A scatter correction method and a scatter correction system

Abstract

The disclosure provides a scattering correction method and a scattering correction system, and belongs to the technical field of medical imaging. The method comprises the following steps: determining a first region and a target beam blocking array scattering image based on an original projection image, wherein the total attenuation of the rays in the first region is greater than or equal to an attenuation threshold, and the original projection image comprises a projection image with a shadow region formed after rays blocked by the beam blocking array are projected towards a detection object; using the pixels corresponding to the first region in the target beam blocking array scattering image to optimize an initial scattering kernel parameter set to obtain an optimized scattering kernel parameter set; determining a scattering kernel superimposed scattering image based on the optimized scattering kernel parameter set and the original projection image; and performing scattering correction on the original projection image based on the scattering kernel superimposed scattering image. The scheme combines the advantages of the BSA method and the SKS method in estimating the scattering distribution, and can improve the precision of scattering correction.

Description

Technical Field

[0001] This disclosure relates to the field of medical imaging technology, and in particular to a scattering correction method and a scattering correction system. Background Technology

[0002] Cone-beam computed tomography (CBCT) offers advantages such as high imaging speed, high spatial resolution, and low radiation dose, and has been widely used in clinical fields such as dentistry, radiotherapy, and interventional surgery. However, during CBCT imaging, the interaction between X-ray photons and the scanned object generates a large number of scattered photons. These scattered photons are mixed with the X-rays and are received by the detector, leading to problems such as decreased contrast, inaccurate CT values, and increased artifacts in the reconstructed images, seriously affecting the accuracy of clinical diagnosis.

[0003] Currently, mainstream scattering correction methods are mainly divided into two categories: hardware correction methods and software correction methods. Beam Stop Array (BSA) is a commonly used hardware correction method. It uses a high-attenuation material dot array to block rays, and measures the scattered signal using the resulting shadow area, then interpolates to obtain the full-image scattering distribution. However, the measurement accuracy of this method is significantly affected by the thickness of the scanned object. In thin tissue regions and at tissue boundaries, due to the penumbra effect and high-frequency changes in the scattering signal, the scattering measurement error is large. Scatter Kernel Superposition (SKS) is a commonly used software correction method. It estimates the scattering distribution by convolving a pre-measured scattering kernel with the projected image. However, this method relies on pre-set fixed kernel parameters, making it difficult to adapt to individual differences in different scanned objects, resulting in poor generalization ability and robustness. Summary of the Invention

[0004] This disclosure provides a scattering correction method and a scattering correction system; by combining the advantages of the BSA method and the SKS method in estimating the scattering distribution, the accuracy of scattering correction can be improved.

[0005] The technical solution disclosed herein is implemented as follows: In a first aspect, this disclosure provides a scattering correction method, which includes: determining a first region and a target beam blocking array scattering image based on an original projection image, wherein the total attenuation of rays in the first region is greater than or equal to an attenuation threshold, and the original projection image includes a projection image with a shadowed region formed after rays projected onto the detection object are blocked by the beam blocking array; optimizing an initial scattering kernel parameter set using pixels in the target beam blocking array scattering image corresponding to the first region to obtain an optimized scattering kernel parameter set; determining a scattering kernel superimposed scattering image based on the optimized scattering kernel parameter set and the original projection image; and performing scattering correction on the original projection image based on the scattering kernel superimposed scattering image.

[0006] Secondly, this disclosure provides a scattering correction apparatus, comprising: a determining part configured to determine a first region and a target beam blocking array scattering image based on an original projection image, wherein the total ray attenuation of the first region is greater than or equal to an attenuation threshold, and the original projection image includes a projection image with a shadowed region formed after the rays projected onto the detection object are blocked by the beam blocking array; an optimizing part configured to optimize an initial scattering kernel parameter set using pixels in the target beam blocking array scattering image corresponding to the first region to obtain an optimized scattering kernel parameter set; the determining part further configured to determine a scattering kernel superimposed scattering image based on the optimized scattering kernel parameter set and the original projection image; and a correction part configured to perform scattering correction on the original projection image based on the scattering kernel superimposed scattering image.

[0007] Thirdly, this disclosure provides a scattering correction system, comprising: an imaging source configured to emit a ray beam toward a detection object; a detector disposed opposite to the imaging source, the detector being configured to receive the ray beam passing through the detection object and generate a raw projection image; a beam blocking array disposed on the ray beam path between the imaging source and the detector and configured to block a portion of the ray beam directed toward the detection object; and an electronic device communicatively connected to the detector to acquire the raw projection image from the detector and perform the steps of the scattering correction method described in the first aspect.

