CT Traceability - Calibration and Accuracy Calibration and Accuracy Prof. Wim Dewulf, Group T - KU Leuven
Outline • Introduction: terminology and procedures • Voxel size calibration • Edge offset calibration • Conclusions
Outline • Introduction: terminology and procedures • Voxel size calibration • Edge offset calibration • Conclusions
CT metrology: no resolution, but accuracy - Medical and material applications are satisfied with resolution - Dimensional metrology requires accuracy and traceability to the unit of length (the meter) Required accuracy: - Medical and material applications: none or 10 -1 -10 -2 (1..5%) - Dimensional metrology: CMM 10 -5 -10 -6 (laserinterferometry 10 -7 -10 -8 ) � µ � � µ � � µ � � µ � ���������������������������������������������� ����������������������������������������� �������������������� µ � �������������������� µ � ���!������� µ ������ ���������"#!$�
Resolution
Traceability
CT metrology procedure Data Processing 3D Reconstruction CT scan (CT Pro) 3D grey voxel model Object 2D X-ray images Data Analysis Thresholding Dimensional measurement Edge detection (VolumeGraphics) (VolumeGraphics) Dimensions Surface model
Influencing parameters
Two main questions Voxel size ? ? Edge
Outline • Introduction: terminology and procedures • Voxel size calibration • Edge offset calibration • Conclusions
Voxel size calibration Influence of workpiece position 2 µm 1 µm
Voxel size calibration Calibration with traceable reference objects
Voxel size calibration Calibration with traceable reference objects Excellent repeatability
Voxel size calibration Repeatability Calibration with traceable reference objects 4,002 Repeatability of distance between spheres en spheres [mm] Distance between s 4,001 0 2 4 6 8
Voxel size calibration Repeatability Calibration with traceable reference objects Sphere center distance [mm] 2.999 2.998 2.997 2.996 2.995 0 1 2 3 4 5 Measurement n°
Voxel size calibration Repeatability Calibration with traceable reference objects
Voxel size calibration Calibration by combining CT and CMM
Voxel size calibration Calibration by combining CT and CMM
Voxel size calibration Calibration by combining CT and CMM
Voxel size calibration Calibration by combining CT and CMM
Voxel size calibration Repeatability Calibration by combining CT and CMM Repeatability of data processing : 2D-3D reconstruction ( CT-pro ), Spreiding van 30 metingen in software (VG) (30 measurements automatic edge detection and plane-plane measurement ( Volume Graphics ) in software) mean value measurements -1 sigma +1 sigma 4,9640 4,9630 value (mm) 4,9620 arde [mm] 4,9610 Meetwaarde Measured v 4,9600 4,9590 4,9580 ≈ � 3 σ σ ≈ ≈ 1µm Range ≈ ≈ 5.7µm ≈ ≈ ≈ ≈ ≈ ≈ ≈ 4,9570 (5 measurements in approx. same Repeatability including data acquisition : conditions: temperature, time of day Range: 15 µm (at 35µA, 180kV) and machine warm-up, but no temperature controlled room) 30 µm (for 45µA and voltage varying from 160 to 190 kV)
Influences on the voxel size Drift of X-ray source Source: Vogeler et al., 2011
Influences on the voxel size Drift of X-ray source Source: Vogeler et al., 2011
Influences on the voxel size Tilt of detector and/or rotation stage
Influences on the voxel size Tilt of detector and/or rotation stage
Outline • Introduction: terminology and procedures • Voxel size calibration • Edge offset calibration • Conclusions
Edge offset calibration Edge detection Partial volume effect
Edge offset calibration Edge detection Background Material ISO50%
Edge offset calibration Edge detection
Edge offset calibration
Edge offset calibration
Edge offset calibration 32
Edge offset calibration 33
Edge offset calibration Edge detection Edge Background Material ISO50%
Edge offset calibration Deviation compared to CMM measurements for Afwijking t.o.v. meting met CMM different grey values i.f.v. grijswaarde voor de rand 0.060 measurements automatische rand MM - CT) [mm] 0.040 Automatic e detection Afwijking t.o.v. meting met CMM ( CMM 0.020 anddetectie (= ISO50%) ion Deviation compared to CMM m edge [mm] 56D 0.000 47D 27500 28000 28500 29000 29500 30000 30500 31000 31500 32000 39D 48D -0.020 -0.040 -0.060 Grijswaarde voor randdetectie Grey value for edge detection
Edge offset calibration Calibration by combining CT and CMM
Edge offset calibration Calibration by combining CT and CMM Source: Carmignato et al., 2009
Edge offset calibration Calibration by combining CT and CMM 38
Edge offset calibration Advanced edge detection
Edge offset calibration Deviation compared to CMM measurements for Afwijking t.o.v. meting met CMM different grey values i.f.v. grijswaarde voor de rand 0.060 measurements automatische rand MM - CT) [mm] 0.040 Automatic e detection Afwijking t.o.v. meting met CMM ( CMM 0.020 anddetectie (= ISO50%) ion Deviation compared to CMM m edge [mm] 56D 0.000 47D 27500 28000 28500 29000 29500 30000 30500 31000 31500 32000 39D 48D -0.020 -0.040 40 -0.060 Grijswaarde voor randdetectie Grey value for edge detection
Edge offset calibration Calibration workpieces
Edge offset calibration Calibration workpieces
Edge offset calibration Influence of workpiece properties (e.g. material, roughness,…) Optical Tactile
Edge offset calibration Influence of spot size and workpiece position
Edge offset calibration Influence of feature orientation
Edge offset calibration Influence of source settings No BHC
Edge offset calibration Influence of beam hardening Initial spectrum After e.g. 1mm
Edge offset calibration Influence of beam hardening
Edge offset calibration Influence of beam hardening
Edge offset calibration Influence of beam hardening Reconstructed slice
Edge offset calibration Influence of beam hardening
Edge offset calibration Influence of beam hardening Source: Wenig et al., 2006
Edge offset calibration Influence of beam hardening Source: Wenig et al., 2006
Edge offset calibration Influence of beam hardening
Edge offset calibration Influence of beam hardening
Edge offset calibration Influence of beam hardening Inwards offset
Edge offset calibration Influence of beam hardening
Edge offset calibration Influence of beam hardening
Edge offset calibration Number Multi-material objects (Source: Shammaa et al., 2010) Grey value
Edge offset calibration Multi-material objects Steel Steel Ceramic Ceramic Advanced calibration Thresholding method on 1 (local thresholding) greyvalue
Outline • Introduction: terminology and procedures • Voxel size calibration • Edge offset calibration • Conclusions
Conclusions • Traceability of CT is challenging • Influencing factors relate to CT device, measurement procedure, data processing, object, and environment • Good measurement procedures, including calibration of both voxel size and edge offset, can reduce measurement uncertainties by a factor 5-10
Recommend
More recommend
