Add glossary

docs/glossary.rst

@@ -0,0 +1,36 @@
+Glossary
+========
+
+A collection of common terms used in :mod:`torchgeo` that may be unfamiliar to either:
+
+1. Deep learning researchers who don't know remote sensing
+2. Remote sensing researchers who don't know deep learning
+
+.. glossary::
+
+   chip
+       Synonym for :term:`patch`. A smaller image sampled from a larger :term:`tile`.
+
+   classification
+       A computer vision task that involves predicting the image class for an entire image or a specific bounding box.
+
+   instance segmentation
+       A computer vision task that involves predicting labels for each pixel in an image such that each object has a unique label.
+
+   object detection
+       A computer vision task that involves predicting bounding boxes around each object in an image.
+
+   patch
+       Synonym for :term:`chip`. A smaller image sampled from a larger :term:`tile`.
+
+   regression
+       A computer vision task that involves predicting a real valued number based on an image.
+
+   semantic segmentation
+       A computer vision task that involves predicting labels for each pixel in an image such that each class has a unique label.
+
+   swath
+       A set of :term:`tiles <tile>` along a satellite trajectory.
+
+   tile
+       A single image file taken by a remote sensor like a satellite.

docs/index.rst

@@ -17,6 +17,12 @@ architectures, and common image transformations for geospatial data.
    samplers
    transforms
 
+.. toctree::
+   :maxdepth: 2
+   :caption: User Documentation
+
+   glossary
+
 .. toctree::
    :maxdepth: 1
    :caption: PyTorch Libraries
@@ -28,9 +34,3 @@ architectures, and common image transformations for geospatial data.
    TorchElastic <https://pytorch.org/elastic/>
    TorchServe <https://pytorch.org/serve>
    PyTorch on XLA Devices <http://pytorch.org/xla/>
-
-
-Indices
--------
-
-* :ref:`genindex`

torchgeo/samplers/samplers.py

@@ -33,7 +33,7 @@ class RandomGeoSampler(GeoSampler):
     """Samples elements from a region of interest randomly.
 
     This is particularly useful during training when you want to maximize the size of
-    the dataset and return as many random chips as possible.
+    the dataset and return as many random :term:`chips <chip>` as possible.
     """
 
 
@@ -42,7 +42,7 @@ class GridGeoSampler(GeoSampler):
 
     This is particularly useful during evaluation when you want to make predictions for
     an entire region of interest. You want to minimize the amount of redundant
-    computation by minimizing overlap between chips.
+    computation by minimizing overlap between :term:`chips <chip>`.
 
     Usually the stride should be slightly smaller than the chip size such that each chip
     has some small overlap with surrounding chips. This is used to prevent `stitching