[0008] Fourthly, this disclosure provides an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the scattering correction method as described in the first aspect.

[0009] Fifthly, this disclosure provides a computer-readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the scattering correction method as described in the first aspect.

[0010] This disclosure provides a scattering correction method and system. The method utilizes the positive correlation between the measurement accuracy of the BSA method and the total X-ray attenuation. Pixels in the first region of the target BSA scattering image with high total X-ray attenuation are used as reference ground values ​​to optimize the initial scattering kernel parameter set. Then, the optimized scattering kernel parameter set is used to calculate the SKS scattering image, which is finally used to correct the scattering of the original projection image. This approach enables accurate scattering estimation results even in regions with low total X-ray attenuation. The resulting superimposed scattering kernel image combines the high-precision measurement results of the BSA method in regions with high total X-ray attenuation with the continuous estimation advantages of the SKS method in regions with low total X-ray attenuation and tissue boundary regions, effectively improving the consistency and accuracy of scattering estimation across the entire region. Attached Figure Description

[0011] Figure 1 This is a schematic diagram of an exemplary scattering correction system that enables the implementation of the technical solution disclosed herein.

[0012] Figure 2 This is a schematic diagram of the arrangement of blocking elements in a BSA provided in this disclosure.

[0013] Figure 3 This is a schematic diagram of the arrangement of the blocking elements in another BSA provided in this disclosure.

[0014] Figure 4 This is a schematic flowchart of a scattering correction method provided in this disclosure.

[0015] Figure 5 This is a flowchart illustrating another scattering correction method provided in this disclosure.

[0016] Figure 6 This is a schematic diagram of Z-position stitching of a BSA scattering image and an SKS scattering image provided in this disclosure.

[0017] Figure 7 This is a schematic diagram of another Z-position stitching of BSA scattering image and SKS scattering image provided in this disclosure.

[0018] Figure 8 This is a schematic diagram of weight allocation for a shaded region provided in this disclosure.

[0019] Figure 9 This is a structural block diagram of a scattering correction device provided in this disclosure.

[0020] Figure 10 This is a schematic diagram of the hardware structure of an electronic device provided in this disclosure. Detailed Implementation

[0021] The technical solutions of this disclosure will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0022] In the description of this disclosure, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this disclosure, the term "at least one" means one or more, and the term "multiple" means two or more, unless otherwise expressly specified.

[0023] In this disclosure, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0024] It should be understood that obtaining B based on A does not mean obtaining B solely based on A; B can also be obtained based on A and / or other information.

[0025] It should be understood that the term "comprising," when used in this disclosure, specifies the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0026] It should also be understood that, in the various embodiments of this disclosure, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of this disclosure.

[0027] See Figure 1 , Figure 1 This is a schematic diagram of an exemplary scattering correction system 1 capable of implementing the technical solution disclosed herein. Figure 1 As shown, the scattering correction system 1 may include: an imaging source 10, a BSA 20, a detector 30, and an electronic device 40.

[0028] In this disclosure, the imaging source 10 can be, but is not limited to, an X-ray source or other radiation source capable of generating X-rays. Taking the scatter correction system 1 as an example of a CBCT device, the imaging source 10 can generate a cone-beam X-ray, as shown by the solid arrow. This cone-beam X-ray can scan the object 50, which can be a human body, a phantom, an animal, etc.—any object that can be imaged under the scatter correction system 1.

[0029] In this disclosure, detector 30 is used to receive the cone-beam rays emitted by imaging source 10, convert the received cone-beam rays into electrical signals through photoelectric conversion, and then convert them into digital signals through an analog / digital converter, thereby obtaining the projected image data corresponding to the object 50 being detected.

[0030] For example, in a specific implementation, the imaging source 10 emits a cone-shaped beam of light along a first direction to scan the object 50. The detector 30 then receives the light rays passing through the object 50, converts the optical signal into an electrical signal via photoelectric conversion, and then converts it into a digital signal via an analog-to-digital converter, thereby obtaining the projected image data of the object 50. After obtaining the projected image data of the object 50, an image can be reconstructed based on this data.

[0031] However, when the imaging source 10 scans the object 50, the cone-beam rays will scatter on the object 50, affecting the quality of the reconstructed image. To reduce the impact of scattering on the quality of the reconstructed image, an image beam shield (BSA) 20 can be provided, for example, between the detector 30 and the imaging source 10. In some examples, the BSA 20 can be positioned on the ray beam path between the imaging source 10 and the object 50. In some examples, the BSA 20 can be positioned between the object 50 and the detector 30; specifically, it can be positioned according to actual needs. Furthermore, the BSA 20 is movably positioned on the ray beam path and switches between a working position that cuts into the ray beam and a non-working position that cuts out the ray beam. That is, when the BSA 20 is in the working position that cuts into the ray beam, it can block part of the ray beam, and when it is in the non-working position that cuts out the ray beam, it will not block the ray beam.

