diff --git a/.github/workflows/release.yaml b/.github/workflows/release.yaml
index ba35c1dffbe..83fb1a90588 100644
--- a/.github/workflows/release.yaml
+++ b/.github/workflows/release.yaml
@@ -7,34 +7,6 @@ on:
     branches:
       - release**
 jobs:
-  datasets:
-    name: datasets
-    runs-on: ubuntu-latest
-    steps:
-      - name: Clone repo
-        uses: actions/checkout@v4.1.4
-      - name: Set up python
-        uses: actions/setup-python@v5.1.0
-        with:
-          python-version: "3.12"
-      - name: Cache dependencies
-        uses: actions/cache@v4.0.2
-        id: cache
-        with:
-          path: ${{ env.pythonLocation }}
-          key: ${{ env.pythonLocation }}-${{ hashFiles('pyproject.toml') }}-datasets
-      - name: Install pip dependencies
-        if: steps.cache.outputs.cache-hit != 'true'
-        run: |
-          pip install .[tests]
-          pip cache purge
-      - name: List pip dependencies
-        run: pip list
-      - name: Run pytest checks
-        run: |
-          pytest --cov=torchgeo --cov-report=xml --durations=10
-          python -m torchgeo --help
-          torchgeo --help
   integration:
     name: integration
     runs-on: ubuntu-latest
diff --git a/.github/workflows/tests.yaml b/.github/workflows/tests.yaml
index c572be4633e..b63eb738db6 100644
--- a/.github/workflows/tests.yaml
+++ b/.github/workflows/tests.yaml
@@ -99,6 +99,39 @@ jobs:
         uses: codecov/codecov-action@v4.3.0
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
+  datasets:
+    name: datasets
+    runs-on: ubuntu-latest
+    env:
+      MPLBACKEND: Agg
+    steps:
+      - name: Clone repo
+        uses: actions/checkout@v4.1.4
+      - name: Set up python
+        uses: actions/setup-python@v5.1.0
+        with:
+          python-version: "3.12"
+      - name: Cache dependencies
+        uses: actions/cache@v4.0.2
+        id: cache
+        with:
+          path: ${{ env.pythonLocation }}
+          key: ${{ env.pythonLocation }}-${{ hashFiles('requirements/required.txt') }}-${{ hashFiles('requirements/tests.txt') }}
+      - name: Install pip dependencies
+        if: steps.cache.outputs.cache-hit != 'true'
+        run: |
+          pip install -r requirements/required.txt -r requirements/tests.txt
+          pip cache purge
+      - name: List pip dependencies
+        run: pip list
+      - name: Run pytest checks
+        run: |
+          pytest --cov=torchgeo --cov-report=xml --durations=10
+          python3 -m torchgeo --help
+      - name: Report coverage
+        uses: codecov/codecov-action@v4.3.0
+        with:
+          token: ${{ secrets.CODECOV_TOKEN }}
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.head.label || github.head_ref || github.ref }}
   cancel-in-progress: true
diff --git a/tests/datamodules/test_usavars.py b/tests/datamodules/test_usavars.py
index 004750c2840..c6d5b7c77bf 100644
--- a/tests/datamodules/test_usavars.py
+++ b/tests/datamodules/test_usavars.py
@@ -14,7 +14,6 @@
 class TestUSAVarsDataModule:
     @pytest.fixture
     def datamodule(self, request: SubRequest) -> USAVarsDataModule:
-        pytest.importorskip('pandas', minversion='1.1.3')
         root = os.path.join('tests', 'data', 'usavars')
         batch_size = 1
         num_workers = 0
diff --git a/tests/datasets/test_advance.py b/tests/datasets/test_advance.py
index efa3ac538f8..9e8ef718a79 100644
--- a/tests/datasets/test_advance.py
+++ b/tests/datasets/test_advance.py
@@ -1,11 +1,9 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import os
 import shutil
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -16,6 +14,8 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import ADVANCE, DatasetNotFoundError
 
+from .utils import importandskip
+
 
 def download_url(url: str, root: str, *args: str) -> None:
     shutil.copy(url, root)
@@ -37,19 +37,8 @@ def dataset(self, monkeypatch: MonkeyPatch, tmp_path: Path) -> ADVANCE:
         transforms = nn.Identity()
         return ADVANCE(root, transforms, download=True, checksum=True)
 
-    @pytest.fixture
-    def mock_missing_module(self, monkeypatch: MonkeyPatch) -> None:
-        import_orig = builtins.__import__
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name == 'scipy.io':
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-
     def test_getitem(self, dataset: ADVANCE) -> None:
-        pytest.importorskip('scipy', minversion='1.6.2')
+        pytest.importorskip('scipy', minversion='1.7.2')
         x = dataset[0]
         assert isinstance(x, dict)
         assert isinstance(x['image'], torch.Tensor)
@@ -71,17 +60,14 @@ def test_not_downloaded(self, tmp_path: Path) -> None:
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             ADVANCE(str(tmp_path))
 
