Last updated: 2020-09-08
Checks: 7 0
Knit directory: baumarten/analysis/
This reproducible R Markdown analysis was created with workflowr (version 1.6.2). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.
Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.
Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.
The command set.seed(20200723) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.
Great job! Recording the operating system, R version, and package versions is critical for reproducibility.
Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.
Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.
Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.
The results in this page were generated with repository version 30d98a7. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.
Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:
Ignored files:
Ignored: .Rhistory
Ignored: .Rproj.user/
Ignored: analysis/.Rhistory
Ignored: data/sen2/
Unstaged changes:
Deleted: analysis/probability.Rmd
Modified: code/workflow_project_setup.R
Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.
These are the previous versions of the repository in which changes were made to the R Markdown (analysis/overview.Rmd) and HTML (docs/overview.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.
| File | Version | Author | Date | Message |
|---|---|---|---|---|
| Rmd | 30d98a7 | wiesehahn | 2020-09-08 | Change filenames, add analysis |
Overview
Random Forest evaluation
(classification steps and improvement possibilities)
Input Data
- Choice of satellite imagery (date, quality)
- Preprocessing steps (atmospheric, topographic, … correction)
- Image masking (clouds, cloud shadows, nodata)
Reference Data
- Choice of reference data (distribution)
- Label reference data
- Extraction of input data for reference areas
- Splitting in train and test data sets
Build Classification Model
- Choice of classification features
- Choice of model hyperparameters
Apply Classification Model
- Input data classification
- Class probability calculation
Postprocessing
- Apply filter (Minimal Mapping Unit, Focal filter)
- Integration of class probabilities
Accuracy Assessment
- Calculate accuracy metrics from validation data
Misclassifications
Masking
Tracks, Ways, ect.
Forest tracks, ways and official streets cut through the forest. Along these lines the spectral reflectances are heavily influenced by the surface material. Hence, the classification certainty is much lower along these features and tree species might be classified wrong.
Sentinel-2
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/sol_s2.jpg")
Preprocessed Sentinel-2 RGB image in the Solling area, which was used for tree species predictions
Probability
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/sol_probability_osm.jpg")
Classification probability in the Solling area (of tree species with highest probability per pixel)
Classification
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/sol_species.jpg")
Tree species classification result in the Solling area
Unstocked forest
In many areas the forest stocking changed in recent years due to drought, wind, bark beetle and other stressors. In these areas where healthy tree cover has been lost (and in these areas where it has never been) the reflectances are totally different from reflectances in healthy forest conditions. Hence, the classification certainty is much lower in these areas and tree species might be classified wrong.
Sentinel-2
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/har_s2.jpg")
Preprocessed Sentinel-2 RGB image in the Harz area, which was used for tree species predictions
Probability
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/har_probability.jpg")
Classification probability in the Harz area (of tree species with highest probability per pixel)
Classification
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/har_species.jpg")
Tree species classification result in the Harz area
Atmospheric disturbances
Atmospheric conditions, namely haze, clouds, contrails and other aerosols, have a strong influence on the reflectance values. In areas where they are not correctly masked or corrected the classification certainty is much lower and tree species might be classified wrong.
Sentinel-2
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/goe_s2.jpg")
Preprocessed Sentinel-2 RGB image around Goettingen, which was used for tree species predictions
Probability
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/goe_probability.jpg")
Classification probability around Goettingen (of tree species with highest probability per pixel)
Classification
knitr::include_graphics("https://raw.githubusercontent.com/wiesehahn/baumarten/master/docs/figure/evaluation.Rmd/goe_species.jpg")
Tree species classification result around Goettingen
sessionInfo()R version 4.0.2 (2020-06-22)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18362)
Matrix products: default
locale:
[1] LC_COLLATE=German_Germany.1252 LC_CTYPE=German_Germany.1252
[3] LC_MONETARY=German_Germany.1252 LC_NUMERIC=C
[5] LC_TIME=German_Germany.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] workflowr_1.6.2
loaded via a namespace (and not attached):
[1] Rcpp_1.0.5 rstudioapi_0.11 whisker_0.4 knitr_1.29
[5] magrittr_1.5 R6_2.4.1 rlang_0.4.7 highr_0.8
[9] stringr_1.4.0 tools_4.0.2 xfun_0.15 git2r_0.27.1
[13] htmltools_0.5.0 ellipsis_0.3.1 rprojroot_1.3-2 yaml_2.2.1
[17] digest_0.6.25 tibble_3.0.3 lifecycle_0.2.0 crayon_1.3.4
[21] later_1.1.0.1 vctrs_0.3.2 promises_1.1.1 fs_1.4.2
[25] glue_1.4.1 evaluate_0.14 rmarkdown_2.3 stringi_1.4.6
[29] compiler_4.0.2 pillar_1.4.6 backports_1.1.7 httpuv_1.5.4
[33] pkgconfig_2.0.3