Self-adaptive weighted filtering reconstruction method and device for double-circle trajectory imaging of radiation source and detector

Abstract

The invention discloses a self-adaptive weighted filtering reconstruction method and device for double-circle trajectory imaging of a radiation source and a detector. The method comprises the following steps: acquiring original projection data; constructing a composite weight matrix by obtaining the dynamic offset of the virtual projection point; generating weighted projection data according to the original projection data and the composite weight matrix; performing periphery filling and image rotation operation on the weighted projection data to obtain an intermediate processing image; performing full-angle self-adaptive frequency spectrum continuous compensation filtering on the intermediate processing image to obtain a compensated image; performing image rotation and intelligent center cutting operation on the compensated image to obtain a final projection; and based on an FDK back projection formula, performing weighted back projection reconstruction on the final projection to obtain a three-dimensional cross-sectional image. The method can realize high-efficiency, high-quality and isotropic-resolution image reconstruction, and can be widely applied to the technical field of three-dimensional cross-sectional image reconstruction.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional tomographic image reconstruction technology, and in particular to an adaptive weighted filtering reconstruction method and apparatus for dual-circular trajectory imaging of a X-ray source and a detector. Background Technology

[0002] Existing computed tomography (CL) techniques employ a tilted scanning geometry (where the X-ray source and detector form a certain angle) to allow X-rays to avoid the long axis of the object, thus achieving layer-by-layer imaging of plate-like objects. Since the detector only translates and does not rotate, the vertical projection point of the X-ray source onto the detector plane is dynamically changing (often outside the imaging area). However, existing techniques assume the projection point is in the center, leading to incorrect weight distribution and gradient artifacts such as "bright on one side and dark on the other," resulting in grayscale distortion. Simultaneously, the large tilt angle of tomography (e.g., 35 degrees) results in extremely long optical path lengths at the "far end" of the detector (the side furthest from the X-ray source), causing severe signal attenuation. Current weighting methods do not specifically address this area, resulting in a large amount of noise at the far end of the reconstructed image and an uneven background. Furthermore, FDK reconstruction theory requires filtering to be performed along the tangent direction of the scan trajectory. Since the detector does not rotate, the angle between its pixel row direction and the scan tangent direction changes continuously during the scan. Most existing techniques use fixed row-wise filtering, violating the central slice theorem and causing anisotropic smearing blur in the image.

[0003] Some techniques employ image rotation filtering to correct the filtering direction. However, they neglect the interpolation smoothing effect during discrete image rotation. For example, bilinear / bicubic interpolation causes the greatest high-frequency loss at 45° rotation and the least at 0°. Existing techniques use fixed filter parameters, leading to inconsistent resolution (anisotropy) in the reconstructed image across different directions, with minute details becoming blurred at specific angles. Some existing techniques use zero-padding, causing a sharp abrupt change in edge data from the background, resulting in strong edge artifacts after high-pass filtering. Furthermore, using a fixed large padding size creates significant computational redundancy (FFT calculation of invalid data) at scan angles of 0° or 90°, significantly reducing reconstruction efficiency. Summary of the Invention

[0004] In view of this, the main objective of the embodiments of the present invention is to provide an adaptive weighted filtering reconstruction method and apparatus for dual circular trajectory imaging of X-ray source and detector, in order to solve at least one of the problems of the prior art. The present invention can achieve high-efficiency, high-quality, and isotropic resolution image reconstruction.

[0005] To achieve the above objectives, one aspect of the present invention provides an adaptive weighted filtering reconstruction method for dual-circular trajectory imaging of a radiation source and a detector, the method comprising: Obtain the raw projection data; A composite weight matrix is ​​constructed by obtaining the dynamic offset of the virtual projection points; Based on the original projection data and the composite weight matrix, weighted projection data is generated; The weighted projection data is subjected to perimeter filling and image rotation operations to obtain an intermediate processed image; The intermediate processed image is subjected to full-angle adaptive spectrum continuous compensation filtering to obtain the compensated image; The compensated image is subjected to image rotation and intelligent center cropping operations to obtain the final projection; Based on the FDK back projection formula, the final projection is reconstructed by weighted back projection to obtain a three-dimensional tomographic image.

[0006] In some embodiments, constructing a composite weight matrix by obtaining the dynamic offset of the virtual projection points includes the following steps: Obtain the vertical distance from the X-ray source to the plane of the detector, the radius of rotation of the center of the detector around the Z-axis, the pixel size of the detector, and the projection angle; Obtain the vertical projection point of the ray source on the plane of the detector to obtain the virtual projection point; The dynamic offset of the virtual projection point is obtained based on the rotation radius, the pixel size, and the projection angle. The true physical distance of a pixel is obtained based on the dynamic offset, the vertical distance, and each pixel on the detector. The relative depth factor is obtained based on the first offset of the dynamic offset in the v-axis direction, the first coordinate value of the pixel in the v-axis direction, the pixel size, and the maximum real physical distance of the pixel. Based on the relative depth factor, suppression intensity coefficient, and nonlinear factor, a nonlinear noise suppression weight is constructed. Based on the vertical distance and the actual physical distance of the pixel, the geometric optical path correction weight is obtained, and the geometric optical path correction weight is fused with the nonlinear noise suppression weight to obtain the composite weight matrix.

