diff --git a/skimage/morphology/convex_hull.py b/skimage/morphology/convex_hull.py
index bb9ca734..31b92a47 100644
--- a/skimage/morphology/convex_hull.py
+++ b/skimage/morphology/convex_hull.py
@@ -1,11 +1,12 @@
-__all__ = ['convex_hull_image', 'convex_hull_object']
-
+"""Convex Hull."""
 import numpy as np
 from ..measure._pnpoly import grid_points_in_poly
 from ._convex_hull import possible_hull
 from ..measure._label import label
 from ..util import unique_rows
 
+__all__ = ['convex_hull_image', 'convex_hull_object']
+
 try:
     from scipy.spatial import Delaunay
 except ImportError:
@@ -33,6 +34,8 @@ def convex_hull_image(image):
     .. [1] http://blogs.mathworks.com/steve/2011/10/04/binary-image-convex-hull-algorithm-notes/
 
     """
+    if image.ndim > 2:
+        raise ValueError("Input must be a 2D image")
 
     if Delaunay is None:
         raise ImportError("Could not import scipy.spatial.Delaunay, "
@@ -85,7 +88,7 @@ def convex_hull_object(image, neighbors=8):
 
     Parameters
     ----------
-    image : ndarray
+    image : (M, N) array
         Binary input image.
     neighbors : {4, 8}, int
         Whether to use 4- or 8-connectivity.
@@ -104,6 +107,8 @@ def convex_hull_object(image, neighbors=8):
     convex_hull_image separately on each object.
 
     """
+    if image.ndim > 2:
+        raise ValueError("Input must be a 2D image")
 
     if neighbors != 4 and neighbors != 8:
         raise ValueError('Neighbors must be either 4 or 8.')
diff --git a/skimage/morphology/tests/test_convex_hull.py b/skimage/morphology/tests/test_convex_hull.py
index c0cc954e..1a41b6c8 100644
--- a/skimage/morphology/tests/test_convex_hull.py
+++ b/skimage/morphology/tests/test_convex_hull.py
@@ -31,20 +31,24 @@ def test_basic():
 
     assert_array_equal(convex_hull_image(image), expected)
 
+    # Test that an error is raised on passing a 3D image:
+    image3d = np.empty((5, 5, 5))
+    assert_raises(ValueError, convex_hull_image, image3d)
+
 
 @skipif(not scipy_spatial)
 def test_qhull_offset_example():
-    nonzeros = (([1367, 1368, 1368, 1368, 1369, 1369, 1369, 1369, 1369, 1370, 1370,
-       1370, 1370, 1370, 1370, 1370, 1371, 1371, 1371, 1371, 1371, 1371,
-       1371, 1371, 1371, 1372, 1372, 1372, 1372, 1372, 1372, 1372, 1372,
-       1372, 1373, 1373, 1373, 1373, 1373, 1373, 1373, 1373, 1373, 1374,
-       1374, 1374, 1374, 1374, 1374, 1374, 1375, 1375, 1375, 1375, 1375,
-       1376, 1376, 1376, 1377]),
-    ([151, 150, 151, 152, 149, 150, 151, 152, 153, 148, 149, 150, 151,
-       152, 153, 154, 147, 148, 149, 150, 151, 152, 153, 154, 155, 146,
-       147, 148, 149, 150, 151, 152, 153, 154, 146, 147, 148, 149, 150,
-       151, 152, 153, 154, 147, 148, 149, 150, 151, 152, 153, 148, 149,
-       150, 151, 152, 149, 150, 151, 150]))
+    nonzeros = (([1367, 1368, 1368, 1368, 1369, 1369, 1369, 1369, 1369, 1370,
+                  1370, 1370, 1370, 1370, 1370, 1370, 1371, 1371, 1371, 1371,
+                  1371, 1371, 1371, 1371, 1371, 1372, 1372, 1372, 1372, 1372,
+                  1372, 1372, 1372, 1372, 1373, 1373, 1373, 1373, 1373, 1373,
+                  1373, 1373, 1373, 1374, 1374, 1374, 1374, 1374, 1374, 1374,
+                  1375, 1375, 1375, 1375, 1375, 1376, 1376, 1376, 1377]),
+                ([151, 150, 151, 152, 149, 150, 151, 152, 153, 148, 149, 150,
+                 151, 152, 153, 154, 147, 148, 149, 150, 151, 152, 153, 154,
+                 155, 146, 147, 148, 149, 150, 151, 152, 153, 154, 146, 147,
+                 148, 149, 150, 151, 152, 153, 154, 147, 148, 149, 150, 151,
+                 152, 153, 148, 149, 150, 151, 152, 149, 150, 151, 150]))
     image = np.zeros((1392, 1040), dtype=bool)
     image[nonzeros] = True
     expected = image.copy()
@@ -139,6 +143,9 @@ def test_object():
 
     assert_raises(ValueError, convex_hull_object, image, 7)
 
+    # Test that an error is raised on passing a 3D image:
+    image3d = np.empty((5, 5, 5))
+    assert_raises(ValueError, convex_hull_object, image3d)
 
 if __name__ == "__main__":
     np.testing.run_module_suite()