CN115760713A - A Method for Measuring Micro-Inner Dimensions of Workpieces with Optical Synergetic Cone Beam CT - Google Patents

Abstract

The invention discloses a method for measuring the tiny inner dimension of a workpiece by optical cooperative cone beam CT, which realizes the calibration of point clouds on the inner and outer contours of CT by utilizing the high precision and the high reliability of optical measurement and taking optical outer contour point clouds as a reference through a genetic algorithm so as to realize the precise measurement of the tiny inner dimension of a measured object. The method for measuring the small inner size of the workpiece by the optical cooperation cone beam CT provided by the invention has good reliability, solves the problem of insufficient precision of measuring the small inner size by the CT, and has great practical value in the field of measuring the small inner size.

Description

A Method for Measuring the Small Inner Dimensions of Workpieces with Optical Synergetic Cone Beam CT

technical field

The invention relates to the field of industrial non-destructive testing related to the application of optical testing and cone-beam CT, and relates to a method for measuring tiny inner dimensions.

Background technique

Cone beam CT (Cone Beam Computed Tomography, CBCT) is to use the cone beam ray source and the area array detector to collect a series of projection images of the measured workpiece at different angles, and reconstruct the continuous serial slice images according to the corresponding reconstruction algorithm. technology. As the most advanced non-destructive testing technology at present, cone beam CT can obtain the projection data of multiple sections of the tested workpiece in one scan, intuitively display the internal structure of the detected object, and quantitatively provide the internal size of the object. Cone beam CT has the characteristics of fast scanning speed, high ray utilization rate, and the same spatial resolution within and between slices. However, due to the complexity of the cone beam CT measurement process, the measurement results are affected by many factors, making it difficult to trace the source of the cone beam CT measurement results. In order to reduce the error of cone-beam CT in measuring tiny dimensions, it is necessary to correct the CT point cloud generated by cone-beam CT.

Optical inspection technology is a non-contact measurement technology, which can transmit the visible features contained on the surface of the measured workpiece to the computer for processing, and calculate the precise three-dimensional coordinates of the visual feature center to form a feature point cloud. Optical inspection technology has the characteristics of high point cloud quality and high measurement accuracy; however, for some parts such as deep holes, pits, grooves, and tiny inner dimensions that cannot be scanned by light, optical inspection technology cannot achieve accurate measurement.

Contents of the invention

Aiming at the problem that cone-beam CT detects small size errors, and optical detection technology cannot effectively detect the inner size, the present invention provides a method for measuring the tiny inner size of workpieces with optical cooperative cone-beam CT, combined with the high precision and high reliability of optical measurement, Improving the accuracy of CT measurement and solving the problem of lack of accuracy that may exist in the measurement of small internal dimensions of workpieces by CT has great practical value in the field of measurement of small internal dimensions.

In order to achieve the above object, the present invention adopts the following technical scheme: a method for measuring the tiny inner dimensions of a workpiece by optical cooperative cone beam CT, the method includes the steps in the following order:

(1) Obtain the CT slice image sequence of the workpiece under test by means of cone-beam CT scanning;

(2) Extract CT inner and outer contour features from the CT slice image sequence;

(3) Position the inner and outer contour features of the CT, and generate a point cloud of the inner and outer contours of the CT;

(4) Obtain the optical contour point cloud of the measured workpiece by means of optical scanning;

(5) Using the genetic algorithm to calibrate the CT inner and outer contour point cloud based on the optical outer contour point cloud;

(6) According to the calibrated CT inner and outer contour point clouds, measure the tiny inner dimensions of the workpiece to be measured.

In described step (2), the specific steps of extracting CT inner and outer contour features include:

1) Perform grayscale processing on the CT slice image sequence to obtain the CT grayscale slice sequence;

2) Perform median filter denoising on the CT grayscale slice sequence to obtain the CT grayscale denoised slice sequence;

3) Using the Zernike moment sub-pixel edge detection method to detect the contour of the CT gray-scale denoising slice sequence, and obtain the CT inner and outer contour features.

In described step (3), the CT inner and outer contour features are positioned, and the specific steps of generating CT inner and outer contour point clouds include:

1) Extract the subvoxel coordinates (x, y, z) of CT inner and outer contour features on the CT grayscale denoising slice sequence;

2) Acquire the actual dimensions a _x , a y , a z corresponding to a single voxel in the x, _y , and _z directions in the CT grayscale denoising slice sequence, and a _x , a _y , a _z are determined separately by the CT scanning device ;

3) Calculate the coordinates (xa _x , ya _y , za _z ) of the CT inner and outer contour point clouds.