[0007] In some embodiments, generating weighted projection data based on the original projection data and the composite weight matrix includes the following steps: The original projection data is fused with the composite weight matrix to obtain the weighted projection data with deep denoising.

[0008] In some embodiments, performing perimeter padding and image rotation operations on the weighted projection data to obtain an intermediate processed image includes the following steps: Obtain the angle between the tangent direction of the scan trajectory and the u-axis direction to get the instantaneous tangent angle; Obtain the first number of pixels of the detector in the u-axis direction and the second number of pixels of the detector in the v-axis direction; A dynamic fill width model is constructed based on the instantaneous tangent angle, the number of the first pixel, the number of the second pixel, and the safety margin; Based on the dynamic fill width model, obtain the minimum fill width; Based on the minimum fill width, the weighted projection data is filled around its perimeter to obtain the filled image; Based on bilinear or bicubic interpolation, the filled image is rotated by an angle equal to the instantaneous tangent angle using a relaxed mode to obtain the intermediate processed image.

[0009] In some embodiments, performing full-angle adaptive spectral continuous compensation filtering on the intermediate processed image to obtain the compensated image includes the following steps: The intermediate processed image is subjected to a Fast Fourier Transform to obtain frequency domain data; Obtain the angle between the tangent direction of the scan trajectory and the u-axis direction to get the instantaneous tangent angle; Based on the instantaneous tangent angle and sensitivity coefficient, a full-angle interpolation loss function is constructed; The angle-specific cutoff frequency is obtained based on the reference cutoff frequency, the maximum compensation bandwidth, and the full-angle interpolation loss function. Based on the Haining window and the angle-specific cutoff frequency, a dynamic filter is constructed; The frequency domain data is filtered using the dynamic filter to obtain the compensated image.

[0010] In some embodiments, performing image rotation and intelligent center cropping operations on the compensated image to obtain the final projection includes the following steps: Obtain the angle between the tangent direction of the scan trajectory and the u-axis direction to get the instantaneous tangent angle; The compensated image is rotated by an angle equal to the instantaneous tangent angle to obtain the restored image; A smart center cropping operation is performed on the restored image to extract the geometric center region of the restored image, thus obtaining the final projection.

[0011] In some embodiments, a weighted backprojection reconstruction is performed on the final projection based on the FDK backprojection formula to obtain a three-dimensional tomographic image. The formula used includes: ; In the formula, Represents a three-dimensional tomographic image; Represents the coordinates of the voxel on the z-axis; Represents the vertical distance from the radiation source to the detector plane; Represents the projection angle The final projection below; Represents the horizontal projection coordinates of the voxel on the detector; Represents the vertical projection coordinates of the voxel on the detector.

[0012] To achieve the above objectives, another aspect of the present invention proposes an adaptive weighted filtering reconstruction device for dual-circular trajectory imaging of a radiation source and a detector, the device comprising: The first module is used to acquire the raw projection data; The second module is used to construct a composite weight matrix by obtaining the dynamic offset of the virtual projection points; The third module is used to generate weighted projection data based on the original projection data and the composite weight matrix; The fourth module is used to perform perimeter filling and image rotation operations on the weighted projection data to obtain an intermediate processed image; The fifth module is used to perform full-angle adaptive spectrum continuous compensation filtering on the intermediate processed image to obtain the compensated image; The sixth module is used to perform image rotation and intelligent center cropping operations on the compensated image to obtain the final projection; The seventh module is used to perform weighted back projection reconstruction on the final projection based on the FDK back projection formula to obtain a three-dimensional tomographic image.

[0013] To achieve the above objectives, another aspect of the present invention provides an electronic device, the electronic device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method described above.

[0014] To achieve the above objectives, another aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0015] To achieve the above objectives, another aspect of the present invention provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions to cause the computer device to perform the aforementioned method.

[0016] The embodiments of the present invention include at least the following beneficial effects: The present invention provides an adaptive weighted filtering reconstruction method and apparatus for dual-circular trajectory imaging of a X-ray source and a detector. This scheme provides a data foundation for subsequent processing by acquiring the original projection data; by acquiring the dynamic offset of the virtual projection points, a composite weight matrix is ​​constructed, and then combined with the original projection data to generate weighted projection data, thereby correcting the geometric optical path error caused by the eccentric translation of the detector; the weighted projection data is subjected to perimeter filling and image rotation operations to obtain an intermediate processed image, eliminating the "frame" artifact caused by edge truncation and aligning the filtering direction; the intermediate processed image is subjected to full-angle adaptive spectral continuous compensation filtering to obtain a compensated image, achieving isotropic resolution; the compensated image is subjected to image rotation and intelligent center cropping operations to obtain the final projection, restore the image coordinates, and automatically remove auxiliary data; based on the FDK back projection formula, the final projection is reconstructed by weighted back projection to obtain a three-dimensional tomographic image, achieving high-efficiency and high-quality image reconstruction. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of the adaptive weighted filtering reconstruction method for dual-circular trajectory imaging of X-ray source and detector provided in an embodiment of the present invention; Figure 2 This is a flowchart of constructing a deeply dependent nonlinear denoising weighted field provided in an embodiment of the present invention; Figure 3 This is a flowchart of dynamic minimum boundary extension based on geometric envelope provided in an embodiment of the present invention; Figure 4 This is a flowchart of the full-angle adaptive spectrum continuous compensation filtering provided in the embodiments of the present invention; Figure 5 This is a flowchart of image rotation and intelligent center cropping provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the adaptive weighted filtering reconstruction process for dual-circular trajectory imaging of the X-ray source and detector provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this invention; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this invention as detailed in the appended claims.