-    def test_mock_missing_module(
-        self, dataset: ADVANCE, mock_missing_module: None
-    ) -> None:
-        with pytest.raises(
-            ImportError,
-            match='scipy is not installed and is required to use this dataset',
-        ):
+    def test_missing_module(self, dataset: ADVANCE) -> None:
+        importandskip('scipy')
+        match = 'scipy is not installed and is required to use this dataset'
+        with pytest.raises(ImportError, match=match):
             dataset[0]
 
     def test_plot(self, dataset: ADVANCE) -> None:
-        pytest.importorskip('scipy', minversion='1.6.2')
+        pytest.importorskip('scipy', minversion='1.7.2')
         x = dataset[0].copy()
         dataset.plot(x, suptitle='Test')
         plt.close()
diff --git a/tests/datasets/test_chabud.py b/tests/datasets/test_chabud.py
index 955104a53fb..f93f036b248 100644
--- a/tests/datasets/test_chabud.py
+++ b/tests/datasets/test_chabud.py
@@ -1,11 +1,9 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import os
 import shutil
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -17,7 +15,7 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import ChaBuD, DatasetNotFoundError
 
-pytest.importorskip('h5py', minversion='3')
+from .utils import importandskip
 
 
 def download_url(url: str, root: str, filename: str, *args: str, **kwargs: str) -> None:
@@ -29,6 +27,7 @@ class TestChaBuD:
     def dataset(
         self, monkeypatch: MonkeyPatch, tmp_path: Path, request: SubRequest
     ) -> ChaBuD:
+        pytest.importorskip('h5py', minversion='3.6')
         monkeypatch.setattr(torchgeo.datasets.chabud, 'download_url', download_url)
         data_dir = os.path.join('tests', 'data', 'chabud')
         url = os.path.join(data_dir, 'train_eval.hdf5')
@@ -47,17 +46,6 @@ def dataset(
             checksum=True,
         )
 
-    @pytest.fixture
-    def mock_missing_module(self, monkeypatch: MonkeyPatch) -> None:
-        import_orig = builtins.__import__
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name == 'h5py':
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-
     def test_getitem(self, dataset: ChaBuD) -> None:
         x = dataset[0]
         assert isinstance(x, dict)
@@ -82,17 +70,16 @@ def test_already_downloaded(self, dataset: ChaBuD) -> None:
         ChaBuD(root=dataset.root, download=True)
 
     def test_not_downloaded(self, tmp_path: Path) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             ChaBuD(str(tmp_path))
 
-    def test_mock_missing_module(
-        self, dataset: ChaBuD, tmp_path: Path, mock_missing_module: None
-    ) -> None:
-        with pytest.raises(
-            ImportError,
-            match='h5py is not installed and is required to use this dataset',
-        ):
-            ChaBuD(dataset.root, download=True, checksum=True)
+    def test_missing_module(self) -> None:
+        importandskip('h5py')
+        root = os.path.join('tests', 'data', 'chabud')
+        match = 'h5py is not installed and is required to use this dataset'
+        with pytest.raises(ImportError, match=match):
+            ChaBuD(root)
 
     def test_invalid_bands(self) -> None:
         with pytest.raises(AssertionError):
diff --git a/tests/datasets/test_chesapeake.py b/tests/datasets/test_chesapeake.py
index 459c6bcdbdf..1cff011441f 100644
--- a/tests/datasets/test_chesapeake.py
+++ b/tests/datasets/test_chesapeake.py
@@ -29,10 +29,9 @@ def download_url(url: str, root: str, *args: str, **kwargs: str) -> None:
 
 
 class TestChesapeake13:
-    pytest.importorskip('zipfile_deflate64')
-
     @pytest.fixture
     def dataset(self, monkeypatch: MonkeyPatch, tmp_path: Path) -> Chesapeake13:
+        pytest.importorskip('zipfile_deflate64')
         monkeypatch.setattr(torchgeo.datasets.chesapeake, 'download_url', download_url)
         md5 = 'fe35a615b8e749b21270472aa98bb42c'
         monkeypatch.setattr(Chesapeake13, 'md5', md5)
@@ -63,6 +62,7 @@ def test_already_extracted(self, dataset: Chesapeake13) -> None:
         Chesapeake13(dataset.paths, download=True)
 
     def test_already_downloaded(self, tmp_path: Path) -> None:
+        pytest.importorskip('zipfile_deflate64')
         url = os.path.join(
             'tests', 'data', 'chesapeake', 'BAYWIDE', 'Baywide_13Class_20132014.zip'
         )
diff --git a/tests/datasets/test_cropharvest.py b/tests/datasets/test_cropharvest.py
index f478cdf53ad..69ef07d025d 100644
--- a/tests/datasets/test_cropharvest.py
+++ b/tests/datasets/test_cropharvest.py
@@ -1,11 +1,9 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import os
 import shutil
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -16,7 +14,7 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import CropHarvest, DatasetNotFoundError
 
-pytest.importorskip('h5py', minversion='3')
+from .utils import importandskip
 
 
 def download_url(url: str, root: str, filename: str, md5: str) -> None:
@@ -24,17 +22,6 @@ def download_url(url: str, root: str, filename: str, md5: str) -> None:
 
 
 class TestCropHarvest:
-    @pytest.fixture
-    def mock_missing_module(self, monkeypatch: MonkeyPatch) -> None:
-        import_orig = builtins.__import__
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name == 'h5py':
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-
     @pytest.fixture
     def dataset(self, monkeypatch: MonkeyPatch, tmp_path: Path) -> CropHarvest:
         monkeypatch.setattr(torchgeo.datasets.cropharvest, 'download_url', download_url)
@@ -62,6 +49,7 @@ def dataset(self, monkeypatch: MonkeyPatch, tmp_path: Path) -> CropHarvest:
         return dataset
 
     def test_getitem(self, dataset: CropHarvest) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         x = dataset[0]
         assert isinstance(x, dict)
         assert isinstance(x['array'], torch.Tensor)
@@ -73,28 +61,27 @@ def test_getitem(self, dataset: CropHarvest) -> None:
     def test_len(self, dataset: CropHarvest) -> None:
         assert len(dataset) == 5
 
