Comments (9)
A couple of QC things you can try:
- Are there valid xy points the loaded xarray Dataset?
- Are the corners of your
ds
populated with values? If you loaded from a sparse segy data volume you may need to run the fill_cdp command first.
There is a good example of scraping and checking the headers here. https://segysak.readthedocs.io/en/latest/examples/example_segy_headers.html#Scraping-Headers
from segysak.
the corners are valid in il/xl but the corners_xy are all nan
... 'measurement_system': 'm', 'd3_domain': None, 'epsg': None, 'corner_points': ((1001, 3001), (1001, 4875), (2234, 4875), (2234, 3001)), 'corner_points_xy': ((nan, nan), (nan, nan), (nan, nan), (nan, nan)), 'source_file': 'path.segy', 'srd': None, 'datatype': None, 'percentiles': [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], 'coord_scalar': -100.0}
from segysak.
X and Y corners are for the CDPX and CDPY values. IL and XL corners you can get directly from the dataset dimension vars.
If your X and Y values are not populating correctly it suggests to me that you might not have X and Y data in your SEGY or you have not selected the correct byte positions for the cdpx
and cdpy
variables.
Can you share the output of segy_header_scan(fn)
to display the full output you might have to do this:
import pandas as pd
from IPython.display import display
with pd.option_context("display.max_rows", 89):
display(segy_header_scan(fn))
from segysak.
scrape = scrape[scan[scan["std"] > 0].index]
scrape
TRACE_SEQUENCE_FILE | TraceNumber | CDP | SourceX | SourceY | CDP_X | CDP_Y | INLINE_3D | CROSSLINE_3D | |
---|---|---|---|---|---|---|---|---|---|
1001 | 1 | 3183 | 43518576 | 49289683 | 43518576 | 49289683 | 1001 | 3183 | |
1001 | 2 | 3184 | 43519826 | 49289683 | 43519826 | 49289683 | 1001 | 3184 | |
1001 | 3 | 3185 | 43521076 | 49289683 | 43521076 | 49289683 | 1001 | 3185 | |
1001 | 4 | 3186 | 43522326 | 49289683 | 43522326 | 49289683 | 1001 | 3186 | |
1001 | 5 | 3187 | 43523576 | 49289683 | 43523576 | 49289683 | 1001 | 3187 | |
... | ... | ... | ... | ... | ... | ... | ... | ... | |
1025 | 9996 | 3548 | 43974826 | 49259683 | 43974826 | 49259683 | 1025 | 3548 | |
1025 | 9997 | 3549 | 43976076 | 49259683 | 43976076 | 49259683 | 1025 | 3549 | |
1025 | 9998 | 3550 | 43977326 | 49259683 | 43977326 | 49259683 | 1025 | 3550 | |
1025 | 9999 | 3551 | 43978576 | 49259683 | 43978576 | 49259683 | 1025 | 3551 | |
1025 | 10000 | 3552 | 43979826 | 49259683 | 43979826 | 49259683 | 1025 | 3552 |
from segysak.
A couple of QC things you can try:
- Are there valid xy points the loaded xarray Dataset?
- Are the corners of your
ds
populated with values? If you loaded from a sparse segy data volume you may need to run the fill_cdp command first.There is a good example of scraping and checking the headers here. https://segysak.readthedocs.io/en/latest/examples/example_segy_headers.html#Scraping-Headers
ds.seis.fill_cdpna() did the trick
here are my corners after that:
((432910.76000091387, 492896.8299149062), (432910.75999001024, 477484.3295504351), (456335.7600132545, 492896.83054377494), (456335.76000235084, 477484.3301793039))
THANKS !
from segysak.
You must have had a sparse volume with missing traces.
from segysak.
correct. That is the case for most data by the way. You should consider running this seis_fill_cdpna() by default !
from segysak.
correct. That is the case for most data by the way. You should consider running this seis_fill_cdpna() by default !
Glad you got it sorted but I don't think this is appropriate, the default behaviour is just to load the data available in the segy file. fill_cdpna
makes some assumptions to populate the empty traces with coordinates and that behaviour shouldn't be enforced on all users.
from segysak.
Ok, you may want to warn users when XYs are na in the xarray Dataset when it is converted to a ZGY.
Otherwise users will create ZGYs with nan corners without any warning.
Since fill_cdpna only alters XYs of null or missing traces, I don't see harm in regenerating the XYs since these traces are dead.
Anyway, thanks for your code, this is my goto to convert SEGY files to ZGY files.
