[refactor] pass sift descriptors through memory (possible since merge…

… of PR MISS3D#193)

s2p/sift.py

@@ -23,9 +23,10 @@
 
 # TODO: This is kind of ugly. Cleaner way to do this is to update
 # LD_LIBRARY_PATH, which we should do once we have a proper config file
-sift4ctypes_library = os.path.join(os.path.dirname(
-    os.path.abspath(__file__)), '../lib/libsift4ctypes.so')
-ctypes.CDLL(sift4ctypes_library)
+here = os.path.dirname(os.path.abspath(__file__))
+sift4ctypes = os.path.join(os.path.dirname(here), 'lib', 'libsift4ctypes.so')
+lib = ctypes.CDLL(sift4ctypes)
+
 
 # Filter warnings from rasterio reading files wihtout georeferencing
 warnings.filterwarnings("ignore", category=rio.errors.NotGeoreferencedWarning)
@@ -51,9 +52,6 @@ def keypoints_from_nparray(arr, thresh_dog=0.0133, nb_octaves=8, nb_scales=3, of
     # retrieve numpy buffer dimensions
     h, w = arr.shape
 
-    # load shared library
-    lib = ctypes.CDLL(sift4ctypes_library)
-
     # Set expected args and return types
     lib.sift.argtypes = (ndpointer(dtype=ctypes.c_float, shape=(h, w)), ctypes.c_uint, ctypes.c_uint, ctypes.c_float,
                          ctypes.c_uint, ctypes.c_uint, ctypes.POINTER(ctypes.c_uint), ctypes.POINTER(ctypes.c_uint))
@@ -69,7 +67,7 @@ def keypoints_from_nparray(arr, thresh_dog=0.0133, nb_octaves=8, nb_scales=3, of
 
     # Transform result into a numpy array
     keypoints = np.asarray([keypoints_ptr[i]
-                            for i in range(0, nb_points.value*desc_size.value)])
+                            for i in range(nb_points.value*desc_size.value)])
 
     # Delete results to release memory
     lib.delete_buffer.argtypes = (ctypes.POINTER(ctypes.c_float)),
@@ -94,12 +92,12 @@ def image_keypoints(im, x, y, w, h, max_nb=None, thresh_dog=0.0133, nb_octaves=8
     http://www.ipol.im/pub/pre/82/
 
     Args:
-        im: path to the input image
+        im (str): path to the input image
         max_nb (optional): maximal number of keypoints. If more keypoints are
             detected, those at smallest scales are discarded
 
     Returns:
-        path to the file containing the list of descriptors
+        numpy array of shape (n, 132) containing, on each row: (y, x, s, o, 128-descriptor)
     """
     # Read file with rasterio
     with rio.open(im) as ds:
@@ -124,10 +122,7 @@ def image_keypoints(im, x, y, w, h, max_nb=None, thresh_dog=0.0133, nb_octaves=8
     if max_nb is not None:
         keypoints = keypoints[:max_nb]
 
-    keyfile = common.tmpfile('.txt')
-    np.savetxt(keyfile, keypoints, delimiter=' ', fmt='%.3f')
-
-    return keyfile
+    return keypoints
 
 
 def keypoints_match(k1, k2, method='relative', sift_thresh=0.6, F=None,
@@ -136,7 +131,10 @@ def keypoints_match(k1, k2, method='relative', sift_thresh=0.6, F=None,
     Find matches among two lists of sift keypoints.
 
     Args:
-        k1, k2: paths to text files containing the lists of sift descriptors
+        k1 (array): numpy array of shape (n, 132), where each row represents a
+            sift keypoint with (y, x, scale, orientation, 128-descriptor)
+        k2 (array): numpy array of shape (m, 132), where each row represents a
+            sift keypoint
         method (optional, default is 'relative'): flag ('relative' or
             'absolute') indicating wether to use absolute distance or relative
             distance
@@ -157,29 +155,14 @@ def keypoints_match(k1, k2, method='relative', sift_thresh=0.6, F=None,
         if any, a numpy 2D array containing the list of inliers matches.
     """
     # compute matches
-    # Open k1 and k2 files containing the two lists of SIFTs
-    k1_keys_descriptors = list()
-    with open(k1) as f:
-        for line in f:
-            ll = [float(val) for val in line.split(' ')[:-1]]
-            k1_keys_descriptors.append(ll)
-    k2_keys_descriptors = list()
-    with open(k2) as f:
-        for line in f:
-            ll = [float(val) for val in line.split(' ')[:-1]]
-            k2_keys_descriptors.append(ll)
-    k1_keys_descriptors = np.asarray(k1_keys_descriptors, dtype=np.float32)
-    k2_keys_descriptors = np.asarray(k2_keys_descriptors, dtype=np.float32)
-
-    matches = keypoints_match_from_nparray(k1_keys_descriptors, k2_keys_descriptors, method, sift_thresh,
+    matches = keypoints_match_from_nparray(k1, k2, method, sift_thresh,
                                            epipolar_threshold, F)
 
