Skip to content

add virtual stations and modify covariance prior #51

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
BORONG-SEIS wants to merge 1 commit into
base: master
Choose a base branch
from
Open

add virtual stations and modify covariance prior #51

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
82 changes: 78 additions & 4 deletions gamma/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,6 +68,54 @@ def convert_picks_csv(picks, stations, config):
timestamp0,
)

from pyproj import Proj
def generate_virtual_picks(config, t_start, t_end, pick_interval=14.1, v=6.0):
proj = Proj(f"+proj=sterea +lon_0={config['longitude0']} +lat_0={config['latitude0']} +units=km")
added_lon = np.array(
[-122.82414652, -122.95262168, -122.37010843, -122.12919507,
-122.02831026, -121.7408374 , -121.44178241, -121.43578306,
-121.13691749, -121.14092716, -120.85093316, -120.8482852 ,
-120.61727685, -120.55754466, -122.52978781, -123.08575001,
-121.06983387, -123.73116629, -123.77364856, -123.46086107,
-123.12292307, -122.98747986, -122.74199062, -120.5 ,
-120.61181805, -120.16376028, -120.05377491, -119.71912299,
-119.72273531, -119.39093763, -119.56159317, -120.09566786,
-120.04413004, -119.98432876, -119.87620299, -119.69317441,
-119.52926792, -119.35415746, -119.25843591, -119.15039916,
-118.91953095, -118.81508146, -118.64005703, -118.65317244,
-118.50068618, -118.40118374, -118.72974964, -118.50167475,
-118.21900452, -119.69835168, -117.75963681, -117.4241694 ,
-120.953675 ]
)
added_lat = np.array(
[39.49946678, 39.8569939 , 39.91253099, 39.60099814, 40.32237081,
40.68558211, 40.41839871, 39.96849462, 39.25028619, 39.70040776,
39.97164021, 39.43154741, 40.15209149, 38.8017273 , 41.39434133,
40.304205 , 38.07980305, 41.54789604, 41.18704828, 40.83537571,
41.29294593, 40.84612492, 40.58107511, 37.09010754, 36.63950457,
36.81930455, 36.45849048, 36.45675922, 36.09631963, 36.00356277,
35.6445694 , 37.19399106, 38.08029438, 38.48539883, 37.71917559,
38.08412928, 37.54235371, 37.89547409, 37.35393146, 37.08243587,
37.26541959, 36.9880404 , 37.2557514 , 36.71529466, 36.98316024,
36.43736513, 35.81529587, 36.08202335, 36.25715685, 38.43878105,
35.07589211, 35.24788274, 37.71994667]
)
added_x = proj(added_lon, added_lat)[0]
added_y = proj(added_lon, added_lat)[1]
virtual_pick_x = []
virtual_pick_y = []
virtual_pick_t = []
pick_interval = 14.1
# virtual_t_num = int((raw_picks['t'].max() - raw_picks['t'].min())//pick_interval)
for x, y in zip(added_x, added_y):
# start_time = np.random.uniform(raw_picks['t'].min(), raw_picks['t'].min() + pick_interval, size=1)
virtual_pick_t.extend(np.arange(t_start, t_end, pick_interval))
virtual_pick_x.extend([x]*len(np.arange(t_start, t_end, pick_interval)))
virtual_pick_y.extend([y]*len(np.arange(t_start, t_end, pick_interval)))
virtual_pick_x = np.array(virtual_pick_x)
virtual_pick_y = np.array(virtual_pick_y)
virtual_pick_t = np.array(virtual_pick_t)
return np.array([virtual_pick_t, virtual_pick_x/v, virtual_pick_y/v]).T

def association(picks, stations, config, event_idx0=0, method="BGMM", **kwargs):
data, locs, phase_type, phase_weight, pick_idx, pick_station_id, timestamp0 = convert_picks_csv(
Expand All @@ -85,10 +133,30 @@ def association(picks, stations, config, event_idx0=0, method="BGMM", **kwargs):

if ("use_dbscan" in config) and config["use_dbscan"]:
# db = DBSCAN(eps=config["dbscan_eps"], min_samples=config["dbscan_min_samples"]).fit(data[:, 0:1])
db = DBSCAN(eps=config["dbscan_eps"], min_samples=config["dbscan_min_samples"]).fit(
np.hstack([data[:, 0:1], locs[:, :2] / np.average(vel["p"])])

eps = 13
min_sample = 9
vel = np.average(vel["p"])
dbscan_data = np.hstack([data[:, 0:1], locs[:, :2] / np.array([vel, vel])])

use_phase_score = True
add_vitual_picks = True

keep_data = np.ones(len(data))
pick_weights = picks['phase_score'].values

if add_vitual_picks:
virtual_data = generate_virtual_picks(config, data[:, 0:1].min(), data[:, 0:1].max(), pick_interval=14, v=vel)
print(np.shape(virtual_data), np.shape(dbscan_data))
dbscan_data = np.vstack([dbscan_data, virtual_data])
keep_data = np.append(keep_data, np.zeros(len(virtual_data)))
pick_weights = np.append(pick_weights, 0.8*np.ones(len(virtual_data)))
db = DBSCAN(eps=eps, min_samples=min_sample).fit(
dbscan_data,
sample_weight=pick_weights if use_phase_score else None
)
labels = db.labels_
labels = labels[keep_data == 1]
unique_labels = set(labels)
unique_labels = unique_labels.difference([-1])
else:
Expand Down Expand Up @@ -227,6 +295,10 @@ def associate(
x_std = np.std(locs_[:, 0])
y_std = np.std(locs_[:, 1])
t_std = np.std(data_[:, 0])
rstd = np.sqrt(x_std**2 + y_std**2)
amp_std = np.std(np.log(10**data_[:, 1]))
# print(f"amp_std: {amp_std}")
# use amplitude
## option 1
# scaler = max(np.sqrt(x_std**2 + y_std**2) / 6.0 , 0.1)
## option 2
Expand All @@ -235,8 +307,10 @@ def associate(
# d, v = 50, 6.0
# scaler = max((np.exp(np.sqrt(x_std**2 + y_std**2)/d) - 1)/(np.exp(1) - 1) * d / v / t_std, 0.2) * 10
## option 4
rstd = np.sqrt(x_std**2 + y_std**2)
scaler = max(10.0, (rstd/6.0)*(rstd/60.0)) # 6.0 km/s, 60 km
scaler1 = max(10.0, (rstd/6.0)*(rstd/60.0)) # 6.0 km/s, 60 km
scaler = scaler1
# scaler2 = min(max(10.0, np.exp(amp_std-2.5) * 10**(amp_std)), 440)
# scaler = min(scaler1, scaler2)
if config["use_amplitude"]:
# covariance_prior_pre = [time_range * 10.0, amp_range * 10.0]
covariance_prior_pre = [scaler, scaler]
Expand Down