[[revorest-ray]] (R-Eye) Scanning Engine

Revorest-Ray is the high-precision scanning engine of the Revorest protocol. It acts as the "eye" that interprets satellite data into structured, scientifically-validated evidence for the Digital MRV (d-MRV) ledger.

🔬 Scientific Foundations

This engine is built upon established forest monitoring and allometric research:

1. Pantropical Biomass Allometry (Chave et al., 2014):
   • Provides the foundation for converting structural tree data (Diameter, Height, Wood Density) into Above-Ground Biomass (AGB). Our model uses satellite proxies to simulate these relationships in tropical Dipterocarp forests.
2. Remote Sensing Allometry (Jucker et al., 2017):
   • Establishes the link between crown architecture (detected via high-res texture and radar) and tree biomass. We use Standard Deviation (SD) of Radar VH as a proxy for the "Architectural Axis" of the canopy.
3. Global Tree Density Baselines (Crowther et al., 2015):
   • Provides the reference for "Global Tree Count" expectations. We use these baselines to estimate tree density (trees per hectare) by combining biomass density with canopy heterogeneity metrics.
4. Tropical Forest Multi-Sensor Fusion:
   • Integrating Sentinel-1 (C-Band Radar) for physical structure and Sentinel-2 (Optical) for photosynthetic health (NDVI) to overcome the "saturation effect" common in high-biomass tropical regions.

⬢ Why We Use Hexagonal Grid

...

Inspired by the Civilization strategy game series, the hexagonal grid (Uber's H3 system) provides several critical advantages for the Revorest protocol:

1. Uniform Adjacency: Unlike squares, every neighbor of a hexagon is at the same distance from its center. This makes risk calculations (e.g., fire spread or "buffer risk" for carbon claims) mathematically consistent in all directions.
2. Organic Representation: Forest edges, rivers, and terrain boundaries are rarely straight 90-degree lines. Hexagons mimic organic shapes more effectively than rectangular grids, reducing "staircase" artifacts at forest borders.
3. Hierarchical Scaling: H3 allows for a nested hierarchy where 1 parent hexagon (Res 10) can contain exactly 49 child hexagons (Res 12). This enables us to maintain a "Single Source of Truth" while scanning at different granularities.
4. Gamified Logic: By discretization of the world into hexagons, we turn the planet into a manageable, "playable" interface for forest restoration, making it intuitive for both scientists and users.

🗺️ Spatial Strategy: Bottom-Up d-MRV Hierarchy

Instead of a top-down averaging approach, Ray uses a high-precision Bottom-Up monitoring logic:

• Scanner Unit (Res 12): ~300 m². The Primary Analysis Unit. All scientific calculations (Biomass, Classification) are performed here to capture sub-pixel dynamics (3 pixels per hex).
• Token Unit (Res 10): ~1.5 hectares. The Ledger Container. A Token is only "VERIFIED" if >90% of its child units are healthy.
• Landscape Unit (Res 8): ~73 hectares. The Regional Overview for macro-level reporting.

📊 Resolution Benchmark: H3 vs. Sentinel-2

H3 Level	Area (Approx)	Sentinel-2 Ratio	Use Case
Res 10	1.5 Ha	150 pixels	Legal/Token Container
Res 12	300 m²	3 pixels	Analysis Sweet Spot
Res 13	44 m²	0.4 pixels	Overkill (Redundant)

📊 Data Extraction & Analysis Logic

Our d-MRV engine operates on a hierarchical validation stack:

1. Scanner Level (Res 12): Precision MRV using Sentinel-1/2 fusion.
2. Container Aggregation (Res 10): Collective Integrity (90% threshold for "Verified" status).
3. Landscape Synthesis (Res 8): 100% mathematical sum of underlying verified tokens.

🔄 Scanning Workflow

graph TD
    A["01_initialize_h3_grid.R"] -->|Generate| B("Res 8, 10, 12 Hierarchy")
    A -->|Verify| A1["01a_validate_h3_grid.R"]
    
    B --> C["02_fetch_satellite_data.R"]
    C -->|Batch Task| D{"Google Earth Engine"}
    D -->|Export CSV| E["/rgee/ Folder"]
    
    E --> F["02a_process_satellite_data.R"]
    F -->|Enrich| G("Hierarchical RDS/GPKG")
    
    G --> H["02b_visualize_satellite_data.R"]
    H -->|HTML Map| I("Satellite Evidence Map")
    
    G --> J["03_process_dmrv_logic.R"]
    J -->|Biomass/MRV| K("Final d-MRV Result")

Loading

📋 Today's Development Plan

1. [Script 01] Hierarchical Initialization: Successfully created Res 8 (Landscape), Res 10 (Token), and Res 12 (Scanner) layers.
2. [Script 02] Multi-Sensor Fetching: Implemented robust Batch Extraction for Sentinel-1 & 2 via Google Earth Engine.
3. [Script 02a/b] Processing & Visualization: Established a pipeline to link real satellite evidence to hexagons and visualize health layers (NDVI).
4. [Script 03] Scientific Scoring: (NEXT) Update the d-MRV logic to incorporate the "Architectural Axis" and detect "Double-Counting" risks.

---

Lead: Ardha | Engine: Revorest Protocol