Add a header to begin generating the table of contents

    At Datumate, we are committed to revolutionizing the construction industry through innovative technology. Our recent research aimed to identify the optimal methods for capturing drone images to enhance the accuracy and efficiency of our DatuBIM platform, which integrates AI-driven data analytics and photogrammetry.

    Research Objectives

     The investigation focused on several key questions:

    1. What is the best accuracy users can achieve with DatuBIM?
    2. How do various flight parameters affect mapping accuracy?
    3. How does the number, density, and distribution of Ground Control Points (GCPs) influence mapping accuracy?

    Experimentation Overview

    The study was conducted at the Nahlal junction in Israel, covering a 2.5 km stretch of open roads and 340,000 sqm of asphalt area. Key parameters tested included:

    • Drone and Camera: Phantom 4 Enterprise RTK with a 1″ CMOS sensor and 20 MPX effective pixels.
    • Flight Trajectories: Perpendicular and parallel flights relative to the road, at angles of 90° and 60-65° from the horizon.
    • Flight Elevations: 40m and 60m, achieving GSD of 1.1 cm/px and 1.65 cm/px, respectively.
    • Overlap: 80% forward and 70% side overlaps.

    Ground Control Points (GCPs) and Checker Points

    • GCPs: 40 points measured using RTK GNSS method, spaced 50-75m apart.
    • Checker Points: 60 additional points were used for validation, also measured using RTK GNSS.

    Key Findings

    1. Accuracy: The research confirmed that the DatuBIM platform can deliver a mapping accuracy of up to 1:250 scale with deviations within 5 cm in 95% confidence probability level (CL), as required by Israeli regulations.
    2. Flight Parameters: Both the camera angle and flight elevation significantly impact accuracy. For the highest accuracy, the following configurations are recommended:
    • Flight Elevation: calc the elevation based on the GCPs accuracies. The pixel finger print should be similar to the GCPs accuracies. For examples:
      • If the GCPs 3D accuracies are ~ 1.5-2 cm in 3D, then using the above camera user can fly in 60 m elevation to achieve the 2.5 cm accuracies in 68% (5cm in 95% C.L.). There is no need for flying the drone lower than 60 m in this case.
      • If the GCPs 3D accuracies are ~ 1-1.5 cm in 3D, then using the above camera user can fly in 40 m elevation to achieve the 1.5-2 cm accuracies in 68% (3.5 cm in 95% C.L.).
    • Camera Angle: 60-65° from the horizon provides a balance of detail and efficiency.
    1. Overlap and GCP Distribution: Maintaining an 80% forward and 70% side overlap is crucial. Additionally, using a double-grid flight pattern enhances the texture model quality, especially for 3D mapping.

    Recommendations

    • Flight Planning: Use double-grid flights for best texture models. Single-grid flights with a camera angle of 60-65° are effective for general mapping. If you need only mapping you don’t need to fly double grid, just keep the camera angle to 60-65° from horizon.
    • GCPs: Optimal GCP density for achieving is up to 1 every every 300m under certain conditions.
    • Environmental Conditions: Perform flights between 10:00 to 14:00 on sunny days and limit drone speed to 5 mph. On cloudy days, reduce speed to 3 mph to mitigate the impact of shadows.

    Conclusion

    Our research at Datumate demonstrates that precise drone image capture, combined with intelligent flight planning, significantly enhances the accuracy and efficiency of construction project mapping. By adhering to the recommended flight parameters and GCP distribution, stakeholders can maximize the benefits of the DatuBIM platform, ensuring high-quality, reliable data for all phases of construction projects. 

    For the full report, click here.