-    def test_already_downloaded(self, dataset: CropHarvest, tmp_path: Path) -> None:
-        CropHarvest(root=str(tmp_path), download=False)
+    def test_already_downloaded(self, dataset: CropHarvest) -> None:
+        CropHarvest(dataset.root)
 
     def test_downloaded_zipped(self, dataset: CropHarvest, tmp_path: Path) -> None:
         feature_path = os.path.join(tmp_path, 'features')
         shutil.rmtree(feature_path)
-        CropHarvest(root=str(tmp_path), download=True)
+        CropHarvest(root=str(tmp_path))
 
     def test_not_downloaded(self, tmp_path: Path) -> None:
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             CropHarvest(str(tmp_path))
 
     def test_plot(self, dataset: CropHarvest) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         x = dataset[0].copy()
         dataset.plot(x, suptitle='Test')
         plt.close()
 
-    def test_mock_missing_module(
-        self, dataset: CropHarvest, tmp_path: Path, mock_missing_module: None
-    ) -> None:
-        with pytest.raises(
-            ImportError,
-            match='h5py is not installed and is required to use this dataset',
-        ):
-            CropHarvest(root=str(tmp_path), download=True)[0]
+    def test_missing_module(self) -> None:
+        importandskip('h5py')
+        root = os.path.join('tests', 'data', 'cropharvest')
+        match = 'h5py is not installed and is required to use this dataset'
+        with pytest.raises(ImportError, match=match):
+            CropHarvest(root)[0]
diff --git a/tests/datasets/test_idtrees.py b/tests/datasets/test_idtrees.py
index e907213c29c..4726f9c957a 100644
--- a/tests/datasets/test_idtrees.py
+++ b/tests/datasets/test_idtrees.py
@@ -1,12 +1,10 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import glob
 import os
 import shutil
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -18,7 +16,7 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import DatasetNotFoundError, IDTReeS
 
-pytest.importorskip('laspy', minversion='2')
+from .utils import importandskip
 
 
 def download_url(url: str, root: str, *args: str, **kwargs: str) -> None:
@@ -50,20 +48,8 @@ def dataset(
         transforms = nn.Identity()
         return IDTReeS(root, split, task, transforms, download=True, checksum=True)
 
-    @pytest.fixture(params=['laspy', 'pyvista'])
-    def mock_missing_module(self, monkeypatch: MonkeyPatch, request: SubRequest) -> str:
-        import_orig = builtins.__import__
-        package = str(request.param)
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name == package:
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-        return package
-
     def test_getitem(self, dataset: IDTReeS) -> None:
+        pytest.importorskip('laspy', minversion='2')
         x = dataset[0]
         assert isinstance(x, dict)
         assert isinstance(x['image'], torch.Tensor)
@@ -91,35 +77,31 @@ def test_already_downloaded(self, dataset: IDTReeS) -> None:
         IDTReeS(root=dataset.root, download=True)
 
     def test_not_downloaded(self, tmp_path: Path) -> None:
+        pytest.importorskip('laspy', minversion='2')
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             IDTReeS(str(tmp_path))
 
     def test_not_extracted(self, tmp_path: Path) -> None:
+        pytest.importorskip('laspy', minversion='2')
         pathname = os.path.join('tests', 'data', 'idtrees', '*.zip')
         root = str(tmp_path)
         for zipfile in glob.iglob(pathname):
             shutil.copy(zipfile, root)
         IDTReeS(root)
 
-    def test_mock_missing_module(
-        self, dataset: IDTReeS, mock_missing_module: str
-    ) -> None:
-        package = mock_missing_module
-
-        if package == 'laspy':
-            with pytest.raises(
-                ImportError,
-                match=f'{package} is not installed and is required to use this dataset',
-            ):
-                IDTReeS(dataset.root, dataset.split, dataset.task)
-        elif package == 'pyvista':
-            with pytest.raises(
-                ImportError,
-                match=f'{package} is not installed and is required to plot point cloud',
-            ):
-                dataset.plot_las(0)
+    def test_missing_module(self, dataset: IDTReeS) -> None:
+        importandskip('laspy')
+        match = 'laspy is not installed and is required to use this dataset'
+        with pytest.raises(ImportError, match=match):
+            dataset[0]
+
+        importandskip('pyvista')
+        match = 'pyvista is not installed and is required to plot point cloud'
+        with pytest.raises(ImportError, match=match):
+            dataset.plot_las(0)
 
     def test_plot(self, dataset: IDTReeS) -> None:
+        pytest.importorskip('laspy', minversion='2')
         x = dataset[0].copy()
         dataset.plot(x, suptitle='Test')
         plt.close()
@@ -136,6 +118,7 @@ def test_plot(self, dataset: IDTReeS) -> None:
             plt.close()
 
