- [Classes example](examples/item-classes-maxar.json): STAC Item with classified raster bands in asset (Maxar)
- [Raster Classes example](examples/item-classes-maxar.json): STAC Item with classified raster bands in asset (Maxar)
- [Bitfields example](examples/item-bitfields-landsat.json): STAC Item with bit fields in asset (Landsat)
- Item level:
- [Model Classes example](examples/item-model-classes.json): STAC Item with classes of a Machine Learning Model output
- Collection level:
- [Item Assets example](examples/collection-item-assets.json): STAC Collection using Item Assets for classed child Items
- [JSON Schema](json-schema/schema.json)
Expand Down Expand Up @@ -41,9 +43,11 @@ position of the first bit in a field is given by its offset. Therefore the first

These classification objects can be used in the following places:

- In a raster Asset object if single band.
- In a [raster Asset object](https://github.com/stac-extensions/raster#item-asset-fields) if single band.
- For multiband rasters, use [`raster:bands`](https://github.com/stac-extensions/raster) and store the classes in
each Band Object.
- In a [Machine Learning Model output object](https://github.com/crim-ca/mlm-extension#model-output-object) to
define the classes represented within that output.
- As an [`item-assets`](https://github.com/stac-extensions/item-assets) field in a Collection object, to indicate
that the classification is used across child Items.

Expand Down
Loading

0 comments on commit bbe04d7

Please sign in to comment.