GEO-CEOS stage 4 validation of the Satellite Image Automatic Mapper lightweight computer program for ESA Earth...

Area of Science:

• Remote Sensing
• Geospatial Analysis
• Computer Vision

Background:

• Earth Observation (EO) Level 2 products are crucial but not systematically generated.
• A gap exists in processing big EO data into validated Level 2 products.
• Existing methods require harmonization between spectral color names and land cover classes.

Purpose of the Study:

• To validate the Satellite Image Automatic Mapper (SIAM) for generating ESA EO Level 2 products.
• To address the information gap from EO big data to ESA EO Level 2 product.
• To comply with GEO-CEOS stage 4 validation guidelines.

Main Methods:

• Validated an off-the-shelf SIAM program on a 30m resolution Landsat time-series (2006-2009).
• Employed a prior knowledge-based decision tree for multi-spectral reflectance space hyperpolyhedralization.
• Developed an original protocol for wall-to-wall thematic map quality assessment without sampling.

Main Results:

• SIAM was validated against GEO-CEOS stage 4 guidelines.
• Annual 30m resolution land cover maps were generated for the conterminous U.S.
• The validation protocol allowed comparison between maps with differing legends.

Conclusions:

• The SIAM program effectively bridges the gap in generating validated EO Level 2 products.
• The developed validation protocol is suitable for large-scale thematic map quality assessment.
• Harmonization of color names and land cover classes is essential for accurate mapping.