KGA Irrigation Toolsដំណើរការជាបាច់

DEM Elevation Correction and Point Sample Report

កែតម្រូវ DEM មួយ ឬច្រើន ធៀបនឹង DEM យោងដែលអ្នកជ្រើស ហើយចេញរបាយការណ៍គំរូចំណុច។

រកឃើញនៅឯណា
Processing ▸ Toolbox ▸ KGA Toolbox ▸ KGA Irrigation Tools ▸ DEM Elevation Correction and Point Sample Report
លេខសម្គាល់ algorithm
kga:demelevationcorrectionandpointsamplereport
ប្រភេទឧបករណ៍
Batch algorithm

ទិដ្ឋភាពទូទៅ

Corrects one or more DEMs against a Reference DEM you choose — your LiDAR, or whatever survey you trust — optionally merges them all into one seamless raster, and builds a Statistics tab and a Visualization tab summarising correction quality and post-correction agreement.

This is the tool for a project whose elevation data arrived from several sources at different dates and different accuracies: a LiDAR block, a drone survey, an older contour-derived grid. Each has its own vertical bias, and stitching them untreated gives steps at every seam.

The Reference DEM is never modified — it is the elevation truth every other DEM is corrected toward. The other DEMs are auto-sorted from finest to coarsest resolution and chained: the finest is corrected directly against the Reference, the next-finest against that already-corrected result, and so on.

The run is resumable via a checkpoint.json in the output folder.

ប៉ារ៉ាម៉ែត្រ

Parameters for DEM Elevation Correction and Point Sample Report
ឈ្មោះប្រភេទតម្លៃលំនាំដើមការពិពណ៌នា
Reference DEM (the elevation truth -- never modified)Raster layer
Reference DEM NoData override (blank = use its own NoData)String, optionalempty
DEMs to correct against the reference (1 or more)Multiple raster layersAuto-sorted finest to coarsest and corrected in a chain.
NoData overrides (comma-separated, matching DEM order)String, optionalemptyOne value per DEM, in the order you selected them. Blank uses each raster's own NoData metadata.
Correction method (all stages)EnumAuto - Compare All (Best CV)Auto - Compare All (Best CV), Global Offset, Polynomial Trend (degree 1), Polynomial Trend (degree 2), IDW (distance-decay). Auto tries them all and keeps whichever cross-validates best.
IDW powerNumber2.0IDW method only.
IDW neighbors (k)Integer12IDW method only.
IDW full-strength distance (m)Number3000.0Beyond this the correction fades out.
Cross-validation foldsInteger5How the method comparison is scored.
Calibration subsample strideInteger4Every Nth pixel is used to fit the correction. Raise it for a faster, coarser fit.
Merge finest DEM + all corrected DEMsBooleanTrue
Merge extentEnumReference DEM extentOr Coarsest DEM extent (full).
Merge resolution (m)Number2.5
Report sample stride (for Statistics/Visualization tabs)Integer (min 1)20
Report max sample pointsInteger (min 100)200000
Force re-run everything (ignore checkpoint.json)BooleanFalse
Output folderFolder destination

របៀបប្រើ

  1. Load the reference DEM and every DEM to correct.
  2. Open KGA Irrigation Tools > DEM Elevation Correction and Point Sample Report.
  3. Set the Reference DEM — the one you trust. It is never written to.
  4. Select the DEMs to correct. Their order does not matter; they are sorted by resolution.
  5. Leave the method on Auto for the first run and let the cross-validation pick. Look at what it chose in the Statistics tab before pinning a method.
  6. Press Run.
  7. Open the Statistics and Visualization tabs in the Processing Results panel — that is where the answer is, not in the log.
  8. If the run is interrupted, run it again with the same settings to resume.

លទ្ធផល

លទ្ធផលដែលប្រកាស

Outputs declared for DEM Elevation Correction and Point Sample Report
ឈ្មោះប្រភេទការពិពណ៌នា
Merged DEM (if merge was enabled)File
Run summaryString
StatisticsHTMLAppears as a tab in the Processing Results panel.
VisualizationHTMLLikewise.

អ្វីដែលអ្នកទទួលបាន

In the output folder:

  • corrected_{name}.tif — one corrected DEM per input.
  • MERGED_FINAL_DEM.tif — the seamless merge, when merging is enabled.
  • statistics_report.html and visualization_report.html — the two report tabs, also on disk.
  • correction_report.json — the machine-readable run record.
  • checkpoint.json — resume state. Safe to delete when finished.

កំណត់សម្គាល់ និងដែនកំណត់

  • The chain matters. Each DEM is corrected against the already-corrected one above it, so an error early in the chain propagates down. Check the Statistics tab stage by stage, not just at the end.
  • Correction does not create accuracy. It removes systematic bias relative to the reference; it cannot fix a DEM that is simply wrong in detail.
  • The reports are sampled, not exhaustive. Report sample stride and Report max sample points bound the cost; the statistics describe that sample.
  • Resumable via checkpoint.json. Tick Force re-run to ignore it.
  • NoData overrides must match the order you selected the DEMs in. Getting the order wrong silently corrects against the wrong nodata mask.
  • For a per-point table across several DEMs, use DEM Point Sample Export to Excel, which is the same sampling in spreadsheet form.

មើលផងដែរ

ត្រឡប់ទៅឯកសារណែនាំ