Add Level-3 and Level-4 bands for Land Cover product. #161
Conversation
…rrestrial vegetation
…rrestrial vegetation
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
…t tests to dask arrays
for more information, see https://pre-commit.ci
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## develop #161 +/- ##
===========================================
+ Coverage 80.02% 82.38% +2.35%
===========================================
Files 49 67 +18
Lines 4361 4951 +590
===========================================
+ Hits 3490 4079 +589
- Misses 871 872 +1 ☔ View full report in Codecov by Sentry. |
| dtype="uint8", | ||
| **{ | ||
| "watseas_trh": water_seasonality_threshold, | ||
| "watersea_nodata": WATER_FREQ_NODATA, |
There was a problem hiding this comment.
the correct nodata=255 from the raster, also it won't work as water_season_mask has been converted to uint8 from last step, -999 was lost had it been the effective nodata.
There was a problem hiding this comment.
Good point, refactored the code to effectively replace "nodata"
| **{"_u": 210}, | ||
| ) | ||
|
|
||
| intertidal_mask = expr_eval( |
There was a problem hiding this comment.
This mask is used in level4 water classification.
There was a problem hiding this comment.
I mean it's redundant in a sense that l3 and l4 doesn't have to be generated in that order. The water persistence and consequently l4 water classification can happen well ahead of l3 aggregation.
odc-stats/odc/stats/plugins/lc_level34.py
Line 92 in 40ab398
odc-stats/odc/stats/plugins/lc_level34.py
Line 107 in 40ab398
There was a problem hiding this comment.
This is addressed now.
| def bare_gradation(xx: xr.Dataset, bare_threshold, veg_cover): | ||
|
|
||
| # Now add the bare gradation | ||
| fcp_nodaata = -999 |
There was a problem hiding this comment.
is it a typo? I don't see anywhere changing the fcp nodata value.
There was a problem hiding this comment.
I've removed the redundant statement.
| ) | ||
| # Apply bare gradation expected output classes | ||
| # Map bare gradation classes | ||
| bs_mapping = {100: 10, 120: 12, 150: 15} |
There was a problem hiding this comment.
it might be easier to allocate a new array instead of the extra mapping, also reduce the confusion. It's bit hard to follow that it's essentially to avoid the mixing of small and big values in the original array.
There was a problem hiding this comment.
Agree. I've modified the code and deleted the mapping here and everywhere else.
|
|
||
| # Define mapping from current output to expected a3 output | ||
| # Map vegetation cover classes | ||
| veg_mapping = {160: 16, 150: 15, 130: 13, 120: 12, 100: 10} |
There was a problem hiding this comment.
Deleted the mapping.
|
|
||
| # Apply water persistence expcted classes | ||
| # Map values to the classes expected in water persistence in land cover Level-4 output | ||
| waterper_wat_mapping = {100: 1, 70: 7, 80: 8, 90: 9} |
There was a problem hiding this comment.
Deleted the mapping.
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
This change will create the level4 and level3 land cover bands.
The fix_l3 will be applied on top of this PR which will fix the nodata issue. This fix will be applied to Level4 as well.
Links to definitions and thresholds used to generate level-4 bands:
https://bitbucket.org/au-eoed/livingearth_lccs/src/1e78111ead2feb1179bf9fb30f6e8744aed65886/colour_schemes/lccs_l4_colour_scheme_golden_dark_au.csv
https://knowledge.dea.ga.gov.au/data/product/dea-land-cover-landsat/?tab=details