[0032] In this disclosure, the BSA 20 is configured to be parallel to the illumination surface of the imaging source 10, i.e., the BSA 20 is configured along a second direction perpendicular to the first direction. In some examples, the BSA 20 can be rotated by a preset angle, such as 45 degrees, around its own central axis (which is parallel to the third direction) in the plane containing the second and third directions.

[0033] In addition, the BSA 20 is provided with multiple blocking elements 21. The position of each blocking element can be arbitrarily set; that is, the position of each blocking element is not restricted, as long as there are empty spaces between the blocking elements. In this disclosure, the blocking elements can be arranged in an m×n array, where m and n are both integers greater than 0, and m can be equal to n or not equal to n. Of course, the blocking elements can also be arranged arbitrarily. For example, as shown... Figure 2 As shown, BSA 20 can be composed of 9×9 (81 in total) lead material blocking elements. That is, BSA 20 has 9 rows of blocking elements along the second direction, and each row includes 9 blocking elements arranged along a third direction. The third direction is perpendicular to both the first and second directions, meaning the first, second, and third directions are orthogonal to each other. Figure 1 In the diagram, the third direction is represented by the symbol "×", which means that the third direction is perpendicular to the plane defined by the first and second directions.

[0034] Exemplarily, the blocking element 21 can be a regular shape such as a cylinder or a sphere, and in this disclosure, the blocking element 21 can be made of a high-attenuation material, such as lead or tungsten, which can prevent rays from passing through, thereby forming a shadow area on the detector 30. On the BSA 20, areas 22 outside the multiple blocking elements 21 (i.e., areas not containing the blocking elements 21) can be made of materials with low attenuation properties, such as tempered glass, carbon fiber, acrylic, etc., which allow rays to pass through these areas onto the detector 30. In some examples, such as... Figure 3 As shown, all the blocking elements are embedded towards the imaging source 10, and the intersection of the central axes of each blocking element coincides with the focal point of the imaging source 10. This ensures that the shadow area of ​​each blocking element 21 is consistent and the ray attenuation path is consistent, ensuring that the blocking elements completely block the rays emitted by the imaging source 10, thereby improving the accuracy of subsequent scattering estimation.

[0035] In this disclosure, when there is no object 50 to be detected between the imaging source 10 and the detector 30, the rays emitted by the imaging source 10 only pass through the BSA 20, and the rays are blocked by the blocking member 21 on the BSA 20. As a result, the projected image formed on the detector 30 includes 81 shadow areas corresponding to the blocking member 21 on the BSA 20. Visually, the shadow areas appear as shadow points.

[0036] In this disclosure, when a detection object 50 exists between the imaging source 10 and the detector 30, for example... Figure 1The object to be detected, 50, is positioned between the BSA 20 and the detector 30. After the rays emitted from the imaging source 10 pass sequentially through the BSA 20 and the object to be detected, the projected image formed on the detector 30 includes the projection of the object to be detected, in addition to the 81 shadowed areas corresponding to the blocking element 21. Since the shadowed areas cannot receive the ray signals emitted by the imaging source 10, when a ray signal is detected within a shadowed area in the projected image, it can be assumed that the ray signal within that shadowed area is a scattered signal caused by the object to be detected 50. In some examples, the signal detected at the center of a shadowed area can be considered as the ray signal detected within the shadowed area.

[0037] Based on this, in order to determine the overall scattering signal distribution in the projected image, this disclosure connects the electronic device 40 to the detector 30 to receive the projected image generated by the detector 30. Using the BSA method and leveraging the low-frequency characteristics of scattering, the overall scattering signal distribution in the projected image can be estimated by interpolation based on the scattering signal from the shadow region. In some examples, each shadow region in the projected image can be considered as a scattering sampling point corresponding to the blocking element. Subsequently, based on the scattering signals of every two adjacent scattering sampling points, interpolation is performed in the gaps between every two adjacent scattering sampling points to obtain interpolated sampling points. Then, based on the scattering signals of each scattering sampling point and each interpolated sampling point, a scattering distribution map corresponding to the projected image is obtained.

