Merge main

annotation_gui_gcp/js/WebView.js

@@ -8,7 +8,7 @@ function initialize_event_source(handlers) {
         if (this.readyState == EventSource.CONNECTING) {
             console.log(`Reconnecting (readyState=${this.readyState})...`);
         } else {
-            console.log("Error has occured.", err);
+            console.log("Error has occurred.", err);
         }
     };
 

annotation_gui_gcp/lib/geometry.py

@@ -1,15 +1,17 @@
 from opensfm import dataset
+from numpy import ndarray
+from typing import Dict, Tuple
 
 
-def get_all_track_observations(gcp_database, track_id):
+def get_all_track_observations(gcp_database, track_id: str) -> Dict[str, ndarray]:
     print(f"Getting all observations of track {track_id}")
     data = dataset.DataSet(gcp_database.path)
     tracks_manager = data.load_tracks_manager()
     track_obs = tracks_manager.get_track_observations(track_id)
     return {shot_id: obs.point for shot_id, obs in track_obs.items()}
 
 
-def get_tracks_visible_in_image(gcp_database, image_key, min_len=5):
+def get_tracks_visible_in_image(gcp_database, image_key, min_len: int=5) -> Dict[str, Tuple[ndarray, int]]:
     print(f"Getting track observations visible in {image_key}")
     data = dataset.DataSet(gcp_database.path)
     tracks_manager = data.load_tracks_manager()

annotation_gui_gcp/lib/views/cad_view.py

@@ -1,17 +1,17 @@
 import json
 import logging
 from pathlib import Path
-from typing import Dict, Any
+from typing import Any, Dict, Tuple
 
 import rasterio
 from annotation_gui_gcp.lib.views.web_view import WebView, distinct_colors
 from flask import send_file
 from PIL import ImageColor
 
-logger = logging.getLogger(__name__)
+logger: logging.Logger = logging.getLogger(__name__)
 
 
-def _load_georeference_metadata(path_cad_model):
+def _load_georeference_metadata(path_cad_model) -> Dict[str, Any]:
     path_metadata = path_cad_model.with_suffix(".json")
 
     if not path_metadata.is_file():
@@ -30,7 +30,7 @@ def __init__(
         route_prefix,
         path_cad_file,
         is_geo_reference=False,
-    ):
+    )-> None:
         super().__init__(main_ui, web_app, route_prefix)
 
         self.main_ui = main_ui
@@ -47,7 +47,7 @@ def __init__(
             view_func=self.get_model,
         )
 
-    def get_model(self):
+    def get_model(self) -> Any:
         return send_file(self.cad_path, mimetype="application/octet-stream")
 
     def process_client_message(self, data: Dict[str, Any]) -> None:
@@ -59,7 +59,7 @@ def process_client_message(self, data: Dict[str, Any]) -> None:
         else:
             raise ValueError(f"Unknown event {event}")
 
-    def add_remove_update_point_observation(self, point_coordinates=None):
+    def add_remove_update_point_observation(self, point_coordinates=None)->None:
         gcp_manager = self.main_ui.gcp_manager
         active_gcp = self.main_ui.curr_point
         if active_gcp is None:
@@ -72,12 +72,9 @@ def add_remove_update_point_observation(self, point_coordinates=None):
         )
 
         # Add the new observation
-        if point_coordinates is not None:
-            lla = (
-                self.xyz_to_latlon(*point_coordinates)
-                if self.is_geo_reference
-                else None
-            )
+        if point_coordinates is not None and self.is_geo_reference is not None:
+            lla = self.xyz_to_latlon(*point_coordinates)
+
             geo = {
                 "latitude": lla[0],
                 "longitude": lla[1],
@@ -97,31 +94,31 @@ def add_remove_update_point_observation(self, point_coordinates=None):
             )
         self.main_ui.populate_gcp_list()
 
-    def display_points(self):
+    def display_points(self) -> None:
         pass
 
-    def refocus(self, lat, lon):
+    def refocus(self, lat, lon)->None:
         x, y, z = self.latlon_to_xyz(lat, lon)
         self.send_sse_message(
             {"x": x, "y": y, "z": z},
             event_type="move_camera",
         )
 
-    def highlight_gcp_reprojection(self, *args, **kwargs):
+    def highlight_gcp_reprojection(self, *args, **kwargs)->None:
         pass
 
-    def populate_image_list(self, *args, **kwargs):
+    def populate_image_list(self, *args, **kwargs)->None:
         pass
 
