fusion4landslide/main_f2s3.py at main · gseg-ethz/fusion4landslide · GitHub

1
2
3
4
5
6
7
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
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
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
158
159
160
161
162
163
164
165
166
import argparse
from utils.common import load_yaml, dir_exist, access_device, run_time
from easydict import EasyDict as edict
from cpp_core.pcd_tiling.build import pcd_tiling
from cpp_core.supervoxel_segmentation.build import supervoxel
from src.functions import point_cloud_tiling
# from src.median_res_compute import compute_median_resolution
import os.path as osp
import numpy as np
import open3d as o3d
from utils.logger import get_logger
from src.f2s3 import Deformation_Analyze
import time
import torch
import glob
from tqdm import tqdm
import gc
import os
import re
from src.models import PointNetFeature, FilteringNetwork

import warnings
warnings.simplefilter("ignore", FutureWarning)


def deformation_analysis_process(config):
    """
    The entire process of deformation analysis
    :return:
    """

    # point cloud tiling
    config.tile_dir = osp.join(config.output_dir, config.output_folder, 'tiled_data')
    if not osp.exists(config.tile_dir):
        point_cloud_tiling(config)
        if config.verbose:
            config.logging.info('Point cloud tiling is done. Save tiles to %s', config.tile_dir)
    else:
        if osp.exists(config.tile_dir):
            if config.verbose:
                config.logging.info('Skip point cloud tiling. Tiles will be loaded from: %s', config.tile_dir)
        else:
            raise NotImplementedError('Tile(s) do not exist in %s', osp.exists(config.tile_dir))


    # load tiled files
    src_tiled_overlap_list = sorted(glob.glob(osp.join(config.tile_dir, 'overlap', "source_tile_*")),
                                    key=lambda x: int(re.search(r'\d+', os.path.basename(x)).group()))
    if config.verbose:
        config.logging.info(f'Num. of tile(s) from source/target point cloud: {len(src_tiled_overlap_list)}', )

    config.device = access_device()
    load_pretrained_models(config)

    # implement a loop computation based each tile
    config.feat_dir = osp.join(config.output_dir, config.output_folder, 'features')
    if config.segment_type == 'supervoxel':
        config.segment_dir = osp.join(config.output_dir, config.output_folder, 'svl_segment')
    elif config.segment_type == 'superpoint':
        config.segment_dir = osp.join(config.output_dir, config.output_folder, 'spt_segment')
    config.corr_dir = osp.join(config.output_dir, config.output_folder, 'correspondences')
    dir_exist(config.feat_dir)
    dir_exist(config.segment_dir)
    dir_exist(config.corr_dir)
    continue_tile = 0
    for tile_i, src_tile_overlap_path in enumerate(tqdm(src_tiled_overlap_list[continue_tile:], ncols=150, desc='Load source tile overlaps')):
        tgt_tile_overlap_path = src_tile_overlap_path.replace('source_tile_', 'target_tile_')
        assert osp.exists(tgt_tile_overlap_path)

        config.tile_id = re.findall(r'\d+', os.path.basename(src_tile_overlap_path))[0]

        deformation_analysis = Deformation_Analyze(config, src_tile_overlap_path, tgt_tile_overlap_path)

        deformation_analysis.compute_features()

        deformation_analysis.implement_segmentation()

        deformation_analysis.correspondence_searching()

        if config.correspondence_pruning:
            deformation_analysis.correspondence_pruning()
        else:
            corres = deformation_analysis.correspondences
            magnitude = np.linalg.norm(corres[:, 3:6] - corres[:, :3], axis=1)
            output_path = osp.join(osp.dirname(config.corr_dir), 'results')
            np.savetxt(os.path.join(output_path, 'displacement_vector_and_magnitude_before_pruning.txt'),
                       np.hstack((corres[:, :3], magnitude[:, None])))

        # deformation_analysis.results_output()


def load_pretrained_models(config):
    # load the network of the feature descriptor
    if config.feat_compute and config.feat_type == 'DIPs':
        feat_model_path = osp.join(config.weight_dir, "local_feature_descriptor_best.pth")
        if config.verbose:
            config.logging.info('Load trained feature descriptor weight (%s) from: %s',
                              config.feat_type, feat_model_path)
        config.feat_desc_nn = PointNetFeature()
        config.feat_desc_nn.load_state_dict(torch.load(feat_model_path))
        config.feat_desc_nn.to(config.device)
        config.feat_desc_nn.eval()

    # load the network of the outlier removal
    if config.outlier_removal and config.removal_type == 'binary_classifier':
        outlier_removal_path = osp.join(config.weight_dir, 'outlier_classifier_best.pt')
        if config.verbose:
            config.logging.info('Load trained outlier removal weight (%s) from : %s',
                                config.removal_type, outlier_removal_path)
        config.outlier_removal_nn = FilteringNetwork()
        config.outlier_removal_nn.load_state_dict(torch.load(outlier_removal_path))
        config.outlier_removal_nn.to(config.device)
        config.outlier_removal_nn.eval()
    return None


def main():
    # load configs
    parser = argparse.ArgumentParser()
    parser.add_argument('--config', type=str,
                        # default='./configs/default_config.yaml',
                        # Mattertal
                        # default='./configs/landslide/f2s3_mattertal.yaml',
                        # Brienz
                        default='./configs/landslide/f2s3_brienz.yaml',
                        # Change detection
                        # default='./configs/change_detection/f2s3_fbk_landslide.yaml',
                        help='Path to config file.')
    args = parser.parse_args()
    cfg = load_yaml(args.config)

    # cfg['snapshot_dir'] = 'snapshot/%s' % cfg['exp_dir']

    # ensure the log dir exists, otherwise create a dir
    log_dir = osp.join(cfg['output_dir'], cfg['output_folder'], 'logs')
    dir_exist(log_dir)
    log_save_path = osp.join(log_dir, 'deform_analysis_{}.log'.format(time.strftime('%Y%m%d_%H%M%S')))
    cfg['logging'] = get_logger(log_save_path)

    cfg = edict(cfg)

    # print log info to the terminal
    cfg.logging.info('----------------------------------------------------------------------')
    log_message = "Config: \n"
    for key, value in cfg.items():
        log_message += f"{key}={value} \n"
    cfg.logging.info(log_message)
    cfg.logging.info('----------------------------------------------------------------------')

    with torch.no_grad():
        # recode the time spent
        start_time = time.time()

        # start deformation analysis
        deformation_analysis_process(cfg)

        end_time = time.time()

    if cfg.verbose:
        cfg.logging.info(f"Deformation analysis is done! Save log information to: '{log_save_path}'.")
        cfg.logging.info(f"Save results to: '{cfg.output_dir}/{cfg.output_folder}'. Total time taken: {(end_time - start_time)/3600:.2f}"
                         f" hours or {(end_time - start_time):.1f} seconds.")


if __name__ == '__main__':
    main()