     def test_plot_las(self, dataset: IDTReeS) -> None:
+        pytest.importorskip('laspy', minversion='2')
         pyvista = pytest.importorskip('pyvista', minversion='0.34.2')
         pyvista.OFF_SCREEN = True
 
diff --git a/tests/datasets/test_landcoverai.py b/tests/datasets/test_landcoverai.py
index 7222ff78bbc..7b2d0b6d3f3 100644
--- a/tests/datasets/test_landcoverai.py
+++ b/tests/datasets/test_landcoverai.py
@@ -80,7 +80,7 @@ class TestLandCoverAI:
     def dataset(
         self, monkeypatch: MonkeyPatch, tmp_path: Path, request: SubRequest
     ) -> LandCoverAI:
-        pytest.importorskip('cv2', minversion='4.4.0')
+        pytest.importorskip('cv2', minversion='4.5.4')
         monkeypatch.setattr(torchgeo.datasets.landcoverai, 'download_url', download_url)
         md5 = 'ff8998857cc8511f644d3f7d0f3688d0'
         monkeypatch.setattr(LandCoverAI, 'md5', md5)
@@ -111,7 +111,7 @@ def test_already_extracted(self, dataset: LandCoverAI) -> None:
         LandCoverAI(root=dataset.root, download=True)
 
     def test_already_downloaded(self, monkeypatch: MonkeyPatch, tmp_path: Path) -> None:
-        pytest.importorskip('cv2', minversion='4.4.0')
+        pytest.importorskip('cv2', minversion='4.5.4')
         sha256 = 'ecec8e871faf1bbd8ca525ca95ddc1c1f5213f40afb94599884bd85f990ebd6b'
         monkeypatch.setattr(LandCoverAI, 'sha256', sha256)
         url = os.path.join('tests', 'data', 'landcoverai', 'landcover.ai.v1.zip')
diff --git a/tests/datasets/test_quakeset.py b/tests/datasets/test_quakeset.py
index 59596f2c0c5..112ec10e0c9 100644
--- a/tests/datasets/test_quakeset.py
+++ b/tests/datasets/test_quakeset.py
@@ -1,11 +1,9 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import os
 import shutil
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -17,6 +15,8 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import DatasetNotFoundError, QuakeSet
 
+from .utils import importandskip
+
 
 def download_url(url: str, root: str, *args: str, **kwargs: str) -> None:
     shutil.copy(url, root)
@@ -27,6 +27,7 @@ class TestQuakeSet:
     def dataset(
         self, monkeypatch: MonkeyPatch, tmp_path: Path, request: SubRequest
     ) -> QuakeSet:
+        pytest.importorskip('h5py', minversion='3.6')
         monkeypatch.setattr(torchgeo.datasets.quakeset, 'download_url', download_url)
         url = os.path.join('tests', 'data', 'quakeset', 'earthquakes.h5')
         md5 = '127d0d6a1f82d517129535f50053a4c9'
@@ -39,25 +40,14 @@ def dataset(
             root, split, transforms=transforms, download=True, checksum=True
         )
 
-    @pytest.fixture
-    def mock_missing_module(self, monkeypatch: MonkeyPatch) -> None:
-        import_orig = builtins.__import__
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name == 'h5py':
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-
-    def test_mock_missing_module(
-        self, dataset: QuakeSet, tmp_path: Path, mock_missing_module: None
-    ) -> None:
+    def test_missing_module(self) -> None:
+        importandskip('h5py')
+        root = os.path.join('tests', 'data', 'quakeset')
         with pytest.raises(
             ImportError,
             match='h5py is not installed and is required to use this dataset',
         ):
-            QuakeSet(dataset.root, download=True, checksum=True)
+            QuakeSet(root)
 
     def test_getitem(self, dataset: QuakeSet) -> None:
         x = dataset[0]
@@ -69,10 +59,11 @@ def test_getitem(self, dataset: QuakeSet) -> None:
     def test_len(self, dataset: QuakeSet) -> None:
         assert len(dataset) == 8
 
-    def test_already_downloaded(self, dataset: QuakeSet, tmp_path: Path) -> None:
-        QuakeSet(root=str(tmp_path), download=True)
+    def test_already_downloaded(self, dataset: QuakeSet) -> None:
+        QuakeSet(dataset.root)
 
     def test_not_downloaded(self, tmp_path: Path) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             QuakeSet(str(tmp_path))
 
diff --git a/tests/datasets/test_skippd.py b/tests/datasets/test_skippd.py
index 01907d22cea..fc4140cb03f 100644
--- a/tests/datasets/test_skippd.py
+++ b/tests/datasets/test_skippd.py
@@ -1,12 +1,10 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import os
 import shutil
 from itertools import product
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -18,7 +16,7 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import SKIPPD, DatasetNotFoundError
 
-pytest.importorskip('h5py', minversion='3')
+from .utils import importandskip
 
 
 def download_url(url: str, root: str, *args: str, **kwargs: str) -> None:
@@ -30,6 +28,7 @@ class TestSKIPPD:
     def dataset(
         self, monkeypatch: MonkeyPatch, tmp_path: Path, request: SubRequest
     ) -> SKIPPD:
+        pytest.importorskip('h5py', minversion='3.6')
         task, split = request.param
 
         monkeypatch.setattr(torchgeo.datasets.skippd, 'download_url', download_url)
@@ -53,31 +52,19 @@ def dataset(
             checksum=True,
         )
 
