Runway Crack Detection Using Drones & ArcGIS Pro

Maintaining airport runway pavement is essential for ensuring flight safety while minimizing operational disruptions. Traditional runway inspections are often labor-intensive, time-consuming, and frequently require temporary runway closures.

By combining drone photogrammetry, high-resolution orthophotos, and ArcGIS Pro Deep Learning, engineers can rapidly identify pavement defects, generate GIS-ready datasets, and prioritize maintenance more efficiently. This modern workflow demonstrates how geospatial technologies are transforming airport infrastructure management.

Workflow Overview

A[Mission Planning] --> B[Drone Survey]
B --> C[Image Acquisition]
C --> D[Pix4D Processing]
D --> E[Orthophoto Generation]
E --> F[ArcGIS Pro]
F --> G[Deep Learning Detection]
G --> H[Validation]
H --> I[Mapping & Reporting]

1. Mission Planning

Every successful runway inspection begins with careful planning.

Key planning activities include:

• Define survey objectives.
• Design an automated flight path.
• Coordinate with airport authorities.
• Obtain flight permissions.
• Select an appropriate drone platform.
• Calibrate sensors and cameras.

For this project, the survey was performed using 85% front overlap, 75% side overlap, and a 70-meter flying height to ensure complete coverage and accurate photogrammetric reconstruction.

2. Drone Data Collection

The drone captured high-resolution nadir (90°) imagery across the runway using a pre-planned automated flight path.

Maintaining consistent overlap between images ensures reliable feature matching during photogrammetric processing, resulting in accurate orthophotos and geospatial products.

3. Photogrammetric Processing

Images --> Alignment
Alignment --> CameraOptimization
CameraOptimization --> DensePointCloud
DensePointCloud --> Mesh
Mesh --> DSM
DSM --> Orthophoto
Orthophoto --> Export

The captured imagery is processed using photogrammetry software such as Pix4D.

Typical outputs include:

• Automatic tie points
• Dense point cloud
• 3D textured mesh
• Digital Surface Model (DSM)
• High-resolution orthomosaic

The final orthophoto is exported as a GeoTIFF for further GIS analysis.

4. Crack Detection Using ArcGIS Pro

The generated orthophoto is imported into ArcGIS Pro and analyzed using Esri's Pavement Crack Detection deep learning model.

Typical Workflow

1. Import the orthophoto. 2. Clean and preprocess the imagery. 3. Load the pretrained Deep Learning model (.dlpk). 4. Run Detect Objects Using Deep Learning. 5. Generate crack feature layers. 6. Validate detected cracks against field observations.

GPU processing is recommended for significantly faster analysis when working with large orthophotos.

5. Mapping & Reporting

Detected pavement cracks are converted into GIS-ready datasets that support:

• Maintenance planning
• Asset management
• Spatial analysis
• Pavement condition mapping
• Engineering reports
• Airport infrastructure monitoring

The resulting spatial database enables engineers to visualize crack locations, classify severity, and prioritize maintenance activities.

Project Results

The automated workflow successfully identified 984 runway pavement cracks.

📊 Project Statistics

• 🛣️ Total Cracks: 984
• 📏 Total Crack Length: 1,797.67 m
• 📐 Average Crack Length: 1.83 m
• 📍 Maximum Crack Length: 63.79 m

Key Findings

• Most detected cracks were short localized fractures.
• Long cracks were concentrated near the center of the runway.
• Edge cracks indicated possible drainage or weather-related deterioration.
• Critical long cracks require immediate maintenance intervention.

Recommendations

Based on the analysis, the following best practices are recommended:

• Repair long cracks immediately.
• Schedule routine maintenance for medium-length cracks.
• Improve runway edge drainage.
• Reinforce high-stress pavement sections.
• Conduct regular drone inspections.
• Integrate predictive maintenance using GIS and Artificial Intelligence.

These measures reduce maintenance costs while improving runway safety and operational reliability.

Technologies Used

• Drone Photogrammetry
• Pix4D
• ArcGIS Pro
• Deep Learning (Pavement Crack Detection)
• Orthophoto Generation
• Digital Surface Model (DSM)
• GIS Mapping
• Machine Learning

Benefits of the Workflow

• Faster runway inspections
• Reduced operational downtime
• High positional accuracy
• Automated crack detection
• GIS-ready deliverables
• Improved maintenance planning
• Scalable inspection workflow

Conclusion

Drone-based runway inspections combined with photogrammetry and deep learning provide a modern, scalable solution for pavement condition assessment. High-resolution orthophotos, AI-powered crack detection, and GIS analysis enable airport authorities to transition from reactive maintenance toward predictive infrastructure management.

As drone technology and geospatial AI continue to advance, these workflows will become increasingly valuable for improving aviation safety, reducing operational disruptions, and extending the lifespan of critical airport infrastructure.