Workflow — DeepSeis

01

Load & Prepare

Point to your SAC directory or load SEG-Y/MAT files. DeepSeis automatically scans stations, builds a manifest, and creates optimized cache slices for fast processing.

Supported Input Formats

SEG-Y .sgy / .segy files with IBM/IEEE float support

MATLAB .mat files read via h5py

SAC Directory Nested date/station structures with auto-discovery

SAC Directory Structure

root/
├── 2024-08-01/
│   ├── 1001.GP2.2024-08-01.sac
│   ├── 1002.GP2.2024-08-01.sac
│   └── ...
├── 2024-08-02/
│   └── ...
└── .deepseis/
    ├── manifest.json
    ├── slices/
    │   └── panel_*.npy
    └── stations/

SAC directory auto-discovery with flat or nested structures
SEG-Y and MATLAB file direct loading
Manifest caching for instant reload
Automatic station metadata extraction from SAC headers
Configurable NT, Panels, Receivers, and dt parameters

02

Quality Control

Review data slices visually and apply AI-assisted classification. Filter out noisy panels using directional index (DI) analysis and manual labels.

QC Methods

Visual Review Inspect slices with colormap switching and amplitude clipping

AI Classification Automated labeling with configurable C1/C2/C3 thresholds

DI Filter Directional Index histogram and scatter diagnostics

Slice Filter Label-based filtering for cross-correlation input

Visual slice review with 7 colormap options
AI batch labeling with configurable confidence thresholds
DI histogram and FK scatter diagnostics
Manual label override and refinement
Slice filter with label range configuration

03

Retrieve Virtual Shots

Run frequency-domain cross-correlation across all station pairs. Monitor progress in the background queue while reviewing intermediate results.

Cross-Correlation Parameters

NT16384 (default)

Panels1310 (default)

Receivers51 (default)

dt0.002 s (default)

Multi-threaded background processing
Real-time virtual shot preview in Result tab
Virtual shot navigation with Prev/Next and dropdown
Configurable stacking windows and parameters
Positive/negative lag merging for symmetric results
Output shape: (4001, N_virtual × N_receivers)

04

Stack & Export

Apply NMO correction with your receiver geometry, preview stacked profiles, and export results as SEG-Y for further interpretation or as images for publication.

NMO Stack Inputs

Shot Directory Virtual shot or cross-correlation output directory

Coords CSV Receiver geometry with shot/receiver coordinates

Velocity File Optional velocity model for NMO correction

Statics File Optional statics correction file

CSV-based coordinate input for flexible deployment
Velocity scanning and constant velocity modes
Stretch mute (SU smute / relative stretch)
FK velocity cut, taper, and bandpass filtering
CMP bin size control for output spatial sampling
Export as SEG-Y, PNG, or PDF

From Raw Data to Image in Four Steps