-    @pytest.fixture
-    def mock_missing_module(self, monkeypatch: MonkeyPatch) -> None:
-        import_orig = builtins.__import__
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name == 'h5py':
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-
-    def test_mock_missing_module(
-        self, dataset: SKIPPD, tmp_path: Path, mock_missing_module: None
-    ) -> None:
-        with pytest.raises(
-            ImportError,
-            match='h5py is not installed and is required to use this dataset',
-        ):
-            SKIPPD(dataset.root, download=True, checksum=True)
+    def test_missing_module(self) -> None:
+        importandskip('h5py')
+        root = os.path.join('tests', 'data', 'skippd')
+        match = 'h5py is not installed and is required to use this dataset'
+        with pytest.raises(ImportError, match=match):
+            SKIPPD(root)
 
     def test_already_extracted(self, dataset: SKIPPD) -> None:
         SKIPPD(root=dataset.root, download=True)
 
     @pytest.mark.parametrize('task', ['nowcast', 'forecast'])
     def test_already_downloaded(self, tmp_path: Path, task: str) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         pathname = os.path.join(
             'tests', 'data', 'skippd', f'2017_2019_images_pv_processed_{task}.zip'
         )
@@ -105,6 +92,7 @@ def test_invalid_split(self) -> None:
             SKIPPD(split='foo')
 
     def test_not_downloaded(self, tmp_path: Path) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             SKIPPD(str(tmp_path))
 
diff --git a/tests/datasets/test_so2sat.py b/tests/datasets/test_so2sat.py
index 6daa4057bed..d97cad5762a 100644
--- a/tests/datasets/test_so2sat.py
+++ b/tests/datasets/test_so2sat.py
@@ -1,10 +1,8 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import os
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -15,12 +13,13 @@
 
 from torchgeo.datasets import DatasetNotFoundError, RGBBandsMissingError, So2Sat
 
-pytest.importorskip('h5py', minversion='3')
+from .utils import importandskip
 
 
 class TestSo2Sat:
     @pytest.fixture(params=['train', 'validation', 'test'])
     def dataset(self, monkeypatch: MonkeyPatch, request: SubRequest) -> So2Sat:
+        pytest.importorskip('h5py', minversion='3.6')
         md5s_by_version = {
             '2': {
                 'train': '56e6fa0edb25b065124a3113372f76e5',
@@ -35,17 +34,6 @@ def dataset(self, monkeypatch: MonkeyPatch, request: SubRequest) -> So2Sat:
         transforms = nn.Identity()
         return So2Sat(root=root, split=split, transforms=transforms, checksum=True)
 
-    @pytest.fixture
-    def mock_missing_module(self, monkeypatch: MonkeyPatch) -> None:
-        import_orig = builtins.__import__
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name == 'h5py':
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-
     def test_getitem(self, dataset: So2Sat) -> None:
         x = dataset[0]
         assert isinstance(x, dict)
@@ -62,14 +50,17 @@ def test_out_of_bounds(self, dataset: So2Sat) -> None:
             dataset[2]
 
     def test_invalid_split(self) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         with pytest.raises(AssertionError):
             So2Sat(split='foo')
 
     def test_invalid_bands(self) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         with pytest.raises(ValueError):
             So2Sat(bands=('OK', 'BK'))
 
     def test_not_downloaded(self, tmp_path: Path) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             So2Sat(str(tmp_path))
 
@@ -90,11 +81,8 @@ def test_plot_rgb(self, dataset: So2Sat) -> None:
         ):
             dataset.plot(dataset[0], suptitle='Single Band')
 
-    def test_mock_missing_module(
-        self, dataset: So2Sat, mock_missing_module: None
-    ) -> None:
-        with pytest.raises(
-            ImportError,
-            match='h5py is not installed and is required to use this dataset',
-        ):
-            So2Sat(dataset.root)
+    def test_missing_module(self) -> None:
+        importandskip('h5py')
+        match = 'h5py is not installed and is required to use this dataset'
+        with pytest.raises(ImportError, match=match):
+            So2Sat()
diff --git a/tests/datasets/test_utils.py b/tests/datasets/test_utils.py
index 092e8864acf..bb4cfedb38a 100644
--- a/tests/datasets/test_utils.py
+++ b/tests/datasets/test_utils.py
@@ -1,7 +1,6 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import glob
 import math
 import os
@@ -38,6 +37,8 @@
     working_dir,
 )
 
+from .utils import importandskip
+
 
 class TestDatasetNotFoundError:
     def test_none(self) -> None:
@@ -85,18 +86,6 @@ def test_paths_download(self) -> None:
             raise DatasetNotFoundError(ds)
 
 
-@pytest.fixture
-def mock_missing_module(monkeypatch: MonkeyPatch) -> None:
-    import_orig = builtins.__import__
-
-    def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-        if name in ['radiant_mlhub', 'rarfile', 'zipfile_deflate64']:
-            raise ImportError()
-        return import_orig(name, *args, **kwargs)
-
-    monkeypatch.setattr(builtins, '__import__', mocked_import)
-
-
 class MLHubDataset:
     def download(self, output_dir: str, **kwargs: str) -> None:
         glob_path = os.path.join(
@@ -127,10 +116,6 @@ def download_url(url: str, root: str, *args: str) -> None:
     shutil.copy(url, root)
 
