dmarsters/microwave-sensing-aesthetic

Microwave Sensing Aesthetic MCP Server

An MCP server that enhances image generation prompts with synthetic aperture radar (SAR) and microwave remote sensing aesthetics. Based on deterministic taxonomy mapping of physical sensing parameters to visual characteristics.

Philosophy: Material Truth

This server implements a "material truth" approach to visual aesthetics - revealing physical properties through electromagnetic interaction rather than optical reflection. We're not seeing light bouncing off surfaces; we're reading how electromagnetic waves interact with material structure, moisture content, roughness, and internal composition.

Architecture

Three-layer olog system:

1. Layer 1: Pure Olog - Deterministic taxonomy (wavelengths, polarizations, subjects, intents)
2. Layer 2: Visual Parameter Mapping - False-color generation, speckle, contrast, geometric distortion
3. Layer 3: Claude Synthesis - Natural language prompt weaving

This achieves ~60-80% cost savings vs pure LLM approaches by separating creative synthesis (Claude) from consistent parameter mapping (deterministic).

Installation

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

# Run server
fastmcp run src/microwave_sensing_aesthetic/server.py

Quick Start

from microwave_sensing_aesthetic.server import enhance_with_microwave_aesthetic

result = enhance_with_microwave_aesthetic(
    base_prompt="Dense Amazon rainforest with winding rivers",
    intent="reveal_subsurface",
    geometric_distortion=4
)

Testing

# Run all tests
bash tests/run_tests.sh

# Or individually
python tests/test_server.py
python tests/test_examples.py

Documentation

See full documentation in the project files:

• tests/README.md - Test documentation
• src/microwave_sensing_aesthetic/ologs/ - Taxonomy specifications

License

MIT

Author

Dal Marsters