search

DaloroAT / first_breaks_picking

First break picking in seismic gather
Apache License 2.0
109 stars 39 forks source link

Improve UX #29

Closed DaloroAT closed 1 year ago

DaloroAT commented 1 year ago

Hey @HHTC0

I published beta version with new functionality. Could you please try and check the new version?

pip install first-breaks-picking==0.3.0b0
DaloroAT commented 1 year ago

Changelist:

To install beta pip install first-breaks-picking==0.3.0b0

HHTC0 commented 1 year ago

Hi @DaloroAT, I have tested it now. Apologies I could not respond you earlier due to some additional workload this and next week.

Thanks again for your excellent work!

DaloroAT commented 1 year ago

@HHTC0 Glad to hear it worked out well!

worked however it did not displayed traces at correct position when plotting. It still kept the traces at regular spacing. Let's say I have receiver at 6, 8, and 12 with 10m missing. It should display the missing trace rather regular spacing between 6, 8 and 12m.

I see, you want to put traces on some spatial grid, instead of plotting them sequentially with corresponding labels. In other words, the horizontal axis should be an axis in the mathematical sense, not just a place to put values.

If possible, if you can allow scale value input that allow entering value rather than scroll. E.g., 10 cm/sec or 20 cm/sec or in inches. Also, similar entry for trace density adjustment can be helpful.

Scrolling by the mouse wheel allows you to zoom in and zoom out for X and T axes simultaneously. But you also can zoom/stretch/scale them separately. You need to press and hold the Right button, then move the mouse vertically to scale T or horizontally to scale X, or in both directions. A precise movement of the mouse diagonally will give an effect similar to scrolling with the mouse wheel. Is that what you want?

DaloroAT commented 1 year ago

@HHTC0 could you check mouse interaction?

HHTC0 commented 1 year ago

@DaloroAT Just tested - works fine for panning, scaling and spacing adjustment. Well done!

HHTC0 commented 1 year ago

@DaloroAT if you planning to develop it further. Having a BPF would be helpful for time picking.

DaloroAT commented 1 year ago

Bandpass filter? Yes, I can do it. Unfortunately, the development is taking a long time, since I can only work on the project on weekends.

Could you list common tools for time picking and tomography? It would be convenient to know ahead of time in order to better design the project.

HHTC0 commented 1 year ago

@DaloroAT yes, bandpass is helpful when dealing with noisy data. Low pass is mainly required. No urgency.

Regarding time picking tools, can you please elaborate further?

DaloroAT commented 1 year ago

Regarding time picking tools, can you please elaborate further?

Yes, but please be more specific

HHTC0 commented 1 year ago

@DaloroAT the neural network is doing great job in initial picking. The manual time picking has already added ease of refining were required. Most of the borehole seismic time picking tool does allow picking on peak, trough, or inflection point tangent.

Additionally, having the ability to reject a trace could be another useful functionality.

DaloroAT commented 1 year ago

Most of the borehole seismic time picking tool does allow picking on peak, trough, or inflection point tangent.

Could you please share some examples of it? With videos, images, or at least with paint 😁 I don't know how it looks.

having the ability to reject a trace could be another useful functionality.

You want to reject the exact list of traces, e.g. [1, 4, 5]. Or e.g. list of traces [1, 4, 5] with period T (or Traces per gather) to ignore [1, 4, 5, 1 + T, 4 + T, 5 + T, 1 + 2T, 4 + 2T, 5 + 2T] ? In a particular case, the list may consist of only one trace. I can implement both options.

I also already implemented the inversion of a list of traces, but this option is not in the desktop application right now.