[0038] In some examples, electronic device 40 can be a standalone server, or a server network or server cluster composed of servers. For example, the computer equipment described in the embodiments of this application includes, but is not limited to, computers, network hosts, single network servers, multiple network server sets, or cloud servers composed of multiple servers. Among them, cloud servers are composed of a large number of computers or network servers based on cloud computing.

[0039] In some examples, the aforementioned electronic device 40 can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a PDA (Personal Digital Assistant), a mobile phone, a tablet computer, a wireless terminal device, a communication device, an embedded device, etc. This embodiment does not limit the type of computer device.

[0040] In some examples, electronic device 40 and detector 30 can communicate through any communication method, including but not limited to: mobile communication based on Wideband Code Division Multiple Access (WCDMA), Evolved Universal Terrestrial Radio Access Network (E-UTRAN), Next Generation Radio Access Network (NG-RAN), Long Term Evolution (LTE), Worldwide Interoperability for Microwave Access (WiMAX), future 5th Generation (5G) systems such as New Radio Access Technology (NR), and future communication systems such as 6G; or computer network communication based on TCP / IP Protocol Suite (TCP / IP), User Datagram Protocol (UDP); or Bluetooth networks, ZigBee networks, Near Field Communication (NFC) networks, or any combination thereof.

[0041] like Figure 1 As shown, in some examples, the scattering correction system 1 may also include a support frame 60 for supporting, positioning, and moving the detection object 50. This can be achieved by placing the detection object 50 on the support frame 60 and moving the support frame 60 to position the detection object 50 between the imaging source 10 and the detector 30, or by moving the support frame 60 to position the detection object 50 outside the imaging source 10 and the detector 30. For example, in a medical context, the support frame 60 may be a three-dimensional, four-dimensional, five-dimensional, or six-dimensional treatment bed or chair, etc.

[0042] It should be noted that the imaging source 10, BSA 20, and detector 30 can also be mounted on the gantry of a radiotherapy device. The gantry is equipped with a radiotherapy head that emits therapeutic rays for radiotherapy of the recipient. Rotation of the gantry allows the imaging source 10, BSA 20, and detector 30 to scan the recipient (i.e., the detection object 50) from multiple angles, and the detector 30 can acquire a non-empty field projection image formed by the rays passing through the detection object.

[0043] In the specific implementation of scattering correction, the BSA method assumes that the blocking element 21 is completely opaque and the X-ray source is an ideal point source. Therefore, only scattered signals are present in the shadow area corresponding to the blocking element 21 on the detector 30. However, in reality, the X-ray source is a focal point with a certain physical size (typically 0.1mm to 1mm). Rays emitted from different positions of the focal point are blocked by different edges of the blocking element 21, forming a penumbra around the shadow where some rays are blocked. The penumbra width is determined only by the system hardware parameters and is independent of the scanned object. The formula for calculating the penumbra width is: ,in For the focal size, The distance from the focal point to BSA 20. This refers to the distance from BSA 20 to detector 30.

[0044] The presence of X-ray signals mixed in within the penumbra can cause BSA measurements to be overestimated. The relative magnitude of this measurement error is related to the total attenuation of the X-rays at that location. When X-rays pass through regions with high total attenuation, the X-ray intensity decreases significantly, and the scattered signal constitutes a large proportion of the total signal. In this case, the intensity of the X-ray signal mixed in within the penumbra is far lower than the actual scattered signal intensity, and its impact on the scattering measurement results is negligible. For example, when X-rays pass through a 20cm thick abdominal wall, the X-ray intensity attenuates by more than 99%. At this point, the scattered signal intensity can reach more than 50% of the original X-ray intensity, and the weak X-ray signal mixed in within the penumbra accounts for only 1% to 5% of the actual scattered value.

[0045] When X-rays pass through regions with low total attenuation, the intensity of the radiation decreases only slightly, resulting in a low proportion of the scattered signal in the total signal. In this case, the intensity of the radiation signal mixed in within the penumbra approaches or even exceeds the true scattered signal intensity, leading to significant deviations in BSA scattering measurements. For example, when radiation passes through a 2cm thick arm, the intensity decreases by only about 10%, resulting in a scattered signal intensity of less than 5% of the radiation intensity. The strong radiation signal mixed in within the penumbra can account for 50% to 200% of the true scattered value.

[0046] The SKS method is an algorithm that estimates scattering distribution through mathematical convolution. It assumes that the scattered signal at a point is the superposition of scattering from all surrounding pixels, mathematically represented as the convolution of the projected image and the scattering kernel. The characteristics of the scattering kernel are defined by a set of kernel parameters. However, the SKS method uses a fixed set of kernel parameters pre-calibrated on a standard phantom, which cannot be dynamically adapted to the characteristics of different scanned objects. In reality, the scattering kernel parameter set is closely related to the individual characteristics of the scanned object, such as the total attenuation distribution of X-rays, its overall size, and tissue composition. These characteristics vary significantly among different scanned objects, and the corresponding optimal scattering kernel parameter set will change accordingly.

