cigChannel/tributary_channel_parallel.py at main · wanggy-1/cigChannel · 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
from functions import *


def parallel(i: int,
             xmin: float, xmax: float,
             ymin: float, ymax: float,
             zmin: float, zmax: float, zcut: float,
             dx: float, dy: float, dz: float,
             channel_params: dict, channel_database: dict,
             replacement: bool = False,
             instance_label: bool = False,
             mute: bool = True, seed_val: int = None,
             output_dir: str = './'):

    print(f'Model{i} begin')

    # Initialize the model.
    model = GeoModel(extent=[xmin, xmax, ymin, ymax, zmin, zmax],
                     resolution=[dx, dy, dz],
                     mute=mute)

    # Create Vp model.
    random.seed(seed_val)
    vp_noise = random.uniform(300, 400)
    model.add_vp(h_layer_list=[(70, 150), (70, 150), (70, 150)],
                 fm_list=[0.3, 0.7],
                 vp_list=[(2500, 3500), (3500, 5000), (5000, 6500)],
                 vp_diff_list=[(500, 800), (800, 1000), (800, 1000)],
                 vp_disturb=vp_noise,
                 smooth=False,
                 seed=seed_val,
                 mute=mute)

    # Add channels.
    model.add_tributary_channel_from_database(params=channel_params,
                                                 database=channel_database,
                                                 instance_label = instance_label,
                                                 replacement=replacement,
                                                 seed=seed_val,
                                                 mute=mute)

    # Smooth Vp model.
    model.smooth(sigma=1.5,
                 param=['vp', 'channel'],
                 mute=mute)

    # Add dipping.
    model.add_dipping(a_range=[0.01, 0.04], b_range=[0.01, 0.04], seed=seed_val, mute=mute)

    # Add folds.
    model.add_fold(N=10, sigma_range=[0.12, 0.15], A_range=[0.02, 0.04], d_fold=3000, zscale=1.0,
                    sync=True, seed=seed_val, mute=mute)

    # Add faults.
    model.add_fault(N=5,
                    theta_range=[60, 90],
                    d_max_range=[0.04, 0.06],
                    d_fault=2000,
                    seed=seed_val,
                    mute=mute)

    # Resampling.
    model.resample_z(param=['vp', 'channel'],
                     z_range=(200, 200+zcut),
                     mute=mute)

    # Compute P-wave impedance.
    model.compute_Ip(rho=2.4)

    # Compute reflection coefficients.
    model.compute_rc(mute=mute)

    # Make synthetic seismic data.
    freq = random.uniform(25, 45)
    model.make_synseis(mark_channel=True,
                       plot_wavelet=False,
                       f_ricker=freq,
                       wavelet_type='ricker',
                       length=0.1,
                       mute=mute)

    os.makedirs(output_dir, exist_ok=True)

    # Save data.
    model.Ip.tofile(output_dir+f'/Ip/{i}.dat')
    model.seismic.tofile(output_dir+f'/Seismic/{i}.dat')
    model.seis_label.tofile(output_dir+f'/Label/{i}.dat')

    print(f'Model {i} done.')


# ----------- main function ------------- #
# Inputs.
n_model = 200  # Number of models.
start_num = 0  # Start number of the models.
seed_val = None  # Random seed value.
xmin, xmax, dx = 0, 6400, 25  # xmin: minimum x-coordinate (m); xmax: maximum x-coordinate (m); dx: cell length (m).
ymin, ymax, dy = 0, 6400, 25  # ymin: minimum y-coordinate (m); ymax: maximum y-coordinate (m); dy: cell width (m).
zmin, zmax, dz = 0, 1600, 5  # zmin: minimum z-coordinate (m); zmax: maximum z-coordinate (m); dz: cell height (m).
zcut = 1280
database_dir = './Topography/Tributary/size256_dx25m'  # Channel topography database directory.
output_dir = './dataset/tributary'  # Model output directory.
channel_params = [dict(n_channel=3, z_range=[0, 1], hfill_range=[20, 20],
                       vpfill_std_range=[0, 0], epsilon_range=[0, 0]),
                  dict(n_channel=3, z_range=[0, 1], hfill_range=[20, 20],
                       vpfill_std_range=[0, 0], epsilon_range=[0, 0.5]),
                  dict(n_channel=3, z_range=[0, 1], hfill_range=[20, 20],
                       vpfill_std_range=[0, 0], epsilon_range=[0.5, 1])]  # Channel parameters.
instance_label = False
replacement=False
mute = True
n_cores = 35  # Number of CPUs used for computing.
if n_cores is None:
    n_cores = multiprocessing.cpu_count()

# Construct channel database.
channel_database = {}
channel_database['id'] = os.listdir(database_dir)
channel_database['dir'] = database_dir

# Generate seismic data.
Parallel(n_jobs=n_cores)(delayed(parallel)(i,
                                           xmin=xmin, xmax=xmax, dx=dx,
                                           ymin=ymin, ymax=ymax, dy=dy,
                                           zmin=zmin, zmax=zmax, dz=dz, zcut=zcut,
                                           channel_params=channel_params,
                                           channel_database=channel_database,
                                           instance_label=instance_label,
                                           replacement=replacement,
                                           mute=mute, seed_val=seed_val,
                                           output_dir=output_dir)
                                           for i in range(start_num, start_num+n_model))