-    def latlon_to_xyz(self, lat, lon):
+    def latlon_to_xyz(self, lat, lon) -> Tuple[float, float, float]:
         xs, ys, zs = rasterio.warp.transform("EPSG:4326", self.crs, [lon], [lat], [0])
         x = xs[0] * self.scale - self.offset[0]
         y = ys[0] * self.scale - self.offset[1]
         z = zs[0] * self.scale - self.offset[2]
         y, z = z, -y
         return x, y, z
 
-    def xyz_to_latlon(self, x, y, z):
+    def xyz_to_latlon(self, x, y, z) -> Tuple[float, float, float]:
         y, z = -z, y
 
         # Add offset (cm) and transform to m
@@ -131,13 +128,13 @@ def xyz_to_latlon(self, x, y, z):
         lons, lats, alts = rasterio.warp.transform(self.crs, "EPSG:4326", [x], [y], [z])
         return lats[0], lons[0], alts[0]
 
-    def load_georeference_metadata(self, path_cad_model):
+    def load_georeference_metadata(self, path_cad_model)->None:
         metadata = _load_georeference_metadata(path_cad_model)
         self.scale = metadata["scale"]
         self.crs = metadata["crs"]
         self.offset = metadata["offset"]
 
-    def sync_to_client(self):
+    def sync_to_client(self)->None:
         """
         Sends all the data required to initialize or sync the CAD view
         """

annotation_gui_gcp/lib/views/image_view.py

@@ -104,7 +104,7 @@ def process_client_message(self, data: Dict[str, Any]) -> None:
             "add_or_update_point_observation",
             "remove_point_observation",
         ):
-            raise ValueError(f"Unknown commmand {command}")
+            raise ValueError(f"Unknown command {command}")
 
         if data["point_id"] != self.main_ui.curr_point:
             print(data["point_id"], self.main_ui.curr_point)

annotation_gui_gcp/main.py

@@ -2,7 +2,9 @@
 import json
 import typing as t
 from collections import OrderedDict, defaultdict
+from os import PathLike
 from pathlib import Path
+from typing import Union
 
 import numpy as np
 from annotation_gui_gcp.lib import GUI
@@ -12,7 +14,7 @@
 from opensfm import dataset, io
 
 
-def get_parser():
+def get_parser() -> argparse.ArgumentParser:
     parser = argparse.ArgumentParser(description=__doc__)
     parser.add_argument("dataset", help="dataset")
     parser.add_argument(
@@ -52,7 +54,7 @@ def get_parser():
     return parser
 
 
-def file_sanity_check(root, seq_dict, fname):
+def file_sanity_check(root, seq_dict, fname) -> t.Set[str]:
     # Images available under ./images for a sanity check
     available_images = {p.name for p in (root / "images").iterdir()}
     keys_in_seq_dict = {im_key for seq_keys in seq_dict.values() for im_key in seq_keys}
@@ -88,7 +90,9 @@ def load_rig_assignments(root: Path) -> t.Dict[str, t.List[str]]:
 
 
 def load_sequence_database_from_file(
-    root, fname="sequence_database.json", skip_missing=False
+    root: Path,
+    fname: Union["PathLike[str]", str] = "sequence_database.json",
+    skip_missing: bool = False,
 ):
     """
     Simply loads a sequence file and returns it.
@@ -119,7 +123,7 @@ def load_sequence_database_from_file(
     return seq_dict
 
 
-def load_shots_from_reconstructions(path, min_ims):
+def load_shots_from_reconstructions(path, min_ims) -> t.List[t.List[str]]:
     data = dataset.DataSet(path)
     reconstructions = data.load_reconstruction()
 
@@ -150,7 +154,9 @@ def load_shots_from_reconstructions(path, min_ims):
     return output
 
 
-def group_by_reconstruction(args, groups_from_sequence_database):
+def group_by_reconstruction(
+    args, groups_from_sequence_database
+) -> t.Dict[str, t.List[str]]:
     all_recs_shots = load_shots_from_reconstructions(
         args.dataset, min_ims=args.min_images_in_reconstruction
     )
@@ -173,7 +179,7 @@ def group_by_reconstruction(args, groups_from_sequence_database):
     return groups
 
 
-def group_images(args):
+def group_images(args) -> t.Dict[str, t.List[str]]:
     """
     Groups the images to be shown in different windows/views
 
@@ -203,11 +209,11 @@ def group_images(args):
         return groups_from_sequence_database
 
 
-def find_suitable_cad_paths(path_cad_files, path_dataset, n_paths=6):
+def find_suitable_cad_paths(path_cad_files: Path, path_dataset, n_paths: int = 6):
     if path_cad_files is None:
         return []
 
