Ticket Contents

Description

Goals

• Differentiate between native and planted forests within Area of Interest (AoI) and Micro Watershed (MWS) boundaries.
• Use datasets from planted forest records and other sources (e.g., Forest Survey of India, remote sensing data) to train classification models.
• Compute forest type rasters on Google Earth Engine (GEE) at 30m resolution.
• Vectorize raster outputs to generate field-level or MWS-level polygons representing forest types.
• Publish raster and vector outputs as Earth Engine assets with proper metadata.
• Enable visualization and temporal monitoring of native vs planted forests.

Expected Outcome

• Raster datasets (30m resolution) showing forest type classification: native vs planted.
• Vectorized polygons for AoI/MWS with attributes:
  • Forest type (native/planted)
  • Area (ha)
  • Confidence score or classification probability (if available)
• Published Earth Engine assets (raster + vector) with metadata (source datasets, processing date, classification method).
• Visualization in GEE with color-coded forest types (e.g., native = dark green, planted = light green).
• Validation report confirming coverage, classification accuracy, and alignment with known planted forest datasets.

Acceptance Criteria

Data Acquisition

• Input datasets (planted forest records, FSI, satellite imagery) must be preprocessed and clipped to AoI/MWS boundaries.
• Resolution standardized to 30m.

Raster Classification

• Raster outputs must classify each pixel as either native or planted forest.
• Entire AoI/MWS must be covered with no gaps.
• Optional: include classification confidence or probability per pixel.

Vectorization

• Raster outputs converted to vector polygons using reduceToVectors() in GEE.
• Each polygon must include:
  • Forest type
  • Area (ha)
  • Confidence score (if available)
• Polygons must align with field/MWS boundaries.

Asset Publishing

• Raster and vector datasets must be published as Earth Engine assets.
• Metadata must include data sources, classification schema, resolution, and processing date.

Quality & Validation

• Coverage check: 100% of AoI/MWS classified.
• Accuracy check: ≥85% agreement with reference datasets (FSI, planted forest records).
• Resolution check: raster outputs at 30m.
• Attribute check: all polygons contain forest type, area, and confidence (if available).
• Visualization in GEE shows correct color-coded classification.

Implementation Details

Data Sources

• Planted forest datasets (official records, plantations).
• Native forest datasets (FSI or other surveys).
• Satellite imagery: Sentinel-2, Landsat-8/9.

Processing

• Train classification model (e.g., Random Forest or Decision Tree) using labeled pixels from datasets.
• Apply model in GEE to classify forest type across AoI/MWS.
• Compute raster outputs at 30m resolution.

Vectorization & Publishing

• Vectorize raster outputs into polygons.
• Include attributes: forest type, area, confidence score.
• Upload raster and vector layers as Earth Engine assets with metadata.

Visualization

• GEE visualization with native forests (dark green), planted forests (light green).
• Overlay with AoI/MWS boundaries for field-level inspection.

Validation

• Compare outputs with ground truth or reference datasets.
• Spot-check polygons for correct classification.
• Document coverage, accuracy, and attribute completeness in validation report.

Mockups/Wireframes

No response

Product Name

KYL

Organisation Name

C4GT

Domain

No response

Tech Skills Needed

Python

Organizational Mentor

@amanodt @ankit-work7 @kapildadheech

Angel Mentor

No response

Complexity

High

Category

Backend