Skip to main content

Scanning Guide using Prescan and Positions Generator

You are here:

Overview

 

This scanning workflow consists of three main steps: Prescan, Positions, and Capture.

Prescan. Create a preliminary 3D reconstruction of the subject to use as a reference for planning the scan.

Positions. Generate camera positions based on the reconstructed model to achieve sufficient coverage and capture details from different angles.

Capture. Move through the generated positions and capture images of the subject for the final 3D reconstruction.

 

The Prescan and the Automatic Positions Generator (APG) features are included under the Arago Control paid license agreement. For inquiries, please contact us at [email protected].

 

1. Prescan

 
Prescan creates a preliminary 3D reconstruction of the subject to help you plan camera positions for the main scan. To open the feature, click Prescan. This feature is available only when Arago is connected and initialised.
 

 
Opening the Prescan panel automatically adds a blue bounding box to the scene.

Connect the RealSense sensor to a USB 3.2 or newer port to ensure sufficient data-transfer bandwidth.

 

Depth View and RGB Camera

 

Enable Depth View to activate the RealSense sensor and display its depth data as a point cloud in the scene.

Enable RGB Camera to display the sensor’s color image in a separate live view. This is optional unless you need to define a custom auto-exposure region.

 

 
ROI Auto-Exp adjusts exposure based on a user-defined region of interest (ROI). Click Set ROI in the RGB Camera live view to define the region. If no ROI is specified, the software attempts to determine one automatically based on the visibility of the bounding box.

Voxel resolution. Controls the level of detail in the Prescan reconstruction. A higher resolution can help capture finer details when scanning smaller objects, but may increase processing time and computational requirements.

Scan mode. Select how many elevations the prescan should do.

  • Quick: One rotation around the bounding box, suitable for obtaining a rough mesh.
  • Regular: Two rotations, one at mid-height and one from above.
  • Full: Three rotations: the two used in Regular mode, plus one from below.

Use Current Position. When enabled, the Prescan uses the RealSense sensor’s current elevation for the scan.

 

Bounding Box

 

When the Prescan panel is opened, an opaque box is also created in the scene. The space enclosed by this box will determine the region processed during mesh reconstruction.
It is essential that your object is fully contained in the box. Re-adjust the size and the location of the box to fully contain the object. To accommodate the turntable spin, it is recommended to make this box larger than the object by 1/3.

 

 

Reconstruction

 

Once ready, click “Start Prescan” and wait for the travel and reconstruction to finish.

  

The remeshing process progress is displayed in realtime in the panel as soon as successfull alignment data becomes available.
In case the alignment process is failing, a red dot will be flashing in the same panel and the process can be canceled and restarted at any time.

Reconstruction Post-processing

 

The initial reconstruction may contain unwanted geometry, such as parts of the turntable stand or other objects in the scene. Use the cleaning and cropping tools to remove unwanted geometry, and Fill Holes to close gaps in the mesh and to make it watertight.

You can save the reconstructed mesh for use in future scans. Alternatively, save the project to preserve the reconstruction and associated scene data.

 

 

reconstruction post processing 2

 

When you have finished editing the mesh, click Finish to complete post-processing and open the Positions panel in Auto mode. This step automatically fills holes and decimates the mesh to reduce its complexity and computational requirements.

Note: Finish cannot be undone. Review the mesh before proceeding.

 

2. Positions – APG

 

The Automatic Positions Generator (APG) generates camera positions using the Prescan reconstruction as a reference or any subject instantiated in the scene. You can generate positions for a selected region of interest or for the entire subject, depending on the desired coverage.

 

Adjust the settings to control the generated positions and repeat the generation process until the result meets your scanning requirements.

 

  

The algorithm uses mesh geometry and ray tracing to determine camera viewpoints. It also uses the primary camera’s sensor size and focal length to calculate the camera’s viewing frustum.

  • Focus Distance. Defines the distance between a camera position and the proxy mesh surface.
  • Spread Angles. Limits the angles at which camera positions can be generated relative to the mesh surface. The range is 0° (zenith) to 180° (nadir).

Area Positions Generator

 

The Area Positions Generator lets you define a region of interest (ROI) on the proxy mesh. APG generates camera positions targeting the selected region.

Click the brush icon to start painting on the mesh. Adjust the brush size using the Brush Size slider or the [ and ] keys.

 

 

To remove paint from a specific area, hold Left  Alt  and click the mesh with the left mouse button  LMB. To erase the entire ROI, click the eraser icon next to the brush.

 

 

 
Occlusion – Scatter controls how a fixed number of camera positions are distributed. Toward Occlusion, the algorithm prioritizes occluded areas of the mesh. Toward Scatter, positions are distributed more uniformly. This setting applies only to Area Positions Generator.
 
Specify the desired number of images for the selected region and click Generate. Click Generate again to add more positions targeting the same region.
 

 
Example of camera positions generated for a painted ROI:
 

 

Full Coverage Generator

 

The Full Coverage Generator creates camera positions across the entire proxy mesh to achieve the desired image overlap. Use this mode when you want to cover the whole subject rather than target a specific region by a specific image overlap count.

 
Images avg. Sets the target average number of overlapping images covering each point on the visible surface when using Generate Optimal Coverage.
 

The default settings are a good starting point for most scanning scenarios. Click Generate Optimal Coverage to generate a new set of positions. If the generated positions are not ideal, click the bin icon to delete all and re-generate.

 

 

 3. Capture

 

Once you have generated and reviewed the camera positions, open the Start panel and click Play to begin the capture sequence.

You can pause and resume the sequence at any time. Once the sequence is complete, you can run it again using the same positions to capture another set of images. This is useful when capturing images under both polarized and non-polarized lighting.

Table of Contents