videos_to_colmap.py 12 KB
Newer Older
Clement Pinard's avatar
Clement Pinard committed
1
2
3
4
5
6
from colmap_util import read_model as rm, database as db
import anafi_metadata as am
from wrappers import FFMpeg, PDraw
from sklearn.cluster import KMeans
from sklearn.metrics import pairwise_distances_argmin_min
from argparse import ArgumentParser, ArgumentDefaultsHelpFormatter
nicolas's avatar
nicolas committed
7
from edit_exif import set_gps_location
Clement Pinard's avatar
Clement Pinard committed
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
from path import Path
import pandas as pd
import numpy as np
from pyproj import Proj
from tqdm import tqdm

parser = ArgumentParser(description='Take all the drone videos of a folder and put the frame '
                                    'location in a COLMAP file for vizualisation',
                        formatter_class=ArgumentDefaultsHelpFormatter)

parser.add_argument('--video_folder', metavar='DIR',
                    help='path to videos', type=Path)
parser.add_argument('--system', default='epsg:2154')
parser.add_argument('--centroid_path', default=None)
parser.add_argument('--output_folder', metavar='DIR', type=Path)
parser.add_argument('--workspace', metavar='DIR', type=Path)
parser.add_argument('--image_path', metavar='DIR', type=Path)
parser.add_argument('--output_format', metavar='EXT', default="bin")
parser.add_argument('--vid_ext', nargs='+', default=[".mp4", ".MP4"])
parser.add_argument('--pic_ext', nargs='+', default=[".jpg", ".JPG", ".png", ".PNG"])
parser.add_argument('--nw', default='',
                    help="native-wrapper.sh file location")
parser.add_argument('--fps', default=1, type=int,
                    help="framerate at which videos will be scanned WITH reconstruction")
parser.add_argument('--num_frames', default=200, type=int)
parser.add_argument('--orientation_weight', default=1, type=float)
parser.add_argument('--resolution_weight', default=1, type=float)
parser.add_argument('--num_neighbours', default=10, type=int)
parser.add_argument('--save_space', action="store_true")


def world_coord_from_frame(frame_qvec, frame_tvec):
    '''
    frame_qvec is written in the NED system (north east down)
    frame_tvec is already is the world system (east norht up)
    '''
    world2NED = np.float32([[0, 1, 0],
                            [1, 0, 0],
                            [0, 0, -1]])
    NED2cam = np.float32([[0, 1, 0],
                          [0, 0, 1],
                          [1, 0, 0]])
    world2cam = NED2cam @ rm.qvec2rotmat(frame_qvec).T @ world2NED
    cam_tvec = - world2cam  @ frame_tvec
    cam_qvec = rm.rotmat2qvec(world2cam)
    return cam_qvec, cam_tvec


nicolas's avatar
nicolas committed
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
def set_gps(frames_list, metadata, image_path):
    for frame in frames_list:
        relative = str(frame.relpath(image_path))
        row = metadata[metadata["image_path"] == relative]
        if len(row) > 0:
            row = row.iloc[0]
            set_gps_location(frame,
                             lat=row["location_latitude"],
                             lng=row["location_longitude"],
                             altitude=row["location_altitude"])


def get_georef(metadata):
    relevant_data = metadata[["location_valid", "image_path", "x", "y", "z"]]
    path_list = []
    georef_list = []
    for _, (gps, path, x, y, alt) in relevant_data.iterrows():
        path_list.append(path)
        if gps == 1:
            georef_list.append("{} {} {} {}\n".format(path, x, y, alt))
    return georef_list, path_list


Clement Pinard's avatar
Clement Pinard committed
79
80
def optimal_sample(metadata, num_frames, orientation_weight, resolution_weight):
    metadata["sampled"] = False
nicolas's avatar
nicolas committed
81
    XYZ = metadata[["x", "y", "z"]].values
Clement Pinard's avatar
Clement Pinard committed
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
    axis_angle = metadata[["frame_quat_x", "frame_quat_y", "frame_quat_z"]].values
    if True in metadata["indoor"].unique():
        diameter = (XYZ.max(axis=0) - XYZ.min(axis=0))
        videos = metadata.loc[metadata["indoor"]]["video"].unique()
        new_centroids = 2 * diameter * np.linspace(0, 10, len(videos)).reshape(-1, 1)
        for centroid, v in zip(new_centroids, videos):
            video_index = (metadata["video"] == v).values
            XYZ[video_index] += centroid

    frame_size = metadata["video_quality"].values
    weighted_point_cloud = np.concatenate([XYZ, orientation_weight * axis_angle], axis=1)

