Industrial CT geometric position parameter correction method

Abstract

The invention relates to the technical field of detection, in particular to an industrial CT geometric position parameter correction method, and the method comprises the steps: collecting a filament projection image; reconstructing upper, middle and lower three-layer images of the filament; defining a target function according to the gray scale concentration ratio; and using an optimization algorithm to search for geometric parameters when the target function takes the minimum value. According to the method, the geometric parameter deviation existing in the CT system is remarkably reduced or eliminated. The deviation is a main reason for distortion and blurring of the reconstructed image. Therefore, by applying the correction method, the reconstruction precision of the image is directly improved, the image of the cone-beam industrial CT is clearer and more accurate, the detection requirement of a higher standard is met, and compared with a traditional mode, the correction process is simplified, manual intervention is reduced, and the correction efficiency is greatly improved. Therefore, not only is the operation difficulty reduced, but also the correction period is shortened, so that the CT equipment can recover to the optimal working state more quickly, and the problem that the precision of a traditional adjustment and correction mode is low is solved.

Description

Technical Field

[0001] The present invention relates to the field of detection technology, and in particular to an industrial CT geometric position parameter correction method. Background Art

[0002] Cone-beam industrial CT is widely used, and has the advantages of fast scanning speed, high spatial resolution and high ray utilization.

[0003] The FDK (Feldkamp-Davis-Kress) algorithm is a classic image reconstruction algorithm for cone-beam CT. It requires that the system geometry must be ideal, that is, the ray source, the detector center, and the rotation center are in a straight line, and the straight line is perpendicular to the detector plane; the rotation axis is parallel to the detector plane, and its projection falls on the middle column of the detector. The actual cone-beam CT system is difficult to achieve the ideal installation accuracy during component installation and system assembly, and there is a large deviation, or the system geometry parameters will deviate after the CT system has been used for a period of time. This system geometry parameter deviation will cause the reconstructed image to be distorted or blurred.

[0004] Usually, geometric parameters are repeatedly adjusted manually, which is not only difficult, but also time-consuming and has low precision. Summary of the invention

[0005] The purpose of the present invention is to provide an industrial CT geometric position parameter correction method, aiming to solve the problem of low accuracy of traditional adjustment methods.

[0006] To achieve the above object, the present invention provides an industrial CT geometric position parameter correction method, comprising the following steps:

[0007] collecting filament projection images;

[0008] Reconstruct the upper, middle and lower layers of the filament;

[0009] Define the objective function based on grayscale concentration;

[0010] An optimization algorithm is used to search for the geometric parameters that minimize the objective function.

[0011] Among them, in "collecting filament projection images", the following steps are included:

[0012] Place the filament sample on the sample stage of the device;

[0013] Set acquisition parameters according to the characteristics of the filament and the desired image resolution;

[0014] The projection image of the filament is collected and stored.

[0015] Among them, in “setting acquisition parameters according to the characteristics of the filament and the required image resolution”, the acquisition parameters include exposure time, focal length, and magnification.

[0016] Among them, in "reconstructing the upper, middle and lower three-layer images of the filament", the following steps are included:

[0017] Preprocessing the collected projection images;

[0018] The projected image is divided into three layers: upper, middle, and lower according to the characteristics of the filaments and the quality of the projected image;

[0019] The tomographic reconstruction algorithm is applied to each layer to obtain the reconstructed images of the upper, middle and lower layers.

[0020] Among them, in “preprocessing the collected projection image”, the preprocessing includes denoising and contrast enhancement.

[0021] Among them, in "Defining the objective function according to the grayscale concentration", it includes:

[0022] Extract the gray value from each pixel of the reconstructed image to obtain a gray matrix;

[0023] Sum all pixel values ​​in the grayscale matrix to get the total grayscale value;

[0024] Calculate the coordinates of the grayscale centroid according to the grayscale matrix;

[0025] For each pixel, calculate its distance to the grayscale centroid, and perform weighted summation according to the grayscale value to obtain the weighted radius sum;

[0026] The objective function is defined according to the total gray value, the total gray value, the coordinates and the sum of the weighted radius and the minimum value is obtained.

