Loop-Autobedding

Automatic bedding measurement approximation for geological modeling

Overview

Loop-Autobedding is a Python package that addresses a critical challenge in implicit geological modeling: the lack of bedding measurements in many geological datasets. This package provides methods to approximate bedding orientations from various data sources when direct field measurements are not available.

The Problem

Tools like LoopStructural and map2loop require bedding measurements (strike, dip, and dip direction) to perform implicit geological modeling. However, many geological datasets lack these critical measurements. This creates a significant barrier to:

• Broad-scale geological modeling
• Rapid prototyping of geological models
• Simulating large numbers of models for uncertainty analysis
• Working with legacy datasets that don't include structural measurements

The Solution

Loop-Autobedding provides multiple methods to approximate bedding measurements from commonly available data:

1. DEM-based approximation: Uses topographic gradients from Digital Elevation Models
2. Contact-based approximation: Derives orientations from geological map contact lines
3. Regional trend approximation: Applies known regional structural trends
4. Interpolation: Densifies sparse measurements through spatial interpolation

Each method includes quality scoring to help users understand the reliability of approximated measurements.

Installation

From Source

git clone https://github.com/RichardScottOZ/Loop-Autobedding.git
cd Loop-Autobedding
pip install -e .

Requirements

• Python >= 3.8
• numpy >= 1.20.0

Optional Dependencies

For development:

pip install -e ".[dev]"

For running examples:

pip install -e ".[examples]"

Quick Start

import numpy as np
from autobedding import AutobeddingEstimator

# Initialize estimator
estimator = AutobeddingEstimator()

# Example 1: Generate from regional trend
x = np.linspace(0, 1000, 20)
y = np.linspace(0, 1000, 20)
X, Y = np.meshgrid(x, y)

regional_data = estimator.add_regional_trend(
    strike=45.0,
    dip=30.0,
    x_coords=X.flatten(),
    y_coords=Y.flatten(),
    name="regional"
)

print(f"Generated {len(regional_data)} measurements")

# Export to CSV for use with LoopStructural
estimator.export_to_csv(regional_data, "bedding_measurements.csv")

Features

Data Sources Supported

1. Digital Elevation Models (DEMs)

Approximates bedding from topographic slopes, useful in areas with differential erosion:

dem_data = estimator.add_dem_estimates(
    dem=elevation_array,
    x_coords=x_coordinates,
    y_coords=y_coordinates,
    cell_size=30.0,
    sample_spacing=10
)

2. Geological Contact Lines

Derives bedding orientations from mapped geological contacts:

contact_points = [(x1, y1, z1), (x2, y2, z2), ...]
contact_data = estimator.add_contact_estimates(
    contact_points=contact_points,
    formation="Sandstone Formation"
)

3. Regional Structural Trends

Applies uniform regional strike and dip across study area:

regional_data = estimator.add_regional_trend(
    strike=45.0,
    dip=30.0,
    x_coords=x_array,
    y_coords=y_array
)

4. Sparse Measurement Interpolation

Densifies existing sparse measurements:

interpolated = estimator.interpolate_from_sparse(
    known_dataset=sparse_measurements,
    target_x=target_x_coords,
    target_y=target_y_coords,
    method="nearest"
)

Quality Control

All measurements include quality scores (0-1) indicating reliability:

from autobedding.utils import quality_score, validate_measurements

# Get quality metrics
metrics = quality_score(dataset)
print(f"Mean quality: {metrics['mean_quality']:.2f}")
print(f"High quality measurements: {metrics['high_quality_count']}")

# Validate measurements
validation = validate_measurements(dataset)
print(f"Valid: {validation['valid_measurements']}")
print(f"Invalid: {validation['invalid_measurements']}")

Data Combination

Combine multiple data sources with quality filtering:

# Combine all datasets with minimum quality threshold
combined = estimator.combine_datasets(min_quality=0.5)

# Filter by formation
formation_data = combined.filter_by_formation("Formation_1")

# Filter by quality
high_quality = combined.filter_by_quality(min_quality=0.7)

Usage Examples

Basic Usage

See examples/basic_usage.py for a simple introduction:

cd examples
python basic_usage.py

Advanced Workflow

See examples/advanced_workflow.py for a complete workflow including:

• Multiple data source integration
• Quality filtering and validation
• Formation-based filtering
• Export for LoopStructural/map2loop

cd examples
python advanced_workflow.py

API Reference

Core Classes

AutobeddingEstimator

Main class for generating and managing bedding measurements.

Methods:

• add_dem_estimates(): Generate from DEM data
• add_contact_estimates(): Generate from geological contacts
• add_regional_trend(): Generate from regional trend
• interpolate_from_sparse(): Interpolate from sparse data
• combine_datasets(): Combine multiple datasets
• export_to_csv(): Export to CSV file

BeddingMeasurement

Represents a single bedding measurement.