[0020] It should be noted that although functional modules are divided in the system diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the system or the order in the flowchart. The terms "first / S100" and "second / S200" in the specification, claims, and the foregoing drawings may be used herein to describe various concepts, but unless specifically stated otherwise, these concepts are not limited by these terms. These terms are used only to distinguish one concept from another. For example, first information may also be referred to as second information without departing from the scope of the embodiments of the invention, and similarly, second information may also be referred to as first information. Depending on the context, the words "if" or "when" as used herein may be interpreted as "when," "in response to a determination," or "in the event of a determination."

[0021] The terms “at least one,” “multiple,” “each,” “any,” etc., used in this invention, “at least one” includes one, two, or more than two; “multiple” includes two or more than two; “each” refers to each of the corresponding multiple; and “any” refers to any one of the multiple.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to limit the invention.

[0023] Before providing a detailed description of the embodiments of the present invention, some of the nouns and terms involved in the embodiments of the present invention will be explained first. The nouns and terms involved in the embodiments of the present invention are subject to the following interpretations.

[0024] Dual-circular trajectory scanning: a mechanical scanning mode in which the X-ray source remains stationary, the object moves in a circle above the X-ray source, and the detector moves in a synchronous circular translation above the object.

[0025] Translational detector: This refers to a detector that moves along a circular trajectory while its own pixel coordinate system axis remains parallel to the world coordinate system axis, meaning the detector does not rotate.

[0026] FDK reconstruction algorithm: It is a classic cone-beam CT analytical reconstruction algorithm based on the theory of filtered back projection.

[0027] Isotropic resolution: refers to the consistent sharpness of an image in all directions. Existing technologies produce directional blur (anisotropy) in translation detector mode.

[0028] Virtual projection point: Since the detector does not rotate, the vertical projection point of the X-ray source on the detector plane is dynamically changing.

[0029] Edge copy extension: an image filling strategy that fills the extended region with the pixel values ​​of the image boundary.

[0030] In the field of modern industrial nondestructive testing, with the increasing integration of electronic components, the demand for internal defect detection in large, flat objects (such as printed circuit boards (PCBs), ball grid array packages (BGAs), IGBT modules, carbon fiber laminates, etc.) is becoming increasingly urgent. Traditional computed tomography (CT) typically requires X-rays to be scanned perpendicular to the rotation axis. For plate-shaped objects, this results in excessively long X-ray penetration paths, causing severe beam hardening and metal artifacts, and is also limited by the object's geometry, preventing rotation. Computed tomography (CL) technology employs an inclined scanning geometry (the X-ray source and detector form a certain angle), allowing the X-rays to avoid the long axis of the object, thus achieving layer-by-layer imaging of plate-shaped objects. In the mechanical implementation of CL systems, to simplify motion control and ensure system stability, a "dual circular trajectory" scanning mode is often used: the X-ray source remains stationary, the object moves in a circle above the X-ray source, and the detector performs a synchronous circular translational motion above the object. A key characteristic of tilted scanning geometry is that while the detector moves along a circular trajectory, its own pixel coordinate system axis remains parallel to the world coordinate system axis (i.e., the detector does not rotate). Because the detector only translates and does not rotate, the vertical projection point of the X-ray source onto the detector plane is dynamically changing (often outside the imaging area). This non-standard CT geometry causes the geometric model of traditional reconstruction algorithms to fail.

[0031] In view of this, this invention provides an adaptive weighted filtering reconstruction method and apparatus for dual-circular trajectory imaging of a ray source and a detector. This scheme solves the problems of gray-scale distortion, uneven noise distribution, directional blur and edge artifacts by introducing depth-dependent noise suppression weighting, geometry-based dynamic filling and full-angle adaptive spectrum compensation filtering. It achieves high-efficiency, high-quality and isotropic resolution image reconstruction without increasing hardware costs.