In described step (5), adopt genetic algorithm, the concrete step that the point cloud of CT inner and outer contour is calibrated comprises:

1) Based on the optical outer contour point cloud, first register the CT inner and outer contour point clouds, calculate the average relative error n of the registration result, and determine the size correction coefficients b _x , b _y , b _z according to the relative error n The approximate interval L, where L=[1-2|n|, 1+2|n|]

2) Select an appropriate population size S, terminate evolution algebra T, crossover probability P ₁ and mutation probability P ₂ , initial iteration number t=0, the binary code of each individual i in the population S is determined by the size correction coefficient b _x , b _y and b _z are jointly determined, and b _x , b _y , b _z ∈ L;

3) Calculate the fitness M _i of each individual i in the population S, and the fitness M _i is defined as the reciprocal of the absolute error m _i of the individual i during registration:

The fitness weight F _i is defined as:

4) Select S individuals according to the size of the fitness weight F _i , perform crossover and mutation operations on these individuals according to the crossover probability _P1 and mutation probability _P2 , and obtain the next generation population _S1 , the number of iterations t=t+1;

5) Compare the number of iterations t with the number of terminated iterations T, if t<T, return to step 3) for the next iteration; if t=T, terminate the iteration, and the individual with the largest fitness M _i in the last generation population S _t corresponds to The size correction factor of is the optimal size correction factor

6) Calibrate the point cloud of the inner and outer contours of the CT, and the coordinates of the point cloud of the inner and outer contours of the CT after calibration are

In described step (6), the concrete steps of measuring the tiny inner dimension of measured workpiece are:

1) Encapsulate the calibrated CT inner and outer contour point cloud to obtain the CT inner and outer contour mesh model;

2) Select a suitable angle to slice the CT inner and outer contour mesh models to obtain 2D feature slices;

3) Fitting and measuring the features in the 2D feature section view to obtain the tiny inner dimensions of the workpiece to be measured.

The beneficial effects of the present invention are: the method for measuring the tiny inner dimensions of workpieces by optical cooperative cone-beam CT provided by the invention can effectively improve the accuracy of CT in measuring tiny inner dimensions, and has great practical value in the field of measurement of tiny inner dimensions , and provide guidance for the production and processing of tiny structural parts.

Description of drawings

Fig. 1 is method flowchart of the present invention;

Fig. 2 is the genetic algorithm flowchart that described step (5) involves;

Fig. 3 is a partial structural diagram of an aero-engine fuel nozzle, including a conical surface (1) and a center hole (2);

Fig. 4 is the point cloud of the aeroengine fuel nozzle CT cone surface and center hole involved in step (3) of the specific embodiment;

Fig. 5 is the point cloud of the optical cone surface of the aeroengine fuel nozzle involved in step (4) of the specific embodiment;

Fig. 6 is the CT cone surface and center hole point cloud after calibration of the aeroengine fuel nozzle involved in step (5) of the specific embodiment;

Fig. 7 is the aeroengine fuel nozzle CT cone surface and central hole grid model involved in step (6) of the specific embodiment;

Fig. 8 is a partial 2D characteristic sectional view of the aeroengine fuel nozzle CT cone surface and center hole grid model involved in step (6) of the specific embodiment.

Detailed ways

As shown in Figure 3, taking the central aperture detection of an aero-engine fuel nozzle as an example, the method of the present invention is used to realize the measurement of the central aperture, and the method comprises the steps of the following order:

(1) Use cone beam CT equipment to scan the fuel nozzle, and obtain the CT slice image sequence of the fuel nozzle;

(2) Perform grayscale processing and median filter denoising processing on all CT slice image sequences to obtain CT grayscale denoising slice sequences, and then use the Zernike moment sub-pixel edge detection method to contour the CT grayscale denoising slice sequences Detection, to obtain CT cone surface, center hole features;

(3) Extract the subvoxel coordinates (x, y, z) of the CT cone surface and central hole features on the CT gray-scale denoising slice sequence, and obtain the individual voxel coordinates (x, y, z) in the CT gray-scale denoising slice sequence. Calculate the coordinates (xa _x , ya _y , za _z ) of the CT cone surface and the point cloud of the center hole corresponding to the actual dimensions a _x , a _y , and a _z in the z direction;

(4) Use the Alicona automatic zoom three-dimensional surface measuring instrument to scan the nozzle cone to obtain the optical cone point cloud of the fuel nozzle;

最后对CT锥面、中心孔点云进行标定，标定后的CT锥面、中心孔点云坐标为

(5) Based on the point cloud of the optical cone surface, the initial registration is performed on the point cloud of the CT cone surface and the center hole, and the average relative error n of the registration result is calculated, and the size correction coefficients b _x and b _y are determined according to the relative error n , the approximate interval L where b _z is located, and then select the appropriate population size S, terminate the evolution algebra T, crossover probability P ₁ and mutation probability P ₂ , initial iteration number t=0, and then perform Calculate the fitness M _i and the fitness weight F _i , select S individuals according to the size of the fitness weight F _i , and perform crossover and mutation operations on these individuals according to the crossover probability _P1 and mutation probability _P2 to obtain the next generation population S ₁ , the number of iterations t=t+1, compare the number of iterations t and the number of terminated iterations T, if t<T, proceed to the next iteration; if t=T, terminate the iteration, and obtain the fitness M of the last generation population S _t The optimal size correction coefficient corresponding to the individual with the largest _i

Finally, the CT cone surface and the center hole point cloud are calibrated, and the coordinates of the calibrated CT cone surface and center hole point cloud are

(6) Encapsulate the calibrated CT cone surface and center hole point cloud to obtain the CT cone surface and center hole grid model, and select the angle perpendicular to the hole wall to compare the CT cone surface and center hole grid model at different heights Slicing is performed to obtain a 2D feature sectional view, and the circular features in all 2D feature sectional views are fitted and measured and averaged to obtain the diameter of the center hole of the fuel nozzle.

