yRenaming to differentiate between cpu and gpu implementations (262cdac2) · Commits · aflab / astrogeology / Autocnet

autocnet/matcher/cpu_decompose.py

+3 −3

Changes for autocnet/matcher/cpu_decompose.py: 3 added lines, 3 removed lines.
Original line number	Diff line number	Diff line
		import numpy as np
		from scipy.spatial.distance import cdist

		from autocnet.matcher.feature import FlannMatcher
		from autocnet.matcher.feature_matcher import match
		from autocnet.matcher.cpu_matcher import FlannMatcher
		from autocnet.matcher.cpu_matcher import match
		from autocnet.transformation.decompose import coupled_decomposition


		@@ -203,6 +203,6 @@ def decompose_and_match(self, k=2, maxiteration=3, size=18, buf_dist=3,**kwargs)
		sidx = skp.query('x >= {} and x <= {} and y >= {} and y <= {}'.format(minsx, maxsx, minsy, maxsy)).index
		didx = dkp.query('x >= {} and x <= {} and y >= {} and y <= {}'.format(mindx, maxdx, mindy, maxdy)).index
		# If the candidates < k, OpenCV throws an error
		if len(sidx) >= k and len(didx) >=k:
		if len(sidx) > k and len(didx) > k:
		match(self, aidx=sidx, bidx=didx)
		match(self, aidx=didx, bidx=sidx)

autocnet/matcher/feature_extractor.py→autocnet/matcher/cpu_extractor.py

+5 −5

Changes for autocnet/matcher/cpu_extractor.py: 5 added lines, 5 removed lines.
Original line number	Diff line number	Diff line
		@@ -14,7 +14,7 @@ except Exception: # pragma: no cover
		pass


		def extract_features(array, method='orb', extractor_parameters={}):
		def extract_features(array, extractor_method='sift', extractor_parameters={}):
		"""
		This method finds and extracts features from an image using the given dictionary of keyword arguments.
		The input image is represented as NumPy array and the output features are represented as keypoint IDs
		@@ -25,7 +25,7 @@ def extract_features(array, method='orb', extractor_parameters={}):
		array : ndarray
		a NumPy array that represents an image

		method : {'orb', 'sift', 'fast', 'surf', 'vl_sift'}
		extractor_method : {'orb', 'sift', 'fast', 'surf', 'vl_sift'}
		The detector method to be used. Note that vl_sift requires that
		vlfeat and cyvlfeat dependencies be installed.

		@@ -45,10 +45,10 @@ def extract_features(array, method='orb', extractor_parameters={}):
		'surf': cv2.xfeatures2d.SURF_create,
		'orb': cv2.ORB_create}

		if method == 'vlfeat' and vlfeat != True:
		if extractor_method == 'vlfeat' and vlfeat != True:
		raise ImportError('VLFeat is not available. Please install vlfeat or use a different extractor.')

		if method == 'vlfeat':
		if extractor_method == 'vlfeat':
		keypoint_objs, descriptors = vl.sift.sift(array,
		compute_descriptor=True,
		float_descriptors=True)
		@@ -59,7 +59,7 @@ def extract_features(array, method='orb', extractor_parameters={}):
		# OpenCV requires the input images to be 8-bit
		if not array.dtype == 'int8':
		array = bytescale(array)
		detector = detectors[method](**extractor_parameters)
		detector = detectors[extractor_method](**extractor_parameters)
		keypoint_objs, descriptors = detector.detectAndCompute(array, None)

		keypoints = np.empty((len(keypoint_objs), 7), dtype=np.float32)

autocnet/matcher/feature.py→autocnet/matcher/cpu_matcher.py

+75 −3

Changes for autocnet/matcher/cpu_matcher.py: 75 added lines, 3 removed lines.
Original line number	Diff line number	Diff line
		@@ -8,6 +8,81 @@ FLANN_INDEX_KDTREE = 1 # Algorithm to set centers,
		DEFAULT_FLANN_PARAMETERS = dict(algorithm=FLANN_INDEX_KDTREE, trees=3)


		def match(self, k=2, **kwargs):
		"""
		Given two sets of descriptors, utilize a FLANN (Approximate Nearest
		Neighbor KDTree) matcher to find the k nearest matches. Nearness is
		the euclidean distance between descriptors.

		The matches are then added as an attribute to the edge object.

		Parameters
		----------
		k : int
		The number of neighbors to find
		"""

		def _add_matches(matches):
		"""
		Given a dataframe of matches, either append to an existing
		matches edge attribute or initially populate said attribute.

		Parameters
		----------
		matches : dataframe
		A dataframe of matches
		"""
		if self.matches is None:
		self.matches = matches
		else:
		df = self.matches
		self.matches = df.append(matches,
		ignore_index=True,
		verify_integrity=True)

		def mono_matches(a, b, aidx=None, bidx=None):
		"""
		Apply the FLANN match_features

		Parameters
		----------
		a : object
		A node object

		b : object
		A node object

		aidx : iterable
		An index for the descriptors to subset

		bidx : iterable
		An index for the descriptors to subset
		"""
		# Subset if requested
		if aidx is not None:
		ad = a.descriptors[aidx]
		else:
		ad = a.descriptors

		if bidx is not None:
		bd = b.descriptors[bidx]
		else:
		bd = b.descriptors

		# Load, train, and match
		fl.add(ad, a['node_id'], index=aidx)
		fl.train()
		matches = fl.query(bd, b['node_id'], k, index=bidx)
		_add_matches(matches)
		fl.clear()

		fl = FlannMatcher()
		mono_matches(self.source, self.destination, **kwargs)
		mono_matches(self.destination, self.source, **kwargs)

		self.matches.sort_values(by=['distance'])


		class FlannMatcher(object):
		"""
		A wrapper to the OpenCV Flann based matcher class that adds
		@@ -122,6 +197,3 @@ class FlannMatcher(object):
		return pd.DataFrame(matched, columns=['source_image', 'source_idx',
		'destination_image', 'destination_idx',
		'distance']).astype(np.float32)

		def cudamatcher():
		pass

autocnet/matcher/outlier_detector.py→autocnet/matcher/cpu_outlier_detector.py

+0 −0

File moved.

View file

autocnet/matcher/cuda_extractor.py

+3 −0

Changes for autocnet/matcher/cuda_extractor.py: 3 added lines, 0 removed lines.
Original line number	Diff line number	Diff line
		@@ -3,6 +3,9 @@ import warnings
		import cudasift as cs

		def extract_features(array, nfeatures=None):
		"""
		A custom docstring.
		"""
		if not nfeatures:
		nfeatures = int(max(array.shape) / 1.75)
		else: