The quality of your 3D Mobile Scan depends on careful planning and execution. Follow these best practices to achieve the most accurate and detailed models.
To use Mobile 3D Scan, you must have the following:
Apple LiDAR-enabled device: iPhone Pro/Max or iPad Pro models post-2020.
The latest version of the DroneDeploy Ground App (available in the App Store).
Plan Your Scan Path
Planning your movement ensures complete coverage and a cohesive 3D model.
Walk through the area first and plan your movement to ensure full coverage
Start at a logical entry point and move in an orderly fashion
If possible, try to close the loop of your scan by having a small overlap of your scan to connect the end to where you started
Maintain Consistent Camera Angles and Steady Motion
Consistent angles and smooth movement provide the reconstruction algorithm with the best data.
Keep a consistent scanning angle relative to the surface you are scanning
Move steadily with your feet - keep the phone steady
Capture at least two angles of the same point when scanning flat or featureless surfaces. This improves depth accuracy and ensures better reconstruction
Avoid abrupt movements, painting-like sweeping motions, or rotating the phone
Maintain Effective Distances
Varying your distance and angle, especially in difficult areas, provides the algorithm with more feature points.
For featureless surfaces or narrow spaces, capture from at least two different angles to ensure the reconstruction algorithm has enough data to work with
Back up for a broader view or adjust the angle to add more feature points
Avoid many photos of the same surface taken from the same angle (for example, a large blank wall), as reconstruction struggles with this
Varying angles improve accuracy
Use Straight Paths and Gentle Arcs
Smooth movement minimizes distortion in the resulting 3D model.
Walk in straight, smooth lines
When capturing corners, curve around them naturally
Avoid erratic movements or changes in distances, as this can distort the model
Close Loops When Possible
Closing the loop aids in unifying the different parts of your scan, which is critical for accurate reconstruction.
Try to complete the loop by ending your scan in a position that can see where you started
Ensure slight overlaps in your path to prevent missing details
For fuller coverage, consider making two loops around, capturing in opposite directions
Use the Best Lighting Conditions
Proper lighting is crucial for accurate scans.
Ensure adequate and even lighting throughout the space to prevent incomplete data collection, which could result in missing details or distorted areas in the model
Poor or uneven lighting can lead to incomplete data collection, resulting in missing details or distorted areas in the model
To optimize results, use external lighting or adjust natural lighting for better coverage
Scan with a Clear Area
Minimize obstructions and movement in the environment you are scanning.
Scanning spaces with furniture, equipment, or moving people can obstruct the process, leading to errors or incomplete models
It is advisable to clear the area as much as possible and minimize movement during scanning for better results
Be Aware of Surface Textures
Certain surface types are more challenging for the device's sensor to process.
Surfaces that are monochromatic, reflective, or have low contrast may be difficult for mobile devices to process accurately
Such materials can impact the scan's depth perception, leading to less detailed models
To improve quality, consider scanning from different angles or adjusting lighting to enhance contrast
How to Export and Upload Your Scan
To ensure your data is processed correctly, follow the steps for exporting and uploading.
Export Raw Data Before Syncing: You can only export raw scan data before syncing the scan.
Tap the unsynced pin on your mobile device to access export options.
The export will be zipped; unzip the folder.
In the unzipped folder, only the photos in the folder titled frames need to be uploaded.
Once the scan is synced, the export option is no longer available.
V2.2