    if resolution_weight == 0:
        weights = None
    else:
        weights = frame_size ** resolution_weight
    km = KMeans(n_clusters=num_frames).fit(weighted_point_cloud, sample_weight=weights)
    closest, _ = pairwise_distances_argmin_min(km.cluster_centers_, weighted_point_cloud)
    metadata.at[closest, "sampled"] = True
    return metadata


def register_new_cameras(cameras_dataframe, database, camera_dict, model_name="PINHOLE"):
    camera_ids = []
    for _, (w, h, f, hfov, vfov) in cameras_dataframe.iterrows():
        fx = w / (2 * np.tan(hfov * np.pi/360))
        fy = h / (2 * np.tan(vfov * np.pi/360))
        params = np.array([fx, fy, w/2, h/2])
        model_id = rm.CAMERA_MODEL_NAMES[model_name].model_id
        db_id = database.add_camera(model_id, w, h, params, prior_focal_length=True)
        camera_ids.append(db_id)
        camera_dict[db_id] = rm.Camera(id=db_id,
                                       model=model_name,
                                       width=int(w),
                                       height=int(h),
                                       params=params)
    ids_series = pd.Series(camera_ids)
    return cameras_dataframe.set_index(ids_series)


def process_video_folder(videos_list, existing_pictures, output_video_folder, image_path, system, centroid,
                         workspace, fps=1, total_frames=500, orientation_weight=1, resolution_weight=1,
                         output_colmap_format="bin", save_space=False, **env):
    proj = Proj(system)
    indoor_videos = []
    final_metadata = []
    video_output_folders = {}
    images = {}
    colmap_cameras = {}
    database_filepath = workspace/"thorough_scan.db"
    path_lists_output = {}
    database = db.COLMAPDatabase.connect(database_filepath)
    database.create_tables()
    to_extract = total_frames - len(existing_pictures)

    print("extracting metadata for {} videos...".format(len(videos_list)))
    for v in tqdm(videos_list):
        width, height, framerate = env["ffmpeg"].get_size_and_framerate(v)
        video_output_folder = output_video_folder / "{}x{}".format(width, height) / v.namebase
        video_output_folder.makedirs_p()
        video_output_folders[v] = video_output_folder

        metadata = am.extract_metadata(v.parent, v, env["pdraw"], proj,
                                       width, height, framerate, centroid)
        final_metadata.append(metadata)
        if metadata["indoor"].iloc[0]:
            indoor_videos.append(v)
    final_metadata = pd.concat(final_metadata, ignore_index=True)
    print("{} outdoor videos".format(len(videos_list) - len(indoor_videos)))
    print("{} indoor videos".format(len(indoor_videos)))

    print("{} frames in total".format(len(final_metadata)))

    cam_fields = ["width", "height", "framerate", "picture_hfov", "picture_vfov"]
    cameras_dataframe = final_metadata[cam_fields].drop_duplicates()
    cameras_dataframe = register_new_cameras(cameras_dataframe, database, colmap_cameras, "PINHOLE")
nicolas's avatar
nicolas committed
158
    print("Cameras : ")
Clement Pinard's avatar
Clement Pinard committed
159
160
161
162
163
164
165
166
    print(cameras_dataframe)
    final_metadata["camera_id"] = 0
    for cam_id, row in cameras_dataframe.iterrows():
        final_metadata.loc[(final_metadata[cam_fields] == row).all(axis=1), "camera_id"] = cam_id
    if any(final_metadata["camera_id"] == 0):
        print("Error")
        print((final_metadata["camera_id"] == 0))

Clément Pinard's avatar
Clément Pinard committed
167
168
169
    if to_extract <= 0:
        final_metadata["sampled"] = False
    elif to_extract < len(final_metadata):
Clement Pinard's avatar
Clement Pinard committed
170
171
172
173
174
175
        print("subsampling based on K-Means, to get {}"
              " frames from videos, for a total of {} frames".format(to_extract, total_frames))
        final_metadata = optimal_sample(final_metadata, total_frames - len(existing_pictures),
                                        orientation_weight,
                                        resolution_weight)
        print("Done.")
Clément Pinard's avatar
Clément Pinard committed
176
177
    else:
        final_metadata["sampled"] = True
Clement Pinard's avatar
Clement Pinard committed
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194

    print("Constructing COLMAP model with {:,} frames".format(len(final_metadata[final_metadata["sampled"]])))

    final_metadata["image_path"] = ""
    for image_id, row in tqdm(final_metadata.iterrows(), total=len(final_metadata)):
        video = row["video"]
        frame = row["frame"]
        camera_id = row["camera_id"]
        current_image_path = video_output_folders[video].relpath(image_path) / video.namebase + "_{:05d}.jpg".format(frame)