[0027] Among them, in "using an optimization algorithm to search for geometric parameters that minimize the objective function", the following steps are included:

[0028] Set the initial parameters of the optimization algorithm;

[0029] Use an optimization algorithm to iteratively search in the search space, and update the geometric parameters according to the value of the objective function in each iteration;

[0030] Determine whether the value of the objective function reaches the predetermined convergence condition. If the convergence condition is met, stop searching and output the geometric parameters that make the objective function achieve the minimum value. Otherwise, continue iterating.

[0031] The present invention provides a method for correcting geometric position parameters of industrial CT, comprising the following steps: acquiring projection images of filaments: reconstructing the upper, middle and lower three-layer images of the filaments: defining an objective function according to grayscale concentration; and using an optimization algorithm to search for geometric parameters when the objective function takes a minimum value. The present invention significantly reduces or eliminates the geometric parameter deviation existing in the CT system. This deviation is the main cause of the distortion and blurring of the reconstructed image. Therefore, the application of the correction method will directly improve the reconstruction accuracy of the image, making the image of the cone-beam industrial CT clearer and more accurate, meeting higher standards of detection requirements, simplifying the correction process compared to traditional methods, reducing manual intervention, and thus greatly improving the correction efficiency. This not only reduces the difficulty of operation, but also shortens the correction cycle, allowing the CT equipment to return to the optimal working state more quickly, thereby solving the problem of low accuracy of traditional adjustment methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] Figure 1 The present invention provides a flow chart of an industrial CT geometric position parameter correction method.

[0034] Figure 2 is a flow chart for acquiring filament projection images.

[0035] Figure 3 It is a flowchart for reconstructing the upper, middle and lower layers of the filament.

[0036] Figure 4 It is a flowchart that defines the objective function based on the grayscale concentration.

[0037] Figure 5 It is a flowchart for searching the geometric parameters that minimize the objective function using the optimization algorithm.

[0038] Figure 6 It is the reconstructed image of the linear model with accurate geometric parameters.

[0039] Figure 7 It is the reconstructed image of the linear model when the geometric parameters are inaccurate. DETAILED DESCRIPTION

[0040] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0041] See also Figures 1 to 7 The present invention provides an industrial CT geometric position parameter correction method, comprising the following steps:

[0042] S1 collects filament projection images;

[0043] S11 places the filament sample on the sample stage of the device;

[0044] Specifically, the operator needs to first carefully place the filament sample on the sample stage of the cone-beam industrial CT device. In order to ensure the quality of the acquired projection image, the filament must be firmly clamped or glued to the sample stage to prevent movement or deformation during the acquisition process. At the same time, the position of the filament should be centered to ensure that its projection can evenly cover the effective area of ​​the detector. In addition, ensure that the filament maintains an appropriate angle with the rotation axis of the device to facilitate subsequent image analysis and processing.

[0045] S12 sets acquisition parameters according to the characteristics of the filament and the required image resolution;

[0046] The acquisition parameters include exposure time, focal length, and magnification.

[0047] Specifically, before acquiring the projection image, the operator needs to set the acquisition parameters according to the characteristics of the filament and the required image resolution. These parameters include but are not limited to exposure time, focal length, and magnification. The exposure time setting should take into account the material and diameter of the filament and the required image contrast. The focal length is adjusted to ensure that the projection of the filament can be clearly presented on the detector to avoid image blur or distortion. The choice of magnification depends on the required image resolution, and a higher magnification can provide finer image details.

[0048] S13 collects the projection image of the filament and stores the collected image.

[0049] Specifically, when all acquisition parameters are set, the operator can start the cone beam industrial CT device to acquire filament projection images. During the acquisition process, the device emits X-rays and penetrates the filament sample to form a projection image. These images are then captured by the detector and converted into digital signals for subsequent processing and analysis. In order to ensure the integrity and traceability of the image data, the acquired images should be stored in the designated storage device in a timely manner. During the storage process, attention should also be paid to the naming and classification of the image files to facilitate subsequent search and use. At the same time, in order to ensure data security, the storage device should also be backed up and maintained regularly.

[0050] S2 reconstructs the upper, middle and lower layers of the filament;

[0051] S21 pre-processes the collected projection image;

[0052] The preprocessing includes denoising and contrast enhancement.