 
-def test_mock_missing_module(mock_missing_module: None) -> None:
-    import sys  # noqa: F401
-
-
 @pytest.mark.parametrize(
     'src',
     [
@@ -150,17 +135,16 @@ def test_extract_archive(src: str, tmp_path: Path) -> None:
     extract_archive(os.path.join('tests', 'data', src), str(tmp_path))
 
 
-def test_missing_rarfile(mock_missing_module: None) -> None:
-    with pytest.raises(
-        ImportError,
-        match='rarfile is not installed and is required to extract this dataset',
-    ):
-        extract_archive(
-            os.path.join('tests', 'data', 'vhr10', 'NWPU VHR-10 dataset.rar')
-        )
+def test_missing_rarfile() -> None:
+    importandskip('rarfile')
+    match = 'rarfile is not installed and is required to extract this dataset'
+    with pytest.raises(ImportError, match=match):
+        path = os.path.join('tests', 'data', 'vhr10', 'NWPU VHR-10 dataset.rar')
+        extract_archive(path)
 
 
-def test_missing_zipfile_deflate64(mock_missing_module: None) -> None:
+def test_missing_zipfile_deflate64() -> None:
+    importandskip('zipfile_deflate64')
     # Should fallback on Python builtin zipfile
     extract_archive(os.path.join('tests', 'data', 'landcoverai', 'landcover.ai.v1.zip'))
 
@@ -196,18 +180,13 @@ def test_download_radiant_mlhub_collection(
     download_radiant_mlhub_collection('', str(tmp_path))
 
 
-def test_missing_radiant_mlhub(mock_missing_module: None) -> None:
-    with pytest.raises(
-        ImportError,
-        match='radiant_mlhub is not installed and is required to download this dataset',
-    ):
+def test_missing_radiant_mlhub() -> None:
+    importandskip('radiant_mlhub')
+    match = 'radiant_mlhub is not installed and is required to download this {}'
+    with pytest.raises(ImportError, match=match.format('dataset')):
         download_radiant_mlhub_dataset('', '')
 
-    with pytest.raises(
-        ImportError,
-        match='radiant_mlhub is not installed and is required to download this'
-        + ' collection',
-    ):
+    with pytest.raises(ImportError, match=match.format('collection')):
         download_radiant_mlhub_collection('', '')
 
 
diff --git a/tests/datasets/test_vhr10.py b/tests/datasets/test_vhr10.py
index de4c6c2d507..4479fc8ae9b 100644
--- a/tests/datasets/test_vhr10.py
+++ b/tests/datasets/test_vhr10.py
@@ -1,11 +1,9 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import os
 import shutil
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -18,7 +16,7 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import VHR10, DatasetNotFoundError
 
-pytest.importorskip('pycocotools')
+from .utils import importandskip
 
 
 def download_url(url: str, root: str, *args: str) -> None:
@@ -30,6 +28,9 @@ class TestVHR10:
     def dataset(
         self, monkeypatch: MonkeyPatch, tmp_path: Path, request: SubRequest
     ) -> VHR10:
+        split = request.param
+        if split == 'positive':
+            pytest.importorskip('pycocotools')
         pytest.importorskip('rarfile', minversion='4')
         monkeypatch.setattr(torchgeo.datasets.vhr10, 'download_url', download_url)
         monkeypatch.setattr(torchgeo.datasets.utils, 'download_url', download_url)
@@ -42,21 +43,9 @@ def dataset(
         md5 = '567c4cd8c12624864ff04865de504c58'
         monkeypatch.setitem(VHR10.target_meta, 'md5', md5)
         root = str(tmp_path)
-        split = request.param
         transforms = nn.Identity()
         return VHR10(root, split, transforms, download=True, checksum=True)
 
-    @pytest.fixture
-    def mock_missing_modules(self, monkeypatch: MonkeyPatch) -> None:
-        import_orig = builtins.__import__
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name in {'pycocotools.coco', 'skimage.measure'}:
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-
     def test_getitem(self, dataset: VHR10) -> None:
         for i in range(2):
             x = dataset[i]
@@ -93,25 +82,25 @@ def test_not_downloaded(self, tmp_path: Path) -> None:
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             VHR10(str(tmp_path))
 
-    def test_mock_missing_module(
-        self, dataset: VHR10, mock_missing_modules: None
-    ) -> None:
-        if dataset.split == 'positive':
-            with pytest.raises(
-                ImportError,
-                match='pycocotools is not installed and is required to use this datase',
-            ):
-                VHR10(dataset.root, dataset.split)
-
-            with pytest.raises(
-                ImportError,
-                match='scikit-image is not installed and is required to plot masks',
-            ):
-                x = dataset[0]
-                dataset.plot(x)
+    def test_missing_module(self) -> None:
+        importandskip('pycocotools')
+        root = os.path.join('tests', 'data', 'vhr10')
+        match = 'pycocotools is not installed and is required to use this datase'
+        with pytest.raises(ImportError, match=match):
+            VHR10(root, 'positive')
+
+    def test_missing_module_plot(self) -> None:
+        importandskip('skimage')
+        root = os.path.join('tests', 'data', 'vhr10')
+        match = 'scikit-image is not installed and is required to plot masks'
+        with pytest.raises(ImportError, match=match):
+            ds = VHR10(root, 'negative')
+            x = ds[0]
+            ds.split = 'positive'
+            ds.plot(x)
 
