Coverage scene counts are available as a CSV file, as an HTML table or as an HTML page with bar charts. Principles used to calculate coverage metrics. Problem:
For Brazil and Mexico, ND area is smaller than the country area. Some agencies submit scenes covering only the ND areas, others provide scenes of the whole country.
Solution:
Provide scene counts for both ND and total country area if they differ.
Problem:
Certain images overlap ND boundaries - eg, one scene can lie in both Brazil and Guyana.
Solution:
Provide separate scene counts for each ND, as well as totals across all NDs.
Problem:
Optical images can have cloud cover that varies from 0 to 100.
Solution:
For the purposes of calculating area covered by each agency, scale the area of the optical images
in proportion to the amount of cloud cover. Eg, the area of a scene with 25% cloud cover is multiplied by 0.75.
Agencies could have several rounds of coverage of the same area within one year, so directly comparing just the area covered does not provide a fair comparison.
Problem:
The scene count is not an ideal metric either, because different instruments have different resolution.
Solution:
Calculate duplication percentage. Examples:
For two scenes that are taken at different times but have identical dimensions, the duplication metric would be 200%.
For two non-overlapping scenes duplication metric is 0%. For two scenes of area X that overlap by exactly half, duplication percentage is 2X / 1.5X = 133%. Click to return to ND browser overview. |
|
|
Coverage metrics |