[0053] Specifically, the filament projection images collected from the cone-beam industrial CT device are preprocessed to optimize the image quality and lay the foundation for subsequent image segmentation and reconstruction. The preprocessing process mainly includes two steps: denoising and contrast enhancement. Denoising is to remove noise points or stray signals in the image through an algorithm. These noises may come from the electronic noise of the equipment, environmental interference or the inhomogeneity of the sample itself. The selection of the denoising algorithm should be based on the type and distribution characteristics of the noise to ensure that the noise is effectively removed while retaining the image details. Contrast enhancement is to adjust the grayscale distribution of the image to make the contrast between the filament and the background more obvious. This step helps to improve the visualization of the image, making the filament more prominent in the image, which is convenient for subsequent image segmentation and reconstruction. Specific methods for contrast enhancement may include histogram equalization, adaptive contrast enhancement, etc.

[0054] S22 divides the projected image into upper, middle, and lower layers according to the characteristics of the filaments and the quality of the projected image;

[0055] Specifically, after the preprocessing is completed, the projection image is segmented into three layers: upper, middle and lower layers according to the characteristics of the filaments and the quality of the projection image. The boundary position of each layer is determined to ensure that each layer contains sufficient filament information and does not overlap with each other. The segmentation process and the application of image segmentation algorithms, such as threshold segmentation and edge detection, are performed according to the diameter, shape and distribution characteristics of the filaments to achieve accurate image segmentation. At the same time, attention should also be paid to the quality of the projection image. During the segmentation process, the image is carefully processed and analyzed to ensure the accuracy and reliability of the segmentation results.

[0056] S23 applies a tomographic reconstruction algorithm to each layer to obtain reconstructed images of the upper, middle and lower layers.

[0057] Specifically, after completing the image segmentation, a tomographic reconstruction algorithm is applied to each layer to obtain the reconstructed images of the upper, middle and lower layers. When applying the tomographic reconstruction algorithm, a suitable reconstruction algorithm and parameters are selected according to the projection image data of each layer. Common tomographic reconstruction algorithms include filtered back projection algorithm, iterative reconstruction algorithm, etc. At the same time, an efficient calculation method and an optimized data storage solution are selected to ensure the smooth progress of the reconstruction process and the quality of the reconstructed image.

[0058] S3 defines the objective function based on grayscale concentration;

[0059] S31 extracts the gray value from each pixel of the reconstructed image to obtain a gray matrix;

[0060] Specifically, the grayscale value is extracted from each pixel of the reconstructed image. The reconstructed image is a two-dimensional matrix composed of multiple pixels. Each pixel has a corresponding grayscale value, which reflects the material density or absorption coefficient at the pixel position. We traverse each pixel of the reconstructed image, extract its grayscale value, and organize these grayscale values ​​into a new two-dimensional matrix, namely the grayscale matrix, according to the position relationship of the pixels. The construction of the grayscale matrix provides basic data for subsequent calculations.

[0061] S32 sums all pixel values ​​in the grayscale matrix to obtain a total grayscale value;

[0062] Specifically, after obtaining the grayscale matrix, all pixel values ​​therein are summed to calculate the total grayscale value. The total grayscale value is the sum of all pixel values ​​in the grayscale matrix.

[0063] S33 calculates the coordinates of the grayscale centroid according to the grayscale matrix;

[0064] Specifically, the grayscale centroid is the position of the weighted average of all pixel values ​​in the grayscale matrix. To calculate the coordinates of the grayscale centroid, the grayscale value of each pixel must be weighted first, and the weight is the grayscale value of the pixel itself. Then, the weighted pixel position coordinates are summed and divided by the total grayscale value to obtain the coordinates of the grayscale centroid.

[0065] S34 calculates the distance from each pixel to the grayscale centroid, and performs weighted summation according to the grayscale value to obtain a weighted radius sum;

[0066] Specifically, the distance from each pixel in the reconstructed image to the grayscale centroid is calculated, and a weighted sum is performed according to the grayscale value of the pixel to obtain the weighted radius sum. The weighted radius sum reflects the overall distance relationship between the material distribution and the grayscale centroid in the reconstructed image, and is an important parameter in the definition of the objective function. During the calculation process, the Euclidean distance from each pixel to the grayscale centroid needs to be calculated, and then the distance is multiplied by the grayscale value of the pixel, and the sum of all pixels is summed to obtain the weighted radius sum.