     # Write this to file
     mfile = common.tmpfile('.txt')
     with open(mfile, 'w') as f:
         np.savetxt(f, matches, delimiter=' ')
 
-    matches = np.loadtxt(mfile)
     if matches.ndim == 2:  # filter outliers with ransac
         if model == 'fundamental' and len(matches) >= 7:
             common.run("ransac fmn 1000 .3 7 %s < %s" % (mfile, mfile))
@@ -196,9 +179,8 @@ def keypoints_match(k1, k2, method='relative', sift_thresh=0.6, F=None,
 
 
 def keypoints_match_from_nparray(k1, k2, method, sift_threshold, epi_threshold, F):
-    # load shared library
-    lib = ctypes.CDLL(sift4ctypes_library)
-
+    """
+    """
     # Set expected args and return types
     lib.matching.argtypes = (ndpointer(dtype=ctypes.c_float, shape=k1.shape),
                              ndpointer(dtype=ctypes.c_float, shape=k2.shape),
@@ -215,9 +197,7 @@ def keypoints_match_from_nparray(k1, k2, method, sift_threshold, epi_threshold,
     length_descr = descr - sift_offset
 
     # Transform information of method into boolean
-    use_relative_method = False
-    if method == 'relative':
-        use_relative_method = True
+    use_relative_method = (method == 'relative')
 
     # Format fundamental matrix
     use_fundamental_matrix = False
@@ -230,24 +210,21 @@ def keypoints_match_from_nparray(k1, k2, method, sift_threshold, epi_threshold,
     nb_matches = ctypes.c_uint()
 
     # Call sift fonction from sift4ctypes.so
-    matches_ptr = lib.matching(k1, k2, length_descr, sift_offset, nb_sift_k1, len(k2),
-                 sift_threshold, epi_threshold,
-                 coeff_mat, use_fundamental_matrix,
-                 use_relative_method,
-                 ctypes.byref(nb_matches)
-                 )
+    matches_ptr = lib.matching(k1.astype('float32'), k2.astype('float32'),
+                               length_descr, sift_offset, len(k1), len(k2),
+                               sift_threshold, epi_threshold, coeff_mat,
+                               use_fundamental_matrix, use_relative_method,
+                               ctypes.byref(nb_matches))
 
     # Transform result into a numpy array
-    matches = np.asarray([matches_ptr[i] for i in range(0, nb_matches.value * 4)])
+    matches = np.asarray([matches_ptr[i] for i in range(nb_matches.value * 4)])
 
     # Delete results to release memory
-    lib.delete_buffer.argtypes = (ctypes.POINTER(ctypes.c_float)),
+    lib.delete_buffer.argtypes = ctypes.POINTER(ctypes.c_float),
     lib.delete_buffer(matches_ptr)
 
     # Reshape keypoints array
-    matches = matches.reshape((nb_matches.value, 4))
-
-    return matches
+    return matches.reshape((nb_matches.value, 4))
 
 
 def matches_on_rpc_roi(im1, im2, rpc1, rpc2, x, y, w, h):

tests/test_s2p.py

@@ -40,9 +40,7 @@ def unit_gdal_version():
 
 def unit_image_keypoints():
 
-    kpts = s2p.sift.image_keypoints('tests/data/input_triplet/img_02.tif', 100, 100, 200, 200)
-
-    test_kpts = np.loadtxt(kpts)
+    test_kpts = s2p.sift.image_keypoints('tests/data/input_triplet/img_02.tif', 100, 100, 200, 200)
     ref_kpts  = np.loadtxt('tests/data/expected_output/units/unit_image_keypoints.txt')
 
     test_set = set(map(tuple,test_kpts[:,0:2]))
@@ -90,7 +88,9 @@ def unit_image_keypoints():
 
 def unit_matching():
 
-    test_matches = s2p.sift.keypoints_match('tests/data/units/sift1.txt','tests/data/units/sift2.txt')
+    k1 = np.loadtxt('tests/data/units/sift1.txt')
+    k2 = np.loadtxt('tests/data/units/sift2.txt')
+    test_matches = s2p.sift.keypoints_match(k1, k2)
     expected_matches = np.loadtxt('tests/data/expected_output/units/unit_keypoints_match.txt')
 
     # Check that numbers of matches are the same