-    def latlon_from_meta(path_cad):
+    def latlon_from_meta(path_cad) -> t.Tuple[float, float]:
         path_meta = path_cad.with_suffix(".json")
         with open(path_meta) as f:
             meta = json.load(f)
@@ -231,7 +237,7 @@ def latlon_from_meta(path_cad):
     return [cad_files[i] for i in ixs_sort]
 
 
-def init_ui():
+def init_ui() -> t.Tuple[Flask, argparse.Namespace]:
     app = Flask(__name__)
     parser = get_parser()
     args = parser.parse_args()

bin/plot_features

@@ -48,7 +48,7 @@ if __name__ == "__main__":
         if not features_data:
             continue
         points = features_data.points
-        print("ploting {0} points".format(len(points)))
+        print("plotting {0} points".format(len(points)))
         plt.figure()
         plt.title('Image: ' + image + ', features: ' + str(len(points)))
         fig = plot_features(data.load_image(image), points)

bin/plot_inliers

@@ -241,7 +241,7 @@ def triangulate_tracks(tracks, reconstruction, graph, min_ray_angle):
     :return: An array of booleans determining if each track was successfully triangulated or not.
     """
     succeeded = []
-    triangulator = reconstruct.TrackTriangulator(graph, reconstruction)
+    triangulator = reconstruct.TrackTriangulator(reconstruction, reconstruct.TrackHandlerTrackManager(graph, reconstruction)
 
     for track in tracks:
         # Triangulate with 1 as reprojection threshold to avoid excluding tracks because of error.

bin/plot_matches.py

@@ -10,9 +10,11 @@
 from opensfm import dataset
 from opensfm import features
 from opensfm import io
+from numpy import ndarray
+from typing import List
 
 
-def plot_matches(im1, im2, p1, p2):
+def plot_matches(im1, im2, p1: ndarray, p2: ndarray) -> None:
     h1, w1, c = im1.shape
     h2, w2, c = im2.shape
     image = np.zeros((max(h1, h2), w1 + w2, 3), dtype=im1.dtype)
@@ -29,7 +31,7 @@ def plot_matches(im1, im2, p1, p2):
     pl.plot(p2[:, 0] + w1, p2[:, 1], "ob")
 
 
-def plot_graph(data):
+def plot_graph(data) -> None:
     cmap = cm.get_cmap("viridis")
     connectivity = {}
     for im1 in images:
@@ -70,7 +72,7 @@ def plot_graph(data):
     pl.savefig(os.path.join(data.data_path, "matchgraph.png"))
 
 
-def plot_matches_for_images(data, image, images):
+def plot_matches_for_images(data, image, images) -> None:
     if image:
         pairs = [(image, o) for o in images if o != image]
     elif images:
@@ -122,10 +124,10 @@ def plot_matches_for_images(data, image, images):
     parser.add_argument(
         "--save_figs", help="save figures instead of showing them", action="store_true"
     )
-    args = parser.parse_args()
+    args: argparse.Namespace = parser.parse_args()
 
     data = dataset.DataSet(args.dataset)
-    images = data.images()
+    images: List[str] = data.images()
 
     if args.graph:
         plot_graph(data)

doc/source/annotation_tool.rst

@@ -138,7 +138,7 @@ Detailed documentation for this is not available as the feature is experimental,
 - Start from a dataset containing more than one reconstruction in ``reconstruction.json``.
 
 - Launch the tool:
-    - If the two reconstructions come from different sequences, lauch as usual.
+    - If the two reconstructions come from different sequences, launch as usual.
     - If the two reconstructions come from the same sequence, launch using the ``--group-by-reconstruction`` argument.
       This will split the images into two windows, one for each reconstruction.
 
@@ -150,7 +150,7 @@ Use the 'Rigid', 'Flex' or 'Full' buttons to run the alignment using the annotat
 - The 'Rigid' option triangulates the control points in each reconstruction independently and finds a rigid transform to align them.
 - The 'Flex' option additionally re-runs bundle adjustment, allowing for some deformation of both reconstructions to fit the annotations.
 - The 'Full' option attempts to obtain positional covariances for each camera pose.
-  If succesful, the frame list on the image views is populated with the positional covariance norm. Lower is better.
+  If successful, the frame list on the image views is populated with the positional covariance norm. Lower is better.
 
 After running analysis, the reprojection errors are overlaid on the image views as shown in :ref:`running-alignment`.
 The aligned reconstructions are saved with new filenames in the root folder and can be viewed in 3D with the OpenSfM viewer.

doc/source/dense.rst

@@ -123,7 +123,7 @@ Undistortion
 
 The dense module assumes that images are taken with perspective projection and no radial distortion.  For perspective images, undistorted versions can be generated by taking into account the computed distortion parameters, :math:`k1` and :math:`k2`.
 