[0067] S35 defines the objective function according to the total gray value, the total gray value, the coordinates and the sum of the weighted radius and obtains the minimum value.

[0068] Specifically, key parameters such as the total grayscale value, the coordinates of the grayscale center of gravity, and the sum of weighted radii have been obtained in the previous steps. In this step, the objective function is defined based on these parameters and its minimum value is solved. The choice of the objective function should be determined according to the specific application scenario and requirements, but it should generally be able to reflect characteristics such as the uniformity or concentration of the material distribution in the reconstructed image. In this example, the ratio of the sum of weighted radii to the total grayscale value is used as the objective function. The smaller the ratio, the more concentrated the material distribution and the higher the grayscale concentration. Then, an optimization algorithm is used to solve the minimum value of the objective function to obtain the optimal grayscale concentration.

[0069] The objective function is defined as:

[0070]

[0071] in:

[0072]

[0073] Where g(i, j) is the gray value of pixel (i, j) in the projected image; represents the total grayscale,value; (hc, vc) represents the coordinates of the grayscale centroid,

[0074]

[0075] Represents the sum of weighted radii from the grayscale center. This objective function can effectively determine the grayscale concentration. For thin lines, when the geometric parameters are accurate, each layer is reconstructed as a small dot, and when the geometric parameters are inaccurate, it is reconstructed as an ellipse.

[0076] S4 uses an optimization algorithm to search for geometric parameters that minimize the objective function.

[0077] S41 sets initial parameters of the optimization algorithm;

[0078] The initial parameters include the number of iterations, the step size, and the search range.

[0079] Specifically, a series of initial parameters are set for the optimization algorithm to ensure the smooth progress of the search process. These initial parameters include the number of iterations, step size, and search range. The number of iterations refers to how many times the optimization algorithm will iterate during the search process. The choice of this parameter should be determined according to the complexity of the specific problem and the required accuracy. If the number of iterations is too small, the global optimal solution may not be found; if the number of iterations is too large, the computing time and resource consumption will increase. The step size is the amplitude by which the optimization algorithm updates the geometric parameters in each iteration. The choice of step size has an important impact on the search efficiency and accuracy. A step size that is too large may cause the search process to skip the global optimal solution, while a step size that is too small will make the search process too slow. The search range defines the interval in which the optimization algorithm searches for geometric parameters in the search space. The choice of this range should be determined according to the constraints and prior knowledge of the actual problem. The search range should be large enough to contain the global optimal solution; but at the same time, it should not be too large to avoid unnecessary computation.

[0080] S42 uses an optimization algorithm to iteratively search in the search space, and each iteration updates the geometric parameters according to the value of the objective function;

[0081] Specifically, an optimization algorithm is used to iteratively search in the search space to find the geometric parameters that minimize the objective function. At each iteration, the algorithm calculates the value of the objective function based on the current geometric parameters and updates the geometric parameters based on the value. The method of updating the geometric parameters depends on the optimization algorithm used. For example, for the gradient descent method, the geometric parameters are updated according to the gradient of the objective function with respect to the current geometric parameters; while for the genetic algorithm, it may be necessary to select and crossover new individuals based on the fitness of the individuals. During the iterative search process, it is necessary to continuously monitor the value of the objective function and adjust the search strategy and parameters according to its changes. If the value of the objective function does not change significantly in multiple consecutive iterations, it may mean that the search process has approached the global optimal solution or has fallen into a local optimal solution. At this time, adjust the step size, search range, or adopt other strategies to speed up the search process.

[0082] S43 determines whether the value of the objective function reaches a predetermined convergence condition. If the convergence condition is met, the search is stopped and the geometric parameters that make the objective function achieve the minimum value are output. Otherwise, the iteration continues.

[0083] The convergence conditions include reaching a minimum threshold and the number of iterations reaching an upper limit.

[0084] Specifically, determine whether the value of the objective function has reached a predetermined convergence condition. Convergence conditions usually include reaching a minimum threshold and reaching an upper limit for the number of iterations. Reaching the minimum threshold means that the value of the objective function has been reduced to below a preset threshold. The selection of this threshold should be determined according to the needs of the specific problem. If the value of the objective function is small enough and continued iterations can no longer significantly reduce its value, it can be considered that the search process has converged to the global optimal solution or a region close to the global optimal solution. Reaching an upper limit for the number of iterations means that the optimization algorithm has performed a preset maximum number of iterations. If the objective function value that meets the convergence conditions is not found within this number, it may be necessary to consider adjusting the search strategy, initial parameters, or redefining the problem. When the convergence conditions are met, the search process is stopped and the geometric parameters that minimize the objective function are output.