[0047] The existing SKS method uses a uniform set of fixed scattering kernel parameters to estimate the scattering of all scanned objects, which cannot match the actual scattering characteristics of different objects. This results in a deviation between the calculated SKS scattering image and the real scattering distribution, thus affecting the final scattering correction effect.

[0048] Therefore, this disclosure provides a scattering correction method that combines the advantages of the BSA and SKS methods, see [link to relevant documentation]. Figure 4 It shows a flowchart of a scattering correction method provided in this disclosure, which is applied to... Figure 1 The electronic device 40 in the dispersion correction system 1 shown includes steps S401 to S404 in the dispersion correction method.

[0049] In step S401, the first region and the target BSA scattering image are determined based on the original projected image.

[0050] In this context, the total X-ray attenuation in the first region is greater than or equal to the attenuation threshold. Total X-ray attenuation refers to the total amount of intensity reduction after X-rays pass through all tissues at the corresponding location of the object being examined, reflecting the overall attenuation capability of the tissue at that location for X-rays. The attenuation threshold is a pre-set value used to divide regions into different total X-ray attenuation regions. It can be determined by measuring the accuracy of BSA scattering images under multiple different total X-ray attenuation values, identifying the total X-ray attenuation value corresponding to a significant decrease in accuracy as the attenuation threshold.

[0051] Because the total intensity attenuation of X-rays varies depending on factors such as density, thickness, or anatomical organ type, the total attenuation can also be characterized by density, thickness (e.g., equivalent water thickness), or anatomical region type. Correspondingly, the attenuation threshold can be characterized by thresholds for density, thickness, or anatomical region type, thus defining the first region based on these factors. For example, a region with a density greater than or equal to a density threshold can be defined as the first region, or a region with a thickness greater than or equal to a thickness threshold can be defined as the first region.

[0052] The original projection image is the projection image received by the detector after passing through the object being detected and the BSA. It includes the projection image with a shadowed area formed after the rays projected onto the object by the BSA. The original projection image includes both ray signals and scattered signals.

[0053] The target BSA scattering image is an image reflecting the distribution of scattered signal intensity at each location, calculated based on the original projection image. In some implementations, the scattering measurement values ​​of the shaded areas in the original projection image can be extracted and then interpolated to obtain the scattering signal distribution of the entire image. Then, the total ray attenuation at each location is calculated based on the original projection image and compared with a preset attenuation threshold to identify the first region where the total ray attenuation is greater than or equal to the attenuation threshold. In this way, the first region with high measurement accuracy using the BSA method can be selected.

[0054] In step S402, the initial scattering kernel parameter set is optimized using the pixels in the target BSA scattering image corresponding to the first region to obtain an optimized scattering kernel parameter set.

[0055] The initial scattering nucleus parameter set refers to a set of values ​​pre-calibrated on a standard phantom before scanning to define the characteristics of the scattering nucleus. It can only reflect the scattering characteristics of the standard phantom and differs from the scattering characteristics of different test objects in actual clinical practice.

[0056] The BSA scattering measurements in the first region are highly accurate and can be used as true values ​​reflecting the actual scattering characteristics of the target object. In some implementations, the initial scattering kernel parameter set can be iteratively adjusted to obtain an optimized scattering kernel parameter set when the difference between the SKS scattering image and the target BSA scattering image in the first region is minimized.

[0057] In step S403, the SKS scattering image is determined based on the optimized scattering kernel parameter set and the original projection image.

[0058] Specifically, the original projected image can be convolved with the scattering kernel corresponding to the optimized scattering kernel parameter set to obtain the full-image SKS scattering image. The scattering image generated by the SKS method through convolution is continuous and does not have the sparse sampling problem of the BSA method. It can also obtain a smooth and continuous scattering distribution in regions with low total X-ray attenuation and tissue boundaries.

[0059] Because of the use of an optimized scattering kernel parameter set, the generated SKS scattering image not only matches the measurement results of the BSA method in the high total ray attenuation region, but also obtains accurate scattering estimation results in the low total ray attenuation region, making up for the lack of accuracy of the BSA method in the low attenuation region.

[0060] In step S404, scattering correction is performed on the original projection image based on the SKS scattering image.