-Spherical images (360 panoramas) however can not be unwarped into a single persepective view.  We need to generate multiple perspective views to cover the field of view of a panorama.
+Spherical images (360 panoramas) however can not be unwarped into a single perspective view.  We need to generate multiple perspective views to cover the field of view of a panorama.
 
 This means that the undistortion process will create new views of the reconstruction.  Thus the undistortion process is one where a reconstruction is taken as input and a new reconstruction is produced as output.  The input may contain radially distorted images and panoramas and the output reconstruction will only have undistorted perspective images.
 

doc/source/gcp.rst

@@ -6,7 +6,7 @@ When EXIF data contains GPS location, it is used by OpenSfM to georeference the
 
 Ground control points (GCP) are landmarks visible on the images for which the geospatial position (latitude, longitude and altitude) is known.  A single GCP can be observed in one or more images.
 
-OpenSfM uses GCP in two steps of the reconstruction process: alignment and bundle adjustment.  In the alignment step, points are used to globaly move the reconstruction so that the observed GCP align with their GPS position.  Two or more observations for each GCP are required for it to be used during the aligment step.
+OpenSfM uses GCP in two steps of the reconstruction process: alignment and bundle adjustment.  In the alignment step, points are used to globaly move the reconstruction so that the observed GCP align with their GPS position.  Two or more observations for each GCP are required for it to be used during the alignment step.
 
 In the bundle adjustment step, GCP observations are used as a constraint to refine the reconstruction.  In this step, all ground control points are used.  No minimum number of observation is required.
 

doc/source/geometry.rst

@@ -144,8 +144,8 @@ Identifier `brown`
     y_n = \frac{y}{z} \\
     r^2 = x_n^2 + y_n^2 \\
     d_r = 1 + k_1 r^2 + k_2 r^4 + k_3 r^6\\
-    d^t_x = 2p_1\ x_n\ y_n + p_2\ (r^2 + 2x)\\
-    d^t_y = 2p_2\ x_n\ y_n + p_1\ (r^2 + 2y)\\
+    d^t_x = 2p_1\ x_n\ y_n + p_2\ (r^2 + 2x^2)\\
+    d^t_y = 2p_2\ x_n\ y_n + p_1\ (r^2 + 2y^2)\\
     u = f_x\ (d_r\ x_n + d^t_x) + c_x \\
     v = f_y\ (d_r\ y_n + d^t_y) + c_y
     \end{array}

doc/source/index.rst

@@ -18,6 +18,7 @@ OpenSfM
    rig
    annotation_tool
    api
+   sensor_database
 
 Indices and tables
 ==================

doc/source/quality_report.rst

@@ -24,11 +24,11 @@ Processing Summary
 |processing|
 
  - Reconstructed Images : reconstructed images over total number of images
- - Reconstructed Images : reconstructed points over total number of points in the `tracks.csv` file (`create_tracks`)
+ - Reconstructed Points : reconstructed points over total number of points in the `tracks.csv` file (`create_tracks`)
  - Reconstructed Components : number of continously reconstructed sets of images
- - Detected Features : median number (accross images) of detected features
- - Reconstructed Features : median number (accross images) of reconstructed features
- - Geographic Reference : indicated wether GPS and/or GCP have been used for geo-alignment
+ - Detected Features : median number (across images) of detected features
+ - Reconstructed Features : median number (across images) of reconstructed features
+ - Geographic Reference : indicated whether GPS and/or GCP have been used for geo-alignment
  - GPS / GCP errors : GPS and/or GCP RMS errors
 
 |topview|
@@ -45,7 +45,7 @@ The heatmap represent the density of detected features : the gradient goes from
 
 |feat2|
 
-The table below lists minimum/maximum/mean and median detected and reconstructed features accross images.
+The table below lists minimum/maximum/mean and median detected and reconstructed features across images.
 
 Reconstruction Details
 ~~~~~~~~~~~~~~~~~~~~~~

doc/source/reporting.rst

@@ -4,7 +4,7 @@
 Reporting
 =========
 
-OpenSfM commands write reports on the work done.  Reports are stored in the ``reports`` folder in json format so that they can be loaded programatically.  Here is the list of reports produced and the data included.
+OpenSfM commands write reports on the work done.  Reports are stored in the ``reports`` folder in json format so that they can be loaded programmatically.  Here is the list of reports produced and the data included.
 
 Feature detection
 -----------------