[0085] Beneficial effects:

[0086] 1. The present invention automatically searches for geometric parameters that minimize the objective function through an optimization algorithm, without the need for repeated manual adjustments, thereby greatly improving the automation of the calibration process and reducing operational difficulty and time consumption.

[0087] 2. The present invention defines grayscale concentration as the objective function and uses an optimization algorithm to search, so as to accurately find the geometric parameters that make the grayscale concentration the highest, thereby achieving accurate correction of the system's geometric parameters and effectively improving the accuracy and quality of the reconstructed image.

[0088] 3. The method of the present invention is not only applicable to cone-beam industrial CT systems, but can also be widely applied to other types of CT systems. It only requires appropriate parameter adjustment and algorithm optimization according to the characteristics of the specific system.

[0089] Fourth, by accurately correcting the geometric parameters, the present invention can improve the overall performance of the cone-beam industrial CT system, including scanning speed, spatial resolution and ray utilization, making it more efficient and reliable in practical applications.

[0090] 5. The calibration method of the present invention is relatively simple and easy to implement. At the same time, due to the use of an automated calibration process, the maintenance cost and time cost of the system are also reduced.

[0091] What is disclosed above is only a preferred embodiment of the industrial CT geometric position parameter correction method of the present invention. Of course, it cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above-mentioned embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for correcting geometric position parameters of industrial CT, characterized by: The following steps are involved: collecting filament projection images; Reconstruct the upper, middle and lower layers of the filament; Define the objective function based on grayscale concentration; An optimization algorithm is used to search for the geometric parameters that minimize the objective function.

2. The method for correcting geometric position parameters of industrial CT according to claim 1, characterized in that: In "Acquiring Filament Projection Images", include the following steps: Place the filament sample on the sample stage of the device; Set acquisition parameters according to the characteristics of the filament and the desired image resolution; The projection image of the filament is collected and stored.

3. The method for correcting geometric position parameters of industrial CT according to claim 2, characterized in that: In "Setting acquisition parameters according to the characteristics of the filament and the required image resolution", the acquisition parameters include exposure time, focal length, and magnification.

4. The method for correcting geometric position parameters of industrial CT according to claim 1, characterized in that: In "Reconstructing the upper, middle and lower three-layer images of the filament", the following steps are included: Preprocessing the collected projection images; The projected image is divided into three layers: upper, middle, and lower according to the characteristics of the filaments and the quality of the projected image; The tomographic reconstruction algorithm is applied to each layer to obtain the reconstructed images of the upper, middle and lower layers.

5. The method for correcting geometric position parameters of industrial CT according to claim 4, characterized in that: In "preprocessing the acquired projection image", the preprocessing includes denoising and contrast enhancement.

6. The method for correcting geometric position parameters of industrial CT according to claim 1, characterized in that: In "Defining the objective function based on grayscale concentration", include: Extract the gray value from each pixel of the reconstructed image to obtain a gray matrix; Sum all pixel values ​​in the grayscale matrix to get the total grayscale value; Calculate the coordinates of the grayscale centroid according to the grayscale matrix; For each pixel, calculate its distance to the grayscale centroid, and perform weighted summation according to the grayscale value to obtain the weighted radius sum; The objective function is defined according to the total gray value, the total gray value, the coordinates and the sum of the weighted radius and the minimum value is obtained.

7. The method for correcting geometric position parameters of industrial CT according to claim 1, characterized in that: In "Using an optimization algorithm to search for geometric parameters that minimize the objective function", the following steps are included: Set the initial parameters of the optimization algorithm; Use an optimization algorithm to iteratively search in the search space, and update the geometric parameters according to the value of the objective function in each iteration; Determine whether the value of the objective function reaches the predetermined convergence condition. If the convergence condition is met, stop searching and output the geometric parameters that make the objective function achieve the minimum value. Otherwise, continue iterating.

Images