Establishing Baseline Dimensional Metrology for Injection Molded Polymers
Establishing baseline polymer metrology requires controlling thermal soak kinetics, rigid GD&T fixturing, and accounting for post-mold crystallization drift.
Sub-Voxel Surface Extraction
Meaning
Computational techniques that locate the boundary of a part with higher precision than the base resolution of a volumetric scan improve the accuracy of industrial tomography results. This process works by analyzing the gray-value gradient across multiple adjacent voxels to find the exact point where the material ends and the air begins. Instead of simply picking the center of a voxel, the software uses interpolation to estimate the surface position to within a fraction of a voxel size.
This allows for the measurement of features that are smaller than the nominal resolution of the scanner. In the world of x-ray computed tomography, this capability is what makes the technology competitive with traditional metrology. It enables the inspection of tiny internal channels and complex lattice structures that are otherwise inaccessible.
Accuracy depends on the signal-to-noise ratio of the scan and the quality of the edge-finding algorithm. Proper application of these techniques is essential for validating high-precision aerospace and medical components.
Edge Detection
Identification of the part boundary relies on finding the steepest change in density within the three-dimensional data set. Sub-voxel surface extraction goes beyond simple thresholding by looking at the shape of the gray-value curve. In a perfect scan, the transition from metal to air would be a sharp step.
In reality, it is a blurred slope caused by the physics of the x-ray source and the detector. The software fits a mathematical model to this slope to find the true surface. This method reduces the influence of noise and improves the repeatability of the measurement.
High-quality edge detection is what allows for the precise measurement of wall thickness in complex castings. Operators must be careful to calibrate the system using a phantom of known material and size. This ensures that the gray-values are interpreted correctly across the entire volume.
Resolution Enhancement
Measurement precision is no longer strictly limited by the physical size of the voxels in the reconstructed image. By using sub-voxel surface extraction, it is possible to achieve measurement accuracies that are five to ten times better than the voxel resolution. This means a scan with a one hundred micron voxel size can yield measurements accurate to ten microns.
This enhancement is vital for inspecting large parts where a high-resolution scan would take too long or produce too much data. It allows the quality team to use a coarser scan for the bulk of the part while still getting the precision needed for critical features. However, this does not replace the need for a good scan.
If the raw data is blurry or has too many artifacts, the sub-voxel calculation will be unreliable.
Accuracy Gradient
Reliability of the extracted surface can vary across the part depending on the local geometry and the presence of neighboring materials. Sub-voxel surface extraction is most accurate on flat or gently curved surfaces where the gradient is clear. Near sharp edges or in areas with multiple materials, the algorithm may struggle to find the correct boundary.
This can lead to errors in the measured dimensions. Advanced software packages allow the user to adjust the extraction parameters for different regions of interest. This ensures that each feature is measured using the most appropriate settings.
Validation of the results is often done by comparing the tomography data to measurements taken with a contact probe. This comparison helps build confidence in the digital model. Long-term monitoring of the system performance ensures that the extraction algorithms remain stable as the x-ray tube ages.