[0061] In some implementations, scattering correction is performed on the original projection image. This involves subtracting the pixel values ​​corresponding to the SKS scattering image from the original projection image pixel by pixel to obtain a corrected projection image containing only the X-ray signal, which is then used for subsequent CT reconstruction. For the first region with high total X-ray attenuation, the correction result inherits the high accuracy of the BSA method; for regions with low total X-ray attenuation, the correction result avoids the penumbra error and sampling error inherent in the BSA method. Furthermore, because the scattering kernel parameter set has been customized for the current detection object, the correction result does not exhibit the systematic bias inherent in the scattering kernel superposition method.

[0062] It should be noted that for each projected image included in the original projected image, a corrected image is obtained through a scattering correction method.

[0063] In this embodiment, leveraging the positive correlation between the measurement accuracy of the BSA method and the total X-ray attenuation, pixels in the first region of the target BSA scattering image with high total X-ray attenuation are used as reference ground values ​​to optimize the initial scattering kernel parameter set. Then, the optimized scattering kernel parameter set is used to calculate the SKS scattering image, which is finally used to correct the scattering of the original projection image. This method enables accurate scattering estimation results even in regions with low total X-ray attenuation. The resulting superimposed scattering kernel image combines the high-precision measurement results of the BSA method in regions with high total X-ray attenuation with the continuous estimation advantages of the SKS method in regions with low total X-ray attenuation and tissue boundary regions, effectively improving the consistency and accuracy of scattering estimation across the entire region.

[0064] Furthermore, the object to be detected may also include a second region where the total attenuation of rays is less than an attenuation threshold. Correspondingly, the method further includes: determining the second region based on the original projection image. The total attenuation of rays in the second region is less than the attenuation threshold; for example, the first region corresponds to a thick-woven area of ​​the object to be detected; the second region corresponds to a thin-woven area or object boundary of the object to be detected, and the first and second regions together cover the entire effective pixel range of the original projection image.

[0065] Since the total attenuation of X-rays varies when they pass through objects of different thicknesses, in some embodiments, a first region and a second region can be defined based on the thickness. Specifically, a thickness image is determined based on the original projection image, and regions in the thickness image with a thickness greater than a thickness threshold, such as thick tissue regions, are defined as the first region; regions in the thickness image with a thickness less than or equal to the thickness threshold, such as thin tissue regions, are defined as the second region.

[0066] A thickness image refers to an image that reflects the distribution of equivalent water thickness of the object under test in various projection directions. Equivalent water thickness refers to the thickness of a water layer with the same X-ray attenuation capability at the corresponding position of the object under test.

[0067] The specific process for determining the thickness image is as follows: acquire the empty field image after the ray is blocked by the BSA; interpolate the initial scattering signals of each shadow region in the original projection image to obtain the initial BSA scattering image; homogenize the initial BSA scattering image to obtain the uniform scattering image; and determine the thickness image based on the original projection image, the empty field image, and the uniform scattering image.

[0068] An empty-field image refers to the projected image of X-rays after being blocked by a BSA (Body Shielding System) when no object is placed. Each pixel value in the empty-field image represents the total intensity of X-rays received by the detector at the corresponding location when there is no object to attenuate the radiation. Empty-field images can be acquired using the exact same hardware parameters as subsequent scans with objects, before a clinical scan begins and without an object being placed.

[0069] After acquiring the open field image, the initial BSA scattering image is determined. For each shadow region in the original projection image, the average pixel value within a preset pixel range at its center is taken as the initial scattering signal for that shadow region. Taking the average value from the center region effectively reduces the impact of noise on the measurement value of individual pixels. The initial scattering signals of all shadow regions are used as discrete sampling points, and a cubic spline interpolation algorithm is used for two-dimensional interpolation to obtain an initial BSA scattering image with the same size as the original projection image. Cubic spline interpolation ensures that the interpolation result has continuous first and second derivatives throughout the entire image range, resulting in a smoother scattering distribution. In some embodiments, a bilinear interpolation algorithm can be used instead of cubic spline interpolation; the specific interpolation algorithm is not limited in this disclosure.

[0070] The initial BSA scattering image is homogenized. The initial BSA scattering image is traversed row by row. For each row of pixels, the minimum value of all pixels in that row is obtained. This minimum value is then copied along the row direction to generate a pixel sequence of the same length as the row. After processing all rows, the sequences are stitched together to obtain a complete uniform scattering image. This is because within the same row (i.e., the same Z-height), the minimum value of the scattered signal is closest to the uniform scattering background level across the entire scanning field of view. By taking the minimum value row by row and copying it, an image reflecting the overall uniform scattering level can be obtained. Alternatively, the process can be traversed column by column, and the minimum value of each column can be copied along the column direction to generate a pixel sequence of the same length as the column. This is suitable for scenarios where the tissue is relatively uniformly distributed vertically within the scanning field of view. Another approach is to take the global minimum value of the initial BSA scattering image and copy it to all pixel positions to generate a uniform scattering image. This is suitable for scenarios where the detected object is small and the scattering distribution is relatively uniform. The specific homogenization method is not specifically limited in this disclosure.