     def test_plot(self, dataset: VHR10) -> None:
-        pytest.importorskip('skimage', minversion='0.18')
+        pytest.importorskip('skimage', minversion='0.19')
         x = dataset[1].copy()
         dataset.plot(x, suptitle='Test')
         plt.close()
diff --git a/tests/datasets/test_zuericrop.py b/tests/datasets/test_zuericrop.py
index 330866b36f0..8984e47f05a 100644
--- a/tests/datasets/test_zuericrop.py
+++ b/tests/datasets/test_zuericrop.py
@@ -1,11 +1,9 @@
 # Copyright (c) Microsoft Corporation. All rights reserved.
 # Licensed under the MIT License.
 
-import builtins
 import os
 import shutil
 from pathlib import Path
-from typing import Any
 
 import matplotlib.pyplot as plt
 import pytest
@@ -16,7 +14,7 @@
 import torchgeo.datasets.utils
 from torchgeo.datasets import DatasetNotFoundError, RGBBandsMissingError, ZueriCrop
 
-pytest.importorskip('h5py', minversion='3')
+from .utils import importandskip
 
 
 def download_url(url: str, root: str, *args: str, **kwargs: str) -> None:
@@ -26,6 +24,7 @@ def download_url(url: str, root: str, *args: str, **kwargs: str) -> None:
 class TestZueriCrop:
     @pytest.fixture
     def dataset(self, monkeypatch: MonkeyPatch, tmp_path: Path) -> ZueriCrop:
+        pytest.importorskip('h5py', minversion='3.6')
         monkeypatch.setattr(torchgeo.datasets.zuericrop, 'download_url', download_url)
         data_dir = os.path.join('tests', 'data', 'zuericrop')
         urls = [
@@ -39,17 +38,6 @@ def dataset(self, monkeypatch: MonkeyPatch, tmp_path: Path) -> ZueriCrop:
         transforms = nn.Identity()
         return ZueriCrop(root=root, transforms=transforms, download=True, checksum=True)
 
-    @pytest.fixture
-    def mock_missing_module(self, monkeypatch: MonkeyPatch) -> None:
-        import_orig = builtins.__import__
-
-        def mocked_import(name: str, *args: Any, **kwargs: Any) -> Any:
-            if name == 'h5py':
-                raise ImportError()
-            return import_orig(name, *args, **kwargs)
-
-        monkeypatch.setattr(builtins, '__import__', mocked_import)
-
     def test_getitem(self, dataset: ZueriCrop) -> None:
         x = dataset[0]
         assert isinstance(x, dict)
@@ -79,19 +67,19 @@ def test_already_downloaded(self, dataset: ZueriCrop) -> None:
         ZueriCrop(root=dataset.root, download=True)
 
     def test_not_downloaded(self, tmp_path: Path) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         with pytest.raises(DatasetNotFoundError, match='Dataset not found'):
             ZueriCrop(str(tmp_path))
 
-    def test_mock_missing_module(
-        self, dataset: ZueriCrop, tmp_path: Path, mock_missing_module: None
-    ) -> None:
-        with pytest.raises(
-            ImportError,
-            match='h5py is not installed and is required to use this dataset',
-        ):
-            ZueriCrop(dataset.root, download=True, checksum=True)
+    def test_missing_module(self) -> None:
+        importandskip('h5py')
+        root = os.path.join('tests', 'data', 'zuericrop')
+        match = 'h5py is not installed and is required to use this dataset'
+        with pytest.raises(ImportError, match=match):
+            ZueriCrop(root)
 
     def test_invalid_bands(self) -> None:
+        pytest.importorskip('h5py', minversion='3.6')
         with pytest.raises(ValueError):
             ZueriCrop(bands=('OK', 'BK'))
 
diff --git a/tests/datasets/utils.py b/tests/datasets/utils.py
new file mode 100644
index 00000000000..ff1fd45b226
--- /dev/null
+++ b/tests/datasets/utils.py
@@ -0,0 +1,21 @@
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License.
+
+import pytest
+
+
+def importandskip(modname: str, reason: str | None = None) -> None:
+    """Exact opposite of :func:`pytest.importorskip`.
+
+    Args:
+        modname: The name of the module to import.
+        reason: If given, this reason is shown as the message when the module can
+            be imported.
+    """
+    try:
+        __import__(modname)
+        if reason is None:
+            reason = f'could import {modname!r}'
+        pytest.skip(reason)
+    except ImportError:
+        pass
diff --git a/tests/trainers/test_classification.py b/tests/trainers/test_classification.py
index ca4da48bb78..3402bd0f3fc 100644
--- a/tests/trainers/test_classification.py
+++ b/tests/trainers/test_classification.py
@@ -89,7 +89,7 @@ def test_trainer(
         self, monkeypatch: MonkeyPatch, name: str, fast_dev_run: bool
     ) -> None:
         if name.startswith('so2sat') or name == 'quakeset':
-            pytest.importorskip('h5py', minversion='3')
+            pytest.importorskip('h5py', minversion='3.6')
 
