LifeWatch ERIC Metadata Catalogue
1st stage Data-driven Classifier
It is a Support Vector Machine, pixel-based, classifier trained for a multi-class problem.
It represents the Step 4 of the Ailanthus Workflow within the Internal Joint Initiative.
Default
Identification
- Date ( Creation )
- 2019-07-04
- Date ( Publication )
- 2021-04-02
- Status
- Under development / Pre operational
- Version
- 1.0
- Keywords
- SVM
- Keywords
- classifier
- Keywords
- data-driven
- Keywords
- supervised classification
- Keywords
- Ailanthus
- Keywords
- IJI
- Access constraints
- License
- Use limitation
- The final license will be available soon
- Protocol
- DOI
- Operation name
- Stack_4dates_L5_30m.tiff
- Description
- It is a multi-layer file containing the four multi-season images each one with its 6 bands. File format: raster geotiff.
- Function
- Input file
- Operation name
- Training_multiclass_singleshape_FAO.shp
- Description
- It represents the reference data for the training of the supervised multi-class data-driven classifier. There is a file for each class in the scene labelled according to FAO-LCCS taxonomy. A numeric code is associated to each FAO_LCCS label. File format: shapefile.
- Function
- Input file
- Operation name
- Test_multiclass_singleshape_FAO.shp
- Description
- It represents the reference data for the validation of the output map obtained from the supervised multi-class data-driven classifier. There is a file for each class in the scene labelled according to FAO-LCCS taxonomy. A numeric code is associated to each FAO_LCCS label. File format: shapefile.
- Function
- Input file
- Operation name
- ConfigurationParameter_SVMmulticlass.txt
- Description
- Parameters to be setted for the training of the SVM classifier: Kernel function; Regularization parameter; penalty parameter. File format: text file.
- Function
- Input file
- Operation name
- LC_map_numeric_multiclass_30m.tiff
- Description
- It represents the output multi-class land cover map obtained from the SVM classifier at 30 meters spatial resolution. It is a multi-class map with a numeric code associated to each different class. File format: raster geotiff.
- Function
- Output file
- Operation name
- FAO-numeric_multiclass.txt
- Description
- Look-up table to identify the correspondence between numeric and FAO-LCCS codes. File format: text file.
- Function
- Output file
- Operation name
- ConfMatr_multiclass.txt
- Description
- Confusion matrix related to the validation dataset. File format: text file.
- Function
- Output file
- Operation name
- Accuracy_Estimators_multiclass.txt
- Description
- Accuracy measures as Overall Accuracy; User's Accuracy; Producer's Accuracy; F1-score. File format: text file.
- Function
- Output file
- Topic Category
- Land cover
- Service Category
- data classification
- Service Language
- eng
- Service TRL
- TRL 9 – Actual system proven in operational environment