[0071] The thickness image is finally determined. It can be determined based on the Beer-Lambert law, using the original projection image, the empty field image, and the uniform scattering image. The specific calculation formula is as follows: ,in, Indicates the thickness of the image in coordinates The equivalent water thickness value at that location, This represents the linear attenuation coefficient of equivalent water under the current X-ray energy spectrum, typically taken as 0.186. This value can be pre-calibrated by scanning a standard water phantom. Indicates the empty field image in coordinates Pixel value at that location, Indicates the original projected image in coordinates Pixel value at that location, Represents the uniform scattering image in coordinates The pixel value at that location. This formula represents the approximate ray signal obtained by subtracting the uniformly scattered background from the original projection signal, and then converting the ray signal into an equivalent water thickness using the Beer-Lambert law.

[0072] Furthermore, standard water phantoms of varying thicknesses can be pre-scanned to establish a mapping curve between the ray signal and the equivalent water thickness. During actual calculations, an approximate ray signal for each pixel is obtained. Calculate the normalized ray signal for each pixel. For each normalized ray signal, a lookup table and linear interpolation are performed on a pre-calibrated mapping curve to obtain the corresponding equivalent water thickness value. This method effectively compensates for the nonlinear error caused by the X-ray energy spectral hardening effect, improving the accuracy of thickness calculation. Other calculation methods that can obtain thickness images can also be used, and this disclosure does not limit the specific methods.

[0073] This embodiment pre-corrects the original projection image by using a uniformly scattered image, eliminating most of the interference of the scattered signal on the thickness calculation. The resulting thickness image can more accurately reflect the true tissue thickness distribution of the detected object.

[0074] It should be noted that the aforementioned thickness threshold is predetermined. Specifically, when the thickness of the object through which the rays pass causes the rays to attenuate to a level comparable to the amount of leakage rays in the penumbra region, the measurement accuracy of BSA begins to decrease significantly. This turning point can be determined as the thickness threshold.

[0075] Since the equivalent water thickness of the tissue is positively correlated with the total attenuation of radiation, the regions segmented based on the thickness image are highly consistent with those segmented based on the total attenuation of radiation. This segmentation method does not require additional calculation of the total attenuation of radiation; it can be completed using the already generated thickness image, resulting in low computational cost and high efficiency. Furthermore, the accuracy of the thickness image has been pre-corrected using a uniform scattering image, further ensuring the accuracy of the region segmentation.

[0076] After obtaining the thickness image, the corresponding, such as Figure 5 As shown, step S401 can be implemented in the following step S501.

[0077] In step S501, the target BSA scattering image is determined based on the original projection image and the thickness image.

[0078] Specifically, for each shadow region in the original projection image, the average thickness of each shadow region in the neighborhood is determined in the thickness image; the distance of each shadow region from the boundary of the detected object is determined; based on the average thickness value, distance and initial scattering signal of the shadow region, the corrected scattering signal corresponding to each shadow region is determined through the physical mapping correction function to obtain the target BSA scattering image.

[0079] First, determine the average thickness, and then use a pre-calibrated BSA mask to obtain the center coordinates of all shadow areas in the original projection image. This coordinate is then mapped onto the thickness image to obtain the thickness coordinates at the corresponding location. A thickness coordinate system is defined with the center coordinates of the shadow region as the center. A square neighborhood is defined. In some implementations, N can range from 15 to 20. For example, for a BSA with a spacing of 1 cm, a 15×15 neighborhood can cover a sufficiently large area around the shadow region while avoiding the introduction of irrelevant information from distant areas. The thickness values ​​of all pixels within the neighborhood are extracted, and the arithmetic mean is calculated as the average thickness value corresponding to the shadow region. The formula for calculating the average thickness value is: Average thickness reflects the overall attenuation characteristics of the tissue surrounding the shaded area.

[0080] Using the neighborhood average thickness instead of the thickness of a single pixel can effectively reduce the impact of noise and local fluctuations in the thickness image on the correction results. At the same time, the scattering signal corresponding to the shadow region is actually the sum of the scattering generated by the tissue in a certain area around it, and the average thickness can better reflect the overall attenuation characteristics of the region.