Thanks again for your work. OpenVDS is next for me.
That might help others, I added this function to your _openzgy.py file (kinda quick and dirty):
def zgy_writer2(seisnc_dataset, filename, datatype="float", datarange=(-215, 215-1), dimension=None):
"""Write a seisnc dataset to ZGY file format.
Args:
seisnc_dataset (xr.Dataset): This should be a seisnc dataset.
filename (pathlib.Path/str): The filename to write to.
datatype (str): "float", "int16" or "int8"
datarange (float,float): (-1500.0, 1500.0)
dimension (str, optional): The dimension to write out, if
None uses available dimenions twt first.
"""
assert seisnc_dataset.seis.is_3d()
# fill gaps in data
seisnc_dataset.seis.fill_cdpna()
seisnc_dataset.seis.calc_corner_points()
if seisnc_dataset.corner_points_xy:
corners = tuple(seisnc_dataset.corner_points_xy[i] for i in [0, 3, 1, 2])
else:
corners=[(0, 0), (0, 0), (0, 0), (0, 0)]
# print( corners )
dimension = _check_dimension(seisnc_dataset, dimension)
if dimension == "twt":
zunitdim = UnitDimension(2001)
elif dimension == "depth":
zunitdim = UnitDimension(2002)
else:
zunitdim = UnitDimension(2000)
try:
coord_scalar_mult = seisnc_dataset.attrs["coord_scalar_mult"]
except KeyError:
coord_scalar_mult = 1.0
# dimensions
ni, nj, nk = (
seisnc_dataset.dims[CoordKeyField.iline],
seisnc_dataset.dims[CoordKeyField.xline],
seisnc_dataset.dims[dimension],
)
# vertical
z0 = int(seisnc_dataset[dimension].values[0])
dz = int(seisnc_dataset.sample_rate)
# annotation
il0 = seisnc_dataset[CoordKeyField.iline].values[0]
xl0 = seisnc_dataset[CoordKeyField.xline].values[0]
dil = int(np.median(np.diff(seisnc_dataset[CoordKeyField.iline].values)))
dxl = int(np.median(np.diff(seisnc_dataset[CoordKeyField.xline].values)))
dtype_float = SampleDataType[datatype]
if datatype == "float":
with ZgyWriter(
str(filename),
size=(ni, nj, nk),
datatype=dtype_float,
zunitdim=zunitdim,
zstart=z0,
zinc=dz,
annotstart=(il0, xl0),
annotinc=(dil, dxl),
corners=corners,
) as writer:
order = (CoordKeyField.iline, CoordKeyField.xline, dimension)
writer.write(
(0, 0, 0),
seisnc_dataset[VariableKeyField.data]
.transpose(*order)
.values.astype(np.float32),
)
else:
with ZgyWriter(
str(filename),
size=(ni, nj, nk),
datatype=dtype_float,
datarange=datarange,
zunitdim=zunitdim,
zstart=z0,
zinc=dz,
annotstart=(il0, xl0),
annotinc=(dil, dxl),
corners=corners,
) as writer:
order = (CoordKeyField.iline, CoordKeyField.xline, dimension)
writer.write(
(0, 0, 0),
seisnc_dataset[VariableKeyField.data]
.transpose(*order)
.values.astype(np.float32),
)
from segysak.
Related Issues (20)
- dependencies import error - more_itertools HOT 1
- 0.3.1 segy_writer bug HOT 4
- Visualizing VSP data HOT 5
- Implement `pyzgy` for ZGY files as `segyio` drop in replacement. HOT 19
- Documentation: Add comments in FAQ and elsewhere regarding exporting of non-convertable values for header.
- Look into zarr and zfp support, HOT 1
- Improve support for sparse volumes HOT 1
- ValueError when writing 2D seismic data to file after reprojecting coordinates HOT 9
- segy_loader failing to broadcast input array shape during conversion HOT 4
- Documentation fro header re-map trace headers HOT 2
- suppress tqdm warning HOT 1
- zgy_writer HOT 8
- huge volume segy data loading HOT 5
- Geometry issue/ pandas key error HOT 7
- Direct convert: segy to netcdf HOT 17
- convert segy to netcdf that could be used as an input in ArcGIS Pro to create voxel layer HOT 2
- Inconcistency in shapes (2D seismic) causes error when calling segy_loader HOT 2
- [Question] Is it possible to directly stream the data from Azure Blob Store HOT 1
- Segysak 0.5 cannot find engine with xr.open_dataset HOT 4
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from segysak.