Claims (5)

1. A method for optically coordinated cone-beam CT to measure the tiny inner dimensions of a workpiece, characterized in that: the method comprises the steps in the following order: (1) Obtain the CT slice image sequence of the workpiece under test by means of cone-beam CT scanning; (2) Extract CT inner and outer contour features from the CT slice image sequence; (3) Position the inner and outer contour features of the CT, and generate a point cloud of the inner and outer contours of the CT; (4) Obtain the optical contour point cloud of the measured workpiece by means of optical scanning; (5) Using the genetic algorithm to calibrate the CT inner and outer contour point cloud based on the optical outer contour point cloud; (6) According to the calibrated CT inner and outer contour point clouds, measure the tiny inner dimensions of the workpiece to be measured. 2. the method for a kind of optical cooperative cone-beam CT measurement workpiece tiny inner size according to claim 1, is characterized in that: in described step (2), the specific step of extracting CT inner and outer contour features comprises: 1) Perform grayscale processing on the CT slice image sequence to obtain the CT grayscale slice sequence; 2) Perform median filter denoising on the CT grayscale slice sequence to obtain the CT grayscale denoised slice sequence; 3) Using the Zernike moment sub-pixel edge detection method to detect the contour of the CT gray-scale denoising slice sequence, and obtain the CT inner and outer contour features. 3. The method for measuring the tiny inner dimensions of workpieces by optical cooperative cone-beam CT according to claim 1, characterized in that: in the step (3), the inner and outer contour features of the CT are positioned to generate the inner and outer contours of the CT. The specific steps of the outline point cloud include: 1) Extract the subvoxel coordinates (x, y, z) of CT inner and outer contour features on the CT grayscale denoising slice sequence; 2) Acquire the actual dimensions a _x , a y , a z corresponding to a single voxel in the x, _y , and _z directions in the CT grayscale denoising slice sequence, and a _x , a _y , a _z are determined separately by the CT scanning device ; 3) Calculate the coordinates (xa _x , ya _y , za _z ) of the CT inner and outer contour point cloud. 4. A method for measuring the tiny inner dimensions of workpieces by optical cooperative cone-beam CT according to claim 1, characterized in that: in the step (5), a genetic algorithm is used to calibrate the CT inner and outer contour point clouds The specific steps include: 1) Based on the optical outer contour point cloud, first register the CT inner and outer contour point clouds, calculate the average relative error n of the registration result, and determine the size correction coefficients b _x , b _y , b _z according to the relative error n The approximate interval L, where L=[1-2|n|,1+2|n|] 2) Select an appropriate population size S, terminate evolution algebra T, crossover probability P ₁ and mutation probability P ₂ , initial iteration number t=0, the binary code of each individual i in the population S is determined by the size correction coefficient b _x , b _y and b _z are jointly determined, and b _x , b _y , b _z ∈ L; 3) Calculate the fitness M _i of each individual i in the population S, and the fitness M _i is defined as the reciprocal of the absolute error m _i of the individual i during registration:

The fitness weight F _i is defined as

4) Select S individuals according to the size of the fitness weight F _i , perform crossover and mutation operations on these individuals according to the crossover probability _P1 and mutation probability _P2 , and obtain the next generation population _S1 , the number of iterations t=t+1; 5) Compare the number of iterations t with the number of terminated iterations T, if t<T, then return to step 3) for the next iteration; if t=T, then terminate the iteration, and the individual with the largest fitness M _i in the last generation population S _t corresponds to The size correction factor of is the optimal size correction factor

6) Calibrate the point cloud of the inner and outer contours of the CT, and the coordinates of the point cloud of the inner and outer contours of the CT after calibration are

5. a kind of optical cooperative cone beam CT according to claim 1 measures the method for the tiny inner dimension of workpiece, it is characterized in that: in described step (6), the concrete steps of measuring the tiny inner dimension of measured workpiece are: 1) Encapsulate the calibrated CT inner and outer contour point cloud to obtain the CT inner and outer contour mesh model; 2) Select a suitable angle to slice the CT inner and outer contour mesh models to obtain 2D feature slices; 3) Fitting and measuring the features in the 2D feature section view to obtain the tiny inner dimensions of the workpiece to be measured.

Images