         config = os.path.join('tests', 'conf', name + '.yaml')
 
diff --git a/tests/trainers/test_regression.py b/tests/trainers/test_regression.py
index ef3c6164d98..c62c808c72f 100644
--- a/tests/trainers/test_regression.py
+++ b/tests/trainers/test_regression.py
@@ -71,7 +71,7 @@ def test_trainer(
         self, monkeypatch: MonkeyPatch, name: str, fast_dev_run: bool
     ) -> None:
         if name == 'skippd':
-            pytest.importorskip('h5py', minversion='3')
+            pytest.importorskip('h5py', minversion='3.6')
 
         config = os.path.join('tests', 'conf', name + '.yaml')
 
diff --git a/tests/trainers/test_segmentation.py b/tests/trainers/test_segmentation.py
index 88cd58c0553..d8b207d5d2d 100644
--- a/tests/trainers/test_segmentation.py
+++ b/tests/trainers/test_segmentation.py
@@ -87,12 +87,16 @@ class TestSemanticSegmentationTask:
     def test_trainer(
         self, monkeypatch: MonkeyPatch, name: str, fast_dev_run: bool
     ) -> None:
-        if name == 'naipchesapeake':
-            pytest.importorskip('zipfile_deflate64')
-
-        if name == 'landcoverai':
-            sha256 = 'ecec8e871faf1bbd8ca525ca95ddc1c1f5213f40afb94599884bd85f990ebd6b'
-            monkeypatch.setattr(LandCoverAI, 'sha256', sha256)
+        match name:
+            case 'chabud':
+                pytest.importorskip('h5py', minversion='3.6')
+            case 'landcoverai':
+                sha256 = (
+                    'ecec8e871faf1bbd8ca525ca95ddc1c1f5213f40afb94599884bd85f990ebd6b'
+                )
+                monkeypatch.setattr(LandCoverAI, 'sha256', sha256)
+            case 'naipchesapeake':
+                pytest.importorskip('zipfile_deflate64')
 
         config = os.path.join('tests', 'conf', name + '.yaml')
 
diff --git a/torchgeo/datasets/cropharvest.py b/torchgeo/datasets/cropharvest.py
index 9887d71db6b..4a383687332 100644
--- a/torchgeo/datasets/cropharvest.py
+++ b/torchgeo/datasets/cropharvest.py
@@ -111,15 +111,7 @@ def __init__(
 
         Raises:
             DatasetNotFoundError: If dataset is not found and *download* is False.
-            ImportError: If h5py is not installed
         """
-        try:
-            import h5py  # noqa: F401
-        except ImportError:
-            raise ImportError(
-                'h5py is not installed and is required to use this dataset'
-            )
-
         self.root = root
         self.transforms = transforms
         self.checksum = checksum
@@ -141,6 +133,9 @@ def __getitem__(self, index: int) -> dict[str, Tensor]:
 
         Returns:
             single pixel time-series array and label at that index
+
+        Raises:
+            ImportError: If h5py is not installed
         """
         files = self.files[index]
         data = self._load_array(files['chip'])
@@ -209,7 +204,12 @@ def _load_array(self, path: str) -> Tensor:
         Returns:
             the image
         """
-        import h5py
+        try:
+            import h5py  # noqa: F401
+        except ImportError:
+            raise ImportError(
+                'h5py is not installed and is required to use this dataset'
+            )
 
         filename = os.path.join(path)
         with h5py.File(filename, 'r') as f:
diff --git a/torchgeo/datasets/idtrees.py b/torchgeo/datasets/idtrees.py
index c17df135759..ffd1d460e48 100644
--- a/torchgeo/datasets/idtrees.py
+++ b/torchgeo/datasets/idtrees.py
@@ -166,7 +166,6 @@ def __init__(
             checksum: if True, check the MD5 of the downloaded files (may be slow)
 
         Raises:
-            ImportError: if laspy is not installed
             DatasetNotFoundError: If dataset is not found and *download* is False.
         """
         assert split in ['train', 'test']
@@ -182,13 +181,6 @@ def __init__(
         self.num_classes = len(self.classes)
         self._verify()
 
-        try:
-            import laspy  # noqa: F401
-        except ImportError:
-            raise ImportError(
-                'laspy is not installed and is required to use this dataset'
-            )
-
         self.images, self.geometries, self.labels = self._load(root)
 
     def __getitem__(self, index: int) -> dict[str, Tensor]:
@@ -199,6 +191,9 @@ def __getitem__(self, index: int) -> dict[str, Tensor]:
 
         Returns:
             data and label at that index
+
+        Raises:
+            ImportError: if laspy is not installed
         """
         path = self.images[index]
         image = self._load_image(path).to(torch.uint8)
@@ -262,7 +257,12 @@ def _load_las(self, path: str) -> Tensor:
         Returns:
             the point cloud
         """
-        import laspy
+        try:
+            import laspy  # noqa: F401
+        except ImportError:
+            raise ImportError(
+                'laspy is not installed and is required to use this dataset'
+            )
 
         las = laspy.read(path)
         array: 'np.typing.NDArray[np.int_]' = np.stack([las.x, las.y, las.z], axis=0)