[0032] Figure 1This is an optional flowchart of an adaptive weighted filtering reconstruction method for dual-circular trajectory imaging of a radiation source and detector provided in an embodiment of the present invention. Figure 1 The method may include, but is not limited to, steps S100 to S700: Step S100: Obtain the raw projection data; Step S200: Construct a composite weight matrix by obtaining the dynamic offset of the virtual projection points; Step S300: Generate weighted projection data based on the original projection data and the composite weight matrix; Step S400: Perform perimeter filling and image rotation operations on the weighted projection data to obtain an intermediate processed image; Step S500: Perform full-angle adaptive spectrum continuous compensation filtering on the intermediate processed image to obtain the compensated image; Step S600: Perform image rotation and intelligent center cropping operations on the compensated image to obtain the final projection; Step S700: Based on the FDK back projection formula, perform weighted back projection reconstruction on the final projection to obtain a three-dimensional tomographic image.

[0033] In step S100 of some embodiments, the acquisition of raw projection data provides a basic data source for subsequent processing, ensuring that the reconstruction process has reliable input.

[0034] In this embodiment of the invention, the following system coordinate system is defined: (1) World coordinate system The origin is located at the radiation source. , The axis is vertically upward; (2) Local coordinate system of the detector The origin is located at the center of the detector. The axis corresponds to the detector's horizontal direction. The axis corresponds to the direction of the detector array (perpendicular) and is always parallel to the world coordinate system. shaft and axis.

[0035] In some embodiments, steps S200 to S300, such as Figure 2As shown, the input consists of original projection data and defined geometric parameters. The reference coordinates for geometric calculation are determined by calculating the dynamic offset of the virtual projection points. The actual physical distance between pixels is used for basic geometric optical path correction. A relative depth factor is calculated to identify far-field regions far from the X-ray source. Using the calculated relative depth factor, nonlinear attenuation is applied to the far-field regions to construct nonlinear noise suppression weights. A composite weighted operation is performed on the original projection data, the geometric optical path correction weights, and the nonlinear noise suppression weights to obtain depth-denoised weighted projection data. By constructing a depth-dependent nonlinear denoising weighted field, geometric optical path errors caused by detector eccentricity translation can be corrected, while simultaneously using nonlinear weights to suppress far-field noise caused by large-angle scanning.

[0036] In some embodiments, step S200 may include, but is not limited to, steps S210 to S270: Step S210: Obtain the vertical distance from the X-ray source to the plane of the detector, the rotation radius of the detector's center around the Z-axis, the pixel size of the detector, and the projection angle. Step S220: Obtain the vertical projection point of the X-ray source on the plane of the detector to obtain the virtual projection point; Step S230: Obtain the dynamic offset of the virtual projection point based on the rotation radius, pixel size, and projection angle; Step S240: Obtain the true physical distance of the pixel based on the dynamic offset, vertical distance, and each pixel on the detector; Step S250: Obtain the relative depth factor based on the first offset of the dynamic offset in the v-axis direction, the first coordinate value of the pixel in the v-axis direction, the pixel size, and the maximum real physical distance of the pixel. Step S260: Construct nonlinear noise suppression weights based on the relative depth factor, suppression intensity coefficient, and nonlinear factor; Step S270: Based on the vertical distance and the actual physical distance of the pixels, obtain the geometric optical path correction weights, and fuse the geometric optical path correction weights with the nonlinear noise suppression weights to obtain a composite weight matrix.

[0037] In step S210 of some embodiments, geometric parameters are predefined, including the vertical distance from the ray source to the detector plane. (Right now The radius of rotation of the detector center around the Z-axis is... The detector pixel size is During the scanning process, when acquiring a single frame of raw projection data, the angular position of the X-ray source-detector system is the projection angle. .

[0038] In steps S220 to S230 of some embodiments, the vertical projection point of the ray source (or rotation axis) on the detector plane is the virtual projection point. This is based on the rotation radius. and pixel size For the first Each projection angle The local coordinate dynamic offset of the vertical projection point of the X-ray source (i.e., the point where the Z-axis passes through the detector plane) relative to the detector center is: The calculation formulas used include: ; .

[0039] In step S240 of some embodiments, based on the dynamic offset Vertical distance And every pixel on the detector The formulas used to calculate the true physical distance of pixels include: ; In the formula, This represents the actual physical distance of a pixel.

[0040] In step S250 of some embodiments, the dynamic offset is determined based on the first offset in the v-axis direction. The first coordinate value of the pixel in the v-axis direction Pixel size and the maximum pixel real physical distance The relative depth factor is calculated. This process identifies the "distance" of a pixel from the ray source. Defining the direction radially away from the source as the depth-increasing direction, the following calculation formula applies: ; In the formula, This represents the relative depth factor.

[0041] In step S260 of some embodiments, based on the relative depth factor Suppression intensity coefficient and nonlinear factors A nonlinear noise suppression weight is constructed. Optionally, the suppression strength coefficient can be set to 0.2~0.4, and the nonlinear factor can be set to 2.0. Through the nonlinear noise suppression weight, the near-end signal remains unchanged, while the high-noise signal at the far end is nonlinearly suppressed. For example, the formula used to calculate the nonlinear noise suppression weight includes: ; In the formula, Represents the weights for suppressing nonlinear noise; This indicates taking the maximum value.