[0081] In some implementations, different weights can be assigned to pixels at different locations within the neighborhood. Pixels closer to the center of the shadow region have higher weights, while pixels farther away have lower weights. For example, a Gaussian weighting function can be used. The size of the neighborhood can also be adaptively adjusted according to the degree of local variation in the thickness image; a larger neighborhood is used in areas with gentle thickness changes, while a smaller neighborhood is used in areas with drastic thickness changes. This approach can maintain noise suppression capabilities while avoiding the introduction of excessive mixing information in boundary regions. The specific method for determining the average thickness is not limited in this disclosure.

[0082] Then, the distance between the shadow region and the boundary is determined. Thresholding segmentation is performed based on the thickness image to obtain a binarized object mask. In some implementations, the Otsu method can be used to automatically determine the segmentation threshold. Pixels with a thickness value greater than the threshold are marked as object regions (value 1), and pixels with a thickness value less than the threshold are marked as background regions (value 0). A distance transform is performed on the binarized object mask, and the Euclidean distance from each pixel to the boundary of the nearest detected object is calculated to obtain a distance image, where the value of each pixel represents the shortest distance from that pixel to the boundary of the detected object. The center coordinates of the shadow region are mapped to the distance image, and the distance value at the corresponding position is extracted as the distance of the shadow region from the boundary of the detected object. The scattered signal changes drastically at the boundary; the closer to the boundary, the more obvious the change in the scattered signal. By calculating the distance from the shadow region to the boundary, the influence of the boundary effect on the scattering measurement can be quantified. The distance between the shadow region and the boundary reflects the degree of drastic change in the scattered signal at the location of the shadow region.

[0083] In some implementations, the Canny edge detection algorithm can be used first to extract the edge contours of objects from the thickness image, and then the Euclidean distance from each shadow region to the nearest boundary pixel can be calculated. This approach can more accurately locate the fine boundaries of objects and is suitable for detecting objects with complex boundaries. Alternatively, the thickness image can be decomposed into multiple scales, with boundaries extracted and distances calculated at different scales, and then the distances from different scales weighted and fused. This approach can simultaneously capture boundary features at both large and small scales, improving the robustness of distance calculation. The specific method for calculating the distance between shadow regions and boundaries is not limited in this disclosure.

[0084] Finally, each shaded area is corrected. The physical mapping correction function describes the mapping relationship between the scattering signal measured by BSA and the actual scattering signal. Through pre-calibrated physical mapping correction functions, the mapping relationship between the measured scattering signal and the actual scattering signal under different thicknesses and boundary distances can be accurately reflected, providing a reliable basis for scattering signal correction during online scanning.

[0085] In some implementations, the physical mapping correction function can be obtained through pre-calibration. The calibration process can be performed according to the following steps: First, for each phantom of different thicknesses (e.g., a water plate phantom), at different horizontal positions relative to each phantom by the BSA, acquire the calibration BSA measurement signal corresponding to the original calibration image formed by the ray being blocked by the BSA and projected onto each phantom. For example, place water plate phantoms of different thicknesses at an isocenter and move the water plate left and right so that the BSA blocking element is located at multiple boundary distances of the water plate phantom. Second, acquire the conical beam projection image with the beamguide in the fully open state and the fan-shaped beam projection image with the beamguide in the head-to-foot direction minimized. Then, based on the difference between the conical beam projection image and the fan-shaped beam projection image, determine the true scattering signal of the calibration beam blocking array. Since there is almost no scattering signal in the fan-shaped beam projection, the difference between the two can be approximated as the true scattering signal. Finally, fit the calibration BSA measurement signal and the corresponding true scattering signal of the calibration beam blocking array to generate the physical mapping correction function.

[0086] Specifically, a two-dimensional quadratic polynomial function can be used as the physical mapping correction function, in the form of: ,in, The corrected scattering signal, The average thickness of the neighborhood of the shaded region. This represents the distance from the shaded area to the boundary. to These are the pre-calibrated polynomial coefficients.

[0087] In some implementations, the ranges of average thickness and distance can be divided into multiple intervals, each corresponding to a linear correction function. For example, the thickness can be divided into three intervals: 0cm to 5cm, 5cm to 15cm, and 15cm to 30cm; the distance can be divided into three intervals: 0cm to 1cm, 1cm to 3cm, and greater than 3cm, for a total of nine intervals. Each interval is labeled with a set of linear coefficients. This scheme is faster than quadratic polynomials and can better fit regions with strong nonlinearity. Machine learning models can also be used as physical mapping correction functions. For example, models such as neural networks and support vector regression can be used, with