        final_metadata.at[image_id, "image_path"] = current_image_path

        if row["sampled"]:
            frame_qvec = row[["frame_quat_w",
                              "frame_quat_x",
                              "frame_quat_y",
                              "frame_quat_z"]].values
nicolas's avatar
nicolas committed
195
196
            x, y, z = row[["x", "y", "z"]]
            frame_tvec = np.array([x, y, z])
Clement Pinard's avatar
Clement Pinard committed
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
            if row["location_valid"]:
                frame_gps = row[["location_longitude", "location_latitude", "location_altitude"]]
            else:
                frame_gps = np.full(3, np.NaN)

            world_qvec, world_tvec = world_coord_from_frame(frame_qvec, frame_tvec)
            db_image_id = database.add_image(current_image_path, int(camera_id), prior_t=frame_gps)
            images[db_image_id] = rm.Image(
                id=db_image_id, qvec=world_qvec, tvec=world_tvec,
                camera_id=camera_id, name=current_image_path,
                xys=[], point3D_ids=[])

    database.commit()
    database.close()
    rm.write_model(colmap_cameras, images, {}, output_video_folder, "." + output_colmap_format)
    print("COLMAP model created")

nicolas's avatar
nicolas committed
214
215
216
217
    thorough_georef, thorough_paths = get_georef(final_metadata[final_metadata["sampled"]])
    path_lists_output["thorough"] = {}
    path_lists_output["thorough"]["frames"] = thorough_paths
    path_lists_output["thorough"]["georef"] = thorough_georef
Clement Pinard's avatar
Clement Pinard committed
218
219
220
221
222

    print("Extracting frames from videos")

    for v in tqdm(videos_list):
        video_metadata = final_metadata[final_metadata["video"] == v]
nicolas's avatar
nicolas committed
223
        by_time = video_metadata.set_index(pd.to_datetime(video_metadata["time"], unit="us"))
Clement Pinard's avatar
Clement Pinard committed
224
225
        video_folder = video_output_folders[v]
        video_metadata.to_csv(video_folder/"metadata.csv")
nicolas's avatar
nicolas committed
226
227
228
229
230
231
        path_lists_output[v] = {}
        video_metadata_1fps = by_time.resample("{:.3f}S".format(1/fps)).first()
        georef, frame_paths = get_georef(video_metadata_1fps)
        path_lists_output[v]["frames_lowfps"] = frame_paths
        path_lists_output[v]["georef_lowfps"] = georef
        path_lists_output[v]["frames_full"] = list(video_metadata["image_path"])
Clement Pinard's avatar
Clement Pinard committed
232
233
        if save_space:
            frame_ids = list(video_metadata[video_metadata["sampled"]]["frame"].values)
Clément Pinard's avatar
Clément Pinard committed
234
            if len(frame_ids) > 0:
nicolas's avatar
nicolas committed
235
                extracted_frames = env["ffmpeg"].extract_specific_frames(v, video_folder, frame_ids)
Clement Pinard's avatar
Clement Pinard committed
236
        else:
nicolas's avatar
nicolas committed
237
238
            extracted_frames = env["ffmpeg"].extract_images(v, video_folder)
        set_gps(extracted_frames, video_metadata, image_path)
Clement Pinard's avatar
Clement Pinard committed
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270

    return path_lists_output, video_output_folders


if __name__ == '__main__':
    args = parser.parse_args()
    env = vars(args)
    env["videos_list"] = sum((list(args.video_folder.walkfiles('*{}'.format(ext))) for ext in args.vid_ext), [])
    args.workspace.makedirs_p()
    output_video_folder = args.output_folder / "Videos"
    output_video_folder.makedirs_p()
    env["image_path"] = args.output_folder
    env["output_video_folder"] = output_video_folder
    env["existing_pictures"] = sum((list(args.output_folder.walkfiles('*{}'.format(ext))) for ext in args.pic_ext), [])
    env["pdraw"] = PDraw(args.nw, quiet=True)
    env["ffmpeg"] = FFMpeg(quiet=True)

    if args.centroid_path is not None:
        centroid = np.loadtxt(args.centroid_path)
    else:
        centroid = np.zeros(3)
    env["centroid"] = centroid
    lists, extracted_video_folders = process_video_folder(**env)

    if lists is not None:
        with open(args.output_folder/"video_frames_for_thorough_scan.txt", "w") as f:
            f.write("\n".join(lists["thorough"]))
        with open(args.output_folder/"georef.txt", "w") as f:
            f.write("\n".join(lists["georef"]))
        for v in env["videos_list"]:
            with open(extracted_video_folders[v] / "to_scan.txt", "w") as f:
                f.write("\n".join(lists[v]))