[0042] In step S270 of some embodiments, based on the vertical distance and pixel-to-pixel physical distance The geometric optical path correction weights can be obtained. And the geometric optical path correction weight and the nonlinear noise suppression weight are combined. By fusing the components, a composite weight matrix can be obtained. The formulas used include: ; In the formula, express.

[0043] In some embodiments, step S300 may include, but is not limited to, step S310: Step S310: The original projection data is fused with the composite weight matrix to obtain the weighted projection data with deep denoising.

[0044] In step S310 of some embodiments, the original projection data is... Composite weight matrix The data is then fused to generate deeply denoised weighted projection data. The formulas used include: ; In the formula, This represents weighted projection data.

[0045] In step S400 of some embodiments, such as Figure 3 As shown, the input scanning angle is used to calculate the instantaneous tangent angle. Then, based on the incremental formula of the rotated circumscribed rectangle, the minimum fill width is analytically calculated. The weighted projection data is then filled around its perimeter according to the minimum fill width to obtain an intermediate processed image. This process abandons fixed fill and maximizes computational efficiency by calculating the minimum fill amount required for each frame in real time.

[0046] In standard cone-beam or fan-beam CT geometry where the detector rotates around the Z-axis, the projection angle and the scanning angle are the same; both refer to the angular position to which the X-ray source-detector system rotates to acquire a projection image. .

[0047] In some embodiments, step S400 may include, but is not limited to, steps S410 to S460: Step S410: Obtain the angle between the tangent direction of the scanning trajectory and the u-axis direction to obtain the instantaneous tangent angle; Step S420: Obtain the first number of pixels of the detector in the u-axis direction and the second number of pixels of the detector in the v-axis direction; Step S430: Construct a dynamic fill width model based on the instantaneous tangent angle, the number of first pixels, the number of second pixels, and the safety margin; Step S440: Obtain the minimum fill width based on the dynamic fill width model; Step S450: Fill the weighted projection data around the perimeter according to the minimum fill width to obtain the filled image; Step S460: Based on bilinear or bicubic interpolation, the filled image is rotated by an instantaneous tangent angle using a relaxed mode to obtain an intermediate processed image.

[0048] In step S410 of some embodiments, the scanning trajectory is a circular motion, and the tangent direction of the scanning trajectory is perpendicular to the radial direction. Regarding the scanning angle... The angle between the tangent direction and the detector's traveling direction (u-axis) is the instantaneous tangent angle, and the following formula applies: ; In the formula, It represents the instantaneous tangent angle.

[0049] In step S420 of some embodiments, the number of first pixels of the detector in the u-axis direction is obtained. That is, the width of the image. And obtain the number of the second pixel of the detector in the v-axis direction. That is, the height of the image.

[0050] In steps S430 to S440 of some embodiments, the image is rotated according to the geometric principle of the circumscribed rectangle of a square rotation. The half-width increment required to avoid information loss is related to the sine of the angle. Therefore, based on the instantaneous tangent angle... Number of first pixels Number of second pixels and safety margin (e.g., 20 pixels), construct the following dynamic fill width model: ; In the formula, Indicates the minimum padding width. When hour, (Very small); when hour, The maximum value is reached. This embodiment of the invention calculates the minimum padding amount required for each frame in real-time, saving an average of approximately 20% to 30% of memory and FFT computation compared to traditional fixed maximum padding.

[0051] In step S450 of some embodiments, during the edge replication padding process, the minimum padding width is used. For width, weighted projection data Fill the four sides to obtain the filled image. Alternatively, edge replication can be used, where the pixel values ​​of the filled region are equal to the pixel values ​​of the image boundary (zero-order hold). ; In the formula, Represents the filled region image; To fill the area; This is the original region.

[0052] In step S460 of some embodiments, the filled image is... Rotation The angle, a negative angle, represents a clockwise rotation within the image coordinate system (assuming a positive angle represents counter-clockwise rotation in standard mathematical convention), thus offsetting the tilt of the original image relative to the scan tangent direction. The rotation transformation causes the output image pixel positions to correspond to non-integer coordinates of the input image; therefore, pixel values ​​are calculated through interpolation, optionally using bilinear or bicubic interpolation. The rotation operation is performed in 'loose' mode, which automatically adjusts the output image canvas size to ensure that the entire rotated image data is fully contained without any information being cropped. The output image size is typically greater than or equal to the input image size. The final output is an intermediate processed image. At this point, the horizontal direction of the image is strictly parallel to the tangent direction of the scan trajectory.

[0053] In step S500 of some embodiments, such as Figure 4 As shown, the input intermediate image undergoes a one-dimensional fast Fourier transform to obtain frequency domain data. The interpolation loss is calculated using the instantaneous tangent angle, and then the angle-specific cutoff frequency is calculated. Dynamic filtering is applied using the frequency domain data and the angle-specific cutoff frequency to generate a compensated image. This full-angle adaptive spectral continuous compensation filtering process addresses the angle-dependent blur caused by discrete image rotation interpolation by establishing a continuous compensation model across the entire angle domain. This ensures that the projection data at any scanning angle achieves a consistent frequency domain response, realizing isotropic resolution.

