Chesapeake Land Cover dataset - Learning on the prior extension

<p>This dataset extends the &quot;Chesapeake Land Cover&quot; dataset at <a href="https://lila.science/datasets/chesapeakelandcover">https://lila.science/datasets/chesapeakelandcover</a> with an additional <em>layer </em>of data containing a prior of observing a 1m Chesapeake Conservancy land cover class label given an NLCD label. Specifically, for each <em>tile </em>in the original &quot;Chesapeake Land Cover&quot; dataset, this dataset contains another tile (named with a &quot;_prior_from_cooccurrences_101_31_no_osm_no_buildings.tif&quot; suffix) containing the prior probabilities of observing a four class version of the land cover&nbsp;classes at a 1m resolution. The prior probability is given by the normalized co-occurrence matrix between NLCD classes and the Chesapeake 1m land cover labels in each state with additional spatial smoothing (using a gaussian filter with a standard deviation of 31 pixels and a cutoff of 101 pixels) to reduce the block artifacts caused by the relatively low-resolution of the NLCD labels (30m) compared to the LC labels (1m). <strong>Note:</strong>&nbsp;the prior is a 4-class mapping of the 6-class labels. In the 4-class version of the dataset&nbsp;the &quot;barren land&quot;, &quot;impervious (other)&quot;, and &quot;impervious (road)&quot; classes are combined.</p> <p>This dataset also includes the per-state co-occurrence matrices. Each of these is a matrix, <span class="math-tex">\(C\)</span>, with size 7 x 17, where an entry <span class="math-tex">\(C_{ij}\)</span> gives the normalized count of land cover class <span class="math-tex">\(i\)</span> for given NLCD label <span class="math-tex">\(j\)</span>. The class indices <span class="math-tex">\(i\)</span> and <span class="math-tex">\(j\)</span> correspond to the indices used in the Chesapeake Conservancy land cover label and NLCD&nbsp;layers.</p> <p>The dataset is packaged in a way such that it can simply be unzipped over an existing copy of the &quot;Chesapeake Land Cover&quot; (i.e. follows the same directory structure). E.g., given a directory that contains the original dataset &quot;cvpr_chesapeake_landcover/&quot;, and the file &quot;cvpr_chesapeake_landcover_prior_extension.zip&quot;, run `unzip&nbsp;cvpr_chesapeake_landcover_prior_extension.zip` to create the merged dataset.</p> <p>&nbsp;</p> <p>Version 1.1 fixes an issue with the geotiff metadata that caused GDAL to interpret the 4th band as a nodata mask in some settings.</p>