End-to-End Pipeline Overview

A seamless 5-phase automated workflow for tax evasion detection

1. Drone Data Capture

RTK-GPS Drones
GCP Survey Data
Flight Planning

2. Photogrammetry

Orthomosaics (2D)
Point Clouds (3D)
DSM & DTM Gen

3. AI Analysis

Mask R-CNN
Building Detection
Floor Calculation

4. Spatial Database

PostgreSQL/PostGIS
Spatial Join
Holdings Database

5. Enforcement

Violation Flagging
Tax Calculation
Web Dashboard

System Capabilities

Accuracy

96.8%

Detection rate

Speed

100 km²

Processed per week

ROI

50-178%

Return on investment

Coverage

Full City

Completed in days

Data Pipeline Flow

Technical workflow from drone capture to enforcement action

Drone
System

Mission Planning

RTK-GPS Image Capture

Metadata Storage

Photogrammetry
Engine

Image Alignment

Dense Point Cloud

DSM & DTM Generation

Orthomosaic

AI Analysis
Model

Mask R-CNN Detection

Roof Segmentation

Obstacle Filtering

Floor Count (nDSM)

Database
Integration

Spatial Join (PostGIS)

Link Approved Plans

Calculate Discrepancy

Enforcement
Action

Color-Coded Dashboard

Auto-Generate Notices

Track Case Status