[0054] In some embodiments, step S500 may include, but is not limited to, steps S510 to S560: Step S510: Perform a Fast Fourier Transform on the intermediate processed image to obtain frequency domain data; Step S520: Obtain the angle between the tangent direction of the scanning trajectory and the u-axis direction to obtain the instantaneous tangent angle; Step S530: Construct a full-angle interpolation loss function based on the instantaneous tangent angle and sensitivity coefficient; Step S540: Obtain the angle-specific cutoff frequency based on the reference cutoff frequency, the maximum compensation bandwidth, and the full-angle interpolation loss function; Step S550: Construct a dynamic filter based on the Haining window and the angle-specific cutoff frequency; Step S560: Filter the frequency domain data according to the dynamic filter to obtain the compensated image.

[0055] In step S510 of some embodiments, the intermediate processed image is... Perform a one-dimensional fast Fourier transform on each row to obtain frequency domain data. .

[0056] In some embodiments, step S520 is described with reference to the embodiment of step S410 described above, and will not be repeated here.

[0057] In step S530 of some embodiments, when the image rotates on the digital grid, the degree of high-frequency attenuation caused by interpolation is related to the instantaneous rotation angle. It exhibits a periodic, continuous, non-linear relationship. In grid alignment ( The loss is minimized when the shear is at its maximum ( ). The loss is maximized when the angle is 0. Therefore, we have the constructed full-angle interpolation loss function. The expression is: .in This is the sensitivity coefficient (usually taken as 1.0~2.0), and the function is in The changes are continuous and smooth.

[0058] In step S540 of some embodiments, to compensate for interpolation ambiguity at different angles, the filter cutoff frequency is set as a continuous function of the angle. Based on the reference cutoff frequency, the maximum compensation bandwidth, and the full-angle interpolation loss function, a linear gain model is used to determine the real-time angle-specific cutoff frequency. calculate: in, Indicates the reference cutoff frequency (e.g., 0.70), used for Low-loss angles are used to ensure the basic signal-to-noise ratio. Indicates the maximum compensated bandwidth (e.g., 0.15), used for Additional enhancement at the same height loss angle. The passband width of the filter changes with the scan angle. Smooth, continuous breathing-like adjustments are made between frames to ensure that the high-frequency retention of each frame is precisely matched with the interpolation loss at that angle.

[0059] In steps S550 to S560 of some embodiments, a variable bandwidth Ramp filter incorporating a Hann window is employed: ;in, This refers to the transformation of a projected image from the "spatial domain" to the "frequency domain" after undergoing a one-dimensional FFT (Fast Fourier Transform). This represents every point on this frequency axis.

[0060] In the formula, ; By filtering the frequency domain data using the constructed dynamic filter, the compensated image can be obtained. The filtering formulas used include: ; in, This operation extracts the real part of the number, forcibly discarding these meaningless imaginary parts to ensure the accuracy of the projected output data. It is a physical quantity (ray intensity value) consisting entirely of real numbers.

[0061] In step S600 of some embodiments, such as Figure 5 As shown, the input is the compensated image and the instantaneous tangent angle. The compensated image is rotated to restore the image coordinates. The restored image is then intelligently cropped to extract the geometric center region. Auxiliary data is automatically removed to generate the final projection.

[0062] In some embodiments, step S600 may include, but is not limited to, steps S610 to S630: Step S610: Obtain the angle between the tangent direction of the scanning trajectory and the u-axis direction to obtain the instantaneous tangent angle; Step S620: Perform an image rotation operation on the compensated image with a rotation angle equal to the instantaneous tangent angle to obtain the restored image; Step S630: Perform intelligent center cropping on the restored image to extract the geometric center region of the restored image and obtain the final projection.

[0063] In some embodiments, step S610 is described with reference to the embodiment of step S410 described above, and will not be repeated here.

[0064] In step S620 of some embodiments, the compensated image is rotated by an angle equal to the instantaneous tangent angle. For example, the compensated image... Rotate back to the original angle Thus, the restored image is obtained. .

[0065] In step S630 of some embodiments, since the filling and image rotation operations on the weighted projection data in the upstream steps are both centrally symmetric operations, the effective data is always located at the geometric center of the image. Based on the original size... It can calculate the clipping window, and the formulas used include: ; in, Represents the starting row index of the cropping window; The starting column index of the cropping window; Represents the filled image Total number of rows; Represents the filled image The total number of columns. Based on this cropping window, an intelligent center cropping operation is performed on the restored image to extract the geometric center region of the restored image, automatically removing [the unwanted columns]. Fill in the data to obtain the final projection data. .

[0066] In step S700 of some embodiments, the generated final projection data is substituted into the FDK back projection formula. This is for reconstructing spatial voxels. The projected coordinates are calculated, and depth-weighted summation is applied to obtain a three-dimensional tomographic image. The formulas used include: ; In the formula, Represents a three-dimensional tomographic image; Represents the coordinates of the voxel on the z-axis; Represents the vertical distance from the radiation source to the detector plane; Represents the projection angle The final projection below; Represents the horizontal projection coordinates of the voxel on the detector; Represents the vertical projection coordinates of the voxel on the detector.

[0067] In some embodiments, such as Figure 6 As shown, the processing procedure of an adaptive weighted filtering reconstruction method for dual-circular trajectory imaging of a X-ray source and detector is as follows: Step 1: By constructing a deep-dependent nonlinear denoising weighted field, the geometric optical path error caused by the detector's eccentric translation is corrected, and at the same time, the nonlinear weight is used to suppress the far-field noise caused by large tilt angle scanning.

[0068] Step 2: Dynamic minimum boundary extension based on geometric envelope. This method abandons the traditional fixed padding and calculates the minimum padding amount required for each frame in real time, maximizing computational efficiency.

[0069] Step 3: Eliminate the "frame" artifacts caused by edge truncation by rotating the image based on edge copy extension and align the filtering direction.

[0070] Step 4: Full-Angle Adaptive Spectral Continuous Compensation Filtering. Specifically, to address the angle-dependent blur caused by discrete rotation interpolation of the image, a continuous compensation model is established across the entire angle domain, ensuring that the projection data at any scanning angle can obtain a consistent frequency domain response, thus achieving isotropic resolution.

[0071] Step 5: Restore image coordinates and automatically remove auxiliary data by rotating and intelligent center cropping.

[0072] Step 6: Reconstruct the three-dimensional tomographic image by using FDK depth-weighted back projection.

[0073] This invention also provides an adaptive weighted filtering reconstruction device for dual-circular trajectory imaging of a radiation source and a detector, which can implement the above-described method. The device includes: The first module is used to acquire the raw projection data; The second module is used to construct a composite weight matrix by obtaining the dynamic offset of the virtual projection points; The third module is used to generate weighted projection data based on the original projection data and the composite weight matrix; The fourth module is used to perform perimeter filling and image rotation operations on the weighted projection data to obtain an intermediate processed image; The fifth module is used to perform full-angle adaptive spectrum continuous compensation filtering on the intermediate processed image to obtain the compensated image; The sixth module is used to perform image rotation and intelligent center cropping operations on the compensated image to obtain the final projection; The seventh module is used to perform weighted backprojection reconstruction on the final projection based on the FDK backprojection formula to obtain a three-dimensional tomographic image.

[0074] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0075] This invention also provides an electronic device, which includes a processor and a memory. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This electronic device can be any smart terminal, including a tablet computer, an in-vehicle computer, or similar device.

[0076] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0077] refer to Figure 7 , Figure 7 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 801 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present invention. The memory 802 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 802 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 802 and is called and executed by the processor 801. The 803 input / output interface is used to implement information input and output. The communication interface 804 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 805 transmits information between various components of the device (e.g., processor 801, memory 802, input / output interface 803, and communication interface 804); The processor 801, memory 802, input / output interface 803, and communication interface 804 are connected to each other within the device via bus 805.

[0078] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0079] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0080] This invention also provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions to cause the computer device to perform the aforementioned method.

[0081] In summary, the adaptive weighted filtering reconstruction method and apparatus for dual-circular trajectory imaging of a radiation source and detector according to embodiments of the present invention have the following advantages: 1. Existing technologies often result in tomographic images reconstructed with uneven brightness and excessive noise at the far end. This invention, through dynamic offset weighting and depth noise suppression, not only corrects the grayscale distribution but also significantly suppresses far-field noise, resulting in a cleaner image background and a substantial improvement in signal-to-noise ratio (SNR).

[0082] 2. Existing technologies typically use fixed filter parameters, resulting in reconstructed images that... The direction is clear, in Direction is blurred due to interpolation. The "full-angle adaptive spectrum compensation" mechanism of this invention calculates and matches the optimal filter bandwidth in real time for every tiny angle change. This continuous, breathing-like frequency domain adjustment not only compensates for the directional blur caused by interpolation, but also avoids interlayer stripes in the reconstructed image that may be caused by abrupt parameter changes, ensuring highly consistent sensitivity of precision defect detection in any direction.

[0083] 3. Existing technologies use zero-fill rotation, resulting in bright white borders tens of pixels wide around the image. The edge replication and extension strategy of this invention completely eliminates this artifact, making the image edge areas as clear and usable as the center, significantly expanding the effective detection range (FOV).

[0084] 4. Compared to existing technologies that blindly use fixed large-size filling, the "dynamic boundary calculation" strategy of this invention automatically optimizes the data volume based on the angle. or Nearby, the amount of invalid data processed was reduced by about 30%, significantly improving the overall reconstruction speed of the system.

[0085] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and sub-operations described as part of a larger operation are executed independently.

[0086] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0087] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0088] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0089] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0090] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0091] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0092] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0093] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. An adaptive weighted filtering reconstruction method for dual-circular trajectory imaging of a radiation source and detector, characterized in that, Includes the following steps: Obtain the raw projection data; A composite weight matrix is ​​constructed by obtaining the dynamic offset of the virtual projection points; Based on the original projection data and the composite weight matrix, weighted projection data is generated; The weighted projection data is subjected to perimeter filling and image rotation operations to obtain an intermediate processed image; The intermediate processed image is subjected to full-angle adaptive spectrum continuous compensation filtering to obtain the compensated image; The compensated image is subjected to image rotation and intelligent center cropping operations to obtain the final projection; Based on the FDK back projection formula, the final projection is reconstructed by weighted back projection to obtain a three-dimensional tomographic image.

