How to... (step-by-step instructions)
  1. Pix4D Documentation
  2. PIX4Dmatic
  3. How to... (step-by-step instructions)

Point cloud registration - PIX4Dmatic

Point cloud registration is the process of aligning two or more point clouds to obtain a single, complete, and accurate point cloud.

What is point cloud registration?

Registering point clouds involves the process of aligning and combining multiple point clouds into a unified, singular point cloud. These point clouds can be generated through photogrammetry or collected by sensors from various locations and angles, for example, a terrestrial laser scanner.

To align point clouds, common elements are needed in the point clouds to be joined.

How can I register point clouds in PIX4Dmatic?

Before registration, point clouds must be present in the PIX4Dmatic project. These point clouds can originate from:

  • Processing the dataset in PIX4Dmatic.
  • Importing the point clouds. For detailed instructions on importing point clouds into PIX4Dmatic, please refer to this article.

pointcloudreg_1.75

After processing or importing the point clouds, the next step is to identify the common elements necessary for their registration. These elements are distinct features that appear in both point clouds, easy to identify. These features must be in the same position in both points cloud, as the software will utilize them as reference points for merging the clouds. Additionally, at least three of these common features must be available to achieve a successful alignment.

This element will be identified as marks (shortcut M). Use this tool to create different markers in all the point clouds.

Tip: Create one layer per point cloud, and creating the marks on each layer can help identify the marker associated with each point cloud before registration.

The markers on the left are the reference markers, and those on the right are the ones to be aligned. This means that the transformation is calculated to move the markers on the right to the reference markers. The selected point cloud is the one to which the transformation calculated from the markers is applied. Once the markers have been created, it is time to match each marker in each point cloud.

pointcloudreg3_1.75The markers in both point clouds should correspond and occupy identical positions within their respective clouds.

  • Reference point cloud. This point cloud is the reference to merge with the imported point cloud.densepcl_1.75
  • Imported point cloud. This is the imported point cloud that will be merged with the reference point cloud.church_1.75
  • Final result after the registration, both point clouds merged. pointcloudreg2_1.75