Attributes:

• x, y, z: Coordinates
• dip: Dip angle (0-90°)
• dip_direction: Dip direction (0-360°)
• formation: Formation name (optional)
• quality: Quality score (0-1)
• source: Data source identifier

BeddingDataset

Collection of bedding measurements.

Methods:

• filter_by_quality(): Filter by quality threshold
• filter_by_formation(): Filter by formation name
• to_arrays(): Convert to numpy arrays

Approximators

• DEMApproximator: DEM-based approximation
• ContactApproximator: Contact-based approximation
• RegionalTrendApproximator: Regional trend approximation
• InterpolationApproximator: Interpolation-based approximation

Utilities

• validate_measurements(): Validate dataset
• quality_score(): Calculate quality metrics
• calculate_spatial_coverage(): Calculate spatial statistics
• detect_outliers(): Detect potential outliers

Use Cases

1. Broad-Scale Regional Modeling

Generate bedding measurements across large areas using regional trends and DEM data:

# Regional trend for background
regional = estimator.add_regional_trend(strike=30, dip=25, ...)

# DEM refinement for local variations
dem_refined = estimator.add_dem_estimates(dem, ...)

# Combine with quality filtering
model_input = estimator.combine_datasets(min_quality=0.4)

2. Legacy Dataset Enhancement

Add missing bedding measurements to legacy geological maps:

# Extract contacts from geological map
contacts = extract_contacts_from_map(geological_map)

# Generate bedding approximations
for formation, contact_line in contacts.items():
    estimator.add_contact_estimates(contact_line, formation=formation)

combined = estimator.combine_datasets()

3. Uncertainty Analysis

Generate multiple realizations with varying parameters:

# Create multiple scenarios
for strike in [30, 35, 40, 45]:
    for dip in [20, 25, 30]:
        dataset = estimator.add_regional_trend(
            strike=strike, dip=dip, ...,
            name=f"scenario_s{strike}_d{dip}"
        )
        # Use each scenario for modeling

4. Data Densification

Interpolate between sparse field measurements:

# Load sparse field measurements
known = load_field_measurements()

# Create dense grid
interpolated = estimator.interpolate_from_sparse(
    known, target_x, target_y, method="nearest"
)

Integration with LoopStructural and map2loop

Loop-Autobedding is designed to work seamlessly with Loop3D tools:

import pandas as pd
from LoopStructural import GeologicalModel

# Generate bedding measurements
estimator = AutobeddingEstimator()
# ... add various estimates ...
dataset = estimator.combine_datasets(min_quality=0.5)

# Export to CSV
estimator.export_to_csv(dataset, "bedding.csv")

# Load into LoopStructural
bedding_df = pd.read_csv("bedding.csv")
model = GeologicalModel(...)
model.add_orientation_data(bedding_df)

Quality Considerations

Quality Factors by Method

• DEM-based: 0.4-0.7 (depends on terrain)

  • Best for areas with differential erosion
  • Lower quality in heavily modified terrain

• Contact-based: 0.6-0.8

  • Good for clear geological contacts
  • Assumes contacts parallel bedding

• Regional trend: 0.3-0.6

  • Useful for broad-scale modeling
  • Lower quality due to lack of local variation

• Interpolated: Quality degraded from source

  • Useful for densification
  • Quality decreases with distance from known points

Best Practices

1. Combine multiple sources: Use different methods and filter by quality
2. Validate results: Always check with validate_measurements()
3. Use quality filtering: Set appropriate thresholds for your use case
4. Check spatial coverage: Use calculate_spatial_coverage() to ensure adequate sampling
5. Detect outliers: Use detect_outliers() to identify suspicious measurements

Testing

Run the test suite:

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest tests/

# Run with coverage
pytest --cov=autobedding tests/

Contributing

Contributions are welcome! Areas for improvement:

• Additional approximation methods (e.g., kriging, machine learning)
• Integration with geospatial file formats (GeoTIFF, Shapefiles)
• Visualization tools
• Additional quality metrics
• Performance optimizations

License

Citation

If you use Loop-Autobedding in your research, please cite:

Loop-Autobedding: Automatic bedding measurement approximation for geological modeling
https://github.com/RichardScottOZ/Loop-Autobedding

Related Projects

• LoopStructural - Implicit geological modeling
• map2loop - Map data extraction for geological modeling
• Loop3D - Open source 3D geological modeling platform

Acknowledgments

This package was developed to support the Loop3D project and the broader geological modeling community.

Contact

For questions, issues, or contributions, please use the GitHub issue tracker.