2. The method according to claim 1, characterized in that, The process of constructing a composite weight matrix by obtaining the dynamic offset of the virtual projection points includes the following steps: Obtain the vertical distance from the X-ray source to the plane of the detector, the radius of rotation of the center of the detector around the Z-axis, the pixel size of the detector, and the projection angle; Obtain the vertical projection point of the ray source on the plane of the detector to obtain the virtual projection point; The dynamic offset of the virtual projection point is obtained based on the rotation radius, the pixel size, and the projection angle. The true physical distance of a pixel is obtained based on the dynamic offset, the vertical distance, and each pixel on the detector. The relative depth factor is obtained based on the first offset of the dynamic offset in the v-axis direction, the first coordinate value of the pixel in the v-axis direction, the pixel size, and the maximum real physical distance of the pixel. Based on the relative depth factor, suppression intensity coefficient, and nonlinear factor, a nonlinear noise suppression weight is constructed. Based on the vertical distance and the actual physical distance of the pixel, the geometric optical path correction weight is obtained, and the geometric optical path correction weight is fused with the nonlinear noise suppression weight to obtain the composite weight matrix.

3. The method according to claim 1, characterized in that, The step of generating weighted projection data based on the original projection data and the composite weight matrix includes the following steps: The original projection data is fused with the composite weight matrix to obtain the weighted projection data with deep denoising.

4. The method according to claim 1, characterized in that, The process of performing perimeter padding and image rotation operations on the weighted projection data to obtain an intermediate processed image includes the following steps: Obtain the angle between the tangent direction of the scan trajectory and the u-axis direction to get the instantaneous tangent angle; Obtain the first number of pixels of the detector in the u-axis direction and the second number of pixels of the detector in the v-axis direction; A dynamic fill width model is constructed based on the instantaneous tangent angle, the number of the first pixel, the number of the second pixel, and the safety margin; Based on the dynamic fill width model, obtain the minimum fill width; Based on the minimum fill width, the weighted projection data is filled around its perimeter to obtain the filled image; Based on bilinear or bicubic interpolation, the filled image is rotated by an angle equal to the instantaneous tangent angle using a relaxed mode to obtain the intermediate processed image.

5. The method according to claim 1, characterized in that, The process of performing full-angle adaptive spectral continuous compensation filtering on the intermediate processed image to obtain the compensated image includes the following steps: The intermediate processed image is subjected to a Fast Fourier Transform to obtain frequency domain data; Obtain the angle between the tangent direction of the scan trajectory and the u-axis direction to get the instantaneous tangent angle; Based on the instantaneous tangent angle and sensitivity coefficient, a full-angle interpolation loss function is constructed; The angle-specific cutoff frequency is obtained based on the reference cutoff frequency, the maximum compensation bandwidth, and the full-angle interpolation loss function. Based on the Haining window and the angle-specific cutoff frequency, a dynamic filter is constructed; The frequency domain data is filtered using the dynamic filter to obtain the compensated image.

6. The method according to claim 1, characterized in that, The step of performing image rotation and intelligent center cropping operations on the compensated image to obtain the final projection includes the following steps: Obtain the angle between the tangent direction of the scan trajectory and the u-axis direction to get the instantaneous tangent angle; The compensated image is rotated by an angle equal to the instantaneous tangent angle to obtain the restored image; A smart center cropping operation is performed on the restored image to extract the geometric center region of the restored image, thus obtaining the final projection.

7. The method according to claim 1, characterized in that, The weighted back-projection reconstruction based on the FDK back-projection formula yields a three-dimensional tomographic image. The formula used includes: ； In the formula, Represents a three-dimensional tomographic image; Represents the coordinates of the voxel on the z-axis; Represents the vertical distance from the radiation source to the detector plane; Represents the projection angle The final projection below; Represents the horizontal projection coordinates of the voxel on the detector; Represents the vertical projection coordinates of the voxel on the detector.

8. An adaptive weighted filtering reconstruction device for dual-circular trajectory imaging of a radiation source and a detector, characterized in that, include: The first module is used to acquire the raw projection data; The second module is used to construct a composite weight matrix by obtaining the dynamic offset of the virtual projection points; The third module is used to generate weighted projection data based on the original projection data and the composite weight matrix; The fourth module is used to perform perimeter filling and image rotation operations on the weighted projection data to obtain an intermediate processed image; The fifth module is used to perform full-angle adaptive spectrum continuous compensation filtering on the intermediate processed image to obtain the compensated image; The sixth module is used to perform image rotation and intelligent center cropping operations on the compensated image to obtain the final projection; The seventh module is used to perform weighted back projection reconstruction on the final projection based on the FDK back projection formula to obtain a three-dimensional tomographic image.

9. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images