Adds the ability to get raw keypoints as an ndarray (429e50d1) · Commits · aflab / astrogeology / Autocnet

autocnet/graph/node.py

+23 −23

Original line number	Diff line number	Diff line
		@@ -164,20 +164,6 @@ class Node(dict, MutableMapping):
		else:
		return 0

		""" @property
		def keypoints(self):
		if hasattr(self, '_keypoints'):
		return self._keypoints.copy()
		else:
		return None

		@property
		def descriptors(self):
		if hasattr(self, '_descriptors'):
		return np.copy(self._descriptors)
		else:
		return None"""

		def coverage(self):
		"""
		Determines the area of keypoint coverage
		@@ -249,7 +235,7 @@ class Node(dict, MutableMapping):
		"""
		if hasattr(self, '_keypoints'):
		if index is not None:
		return self._keypoints.loc[index]
		return self._keypoints.ix[index]
		else:
		return self._keypoints

		@@ -278,6 +264,15 @@ class Node(dict, MutableMapping):

		return keypoints

		def get_raw_keypoint_coordinates(self, index):
		"""
		The performance of get_keypoint_coordinates can be slow
		due to the ability for fancier indexing. This method
		returns coordinates using numpy array accessors.
		"""
		index = index.astype(np.int)
		return self._keypoints.values[index,:2]

		@staticmethod
		def _extract_features(args, *kwargs):
		"""
		@@ -391,23 +386,28 @@ class Node(dict, MutableMapping):

		# Add the point object onto the node
		point = Point(pid)

		covered_edges = list(map(tuple, g[['source_image', 'destination_image']].values))
		#print(g[['source_image', 'destination_image']])
		#covered_edges = list(map(tuple, g[['source_image', 'destination_image']].values))
		s = g['source_image'].iat[0]
		d = g['destination_image'].iat[0]
		# The reference edge that we are deepening with
		ab = cg.edge[covered_edges[0][0]][covered_edges[0][1]]
		ab = cg.edge[s][d]

		# Get the coordinates of the search correspondence
		ab_keypoints = ab.source.get_keypoint_coordinates(index=g['source_idx'])
		ab_keypoints = ab.source.get_raw_keypoint_coordinates(index=g['source_idx'])
		ab_x = None

		for j, (r_idx, r) in enumerate(g.iterrows()):
		kp = ab_keypoints.iloc[j].values

		if len(g) == 1:
		kp = ab_keypoints
		else:
		kp = ab_keypoints[j]
		# Homogenize the coord used for epipolar projection
		if ab_x is None:
		ab_x = np.array([kp[0], kp[1], 1.])

		kpd = ab.destination.get_keypoint_coordinates(index=g['destination_idx']).values[0]
		kpd = ab.destination.get_raw_keypoint_coordinates(index=g['destination_idx'])
		if len(kpd.shape) > 1:
		kpd = kpd[0]
		# Add the existing source and destination correspondences
		self.point_to_correspondence[point].add((r['source_image'],
		Correspondence(r['source_idx'],