Registration, Best fit, Alignment, RPS
3DReshaper® provides the ability to place the model in a coherent coordinate system. This feature, also called "registration", "3D alignment" or "RPS alignment (Reference Point System)", can be used on any object: cloud, mesh, polyline, etc. to face many situations as for example:
- When you want to place your model in a specific coordinate system according to geometrical characteristics: symmetry planes, axis, centers, etc. In this case you can make accurate best-fit of geometrical entities like planes, circles, sphere, and make accurate positioning using these elements.
- When you have an existing theoretical geometry, coming for example from an IGES or STEP CAD model and you want to compare or reshape a model using an existing model as reference. In this case you can use some particular points called "Reference Point" (for example hole centers or axis) to define your coordinate system. It is the reason why this method is also called RPS "Reference Point System" or "Reference Positioning System" in the automobile industry. 3DReshaper® provides some tools like:
- Axis alignment: we lock successfully
- 3 translations using one point also called "hard point" or "origin point",
- 2 rotations using a line segment or an axis,
- The last rotation using a plane.
- Point alignment: we enter some point couples (measurement, reference) and 3DReshaper® find the best position so that each measurement point matches the corresponding reference point.
- 3-2-1: This command simulates the layout of a model on an isostatic support. To obtain
the best registration accuracy, point couples should be entered in a strict order called 3-2-1
method that successfully locks the 6 degrees of freedom. Overconstraint systems are also accepted and in this case the best solution is computed in a least squares sense.
- When you have a complex shape you can make a best fit to compute the best position of the raw data (measurement point) in the least squares sense.
- When you get different raw data (measurement point clouds) made in several coordinate systems, for example if the digitalization machine or the model has moved between different acquisitions. In this case, you can make a best fit "all together" so that 3DReshaper® analyses the multiple overlapping areas to find the best position of each individual part according to the others.
In addition, with all registration algorithms it is possible to make ball radius compensations for 3D digitalizers or scanners using mechanical sensors.
|
| Making an accurate laser point clouds alignment. Here using overlapping surfaces of four clouds together with the 3D best-fit registration algorithm. |
| "3-2-1" alignment of a wireframe measurements set: the six arrows represent the movement of each departure point on a surface target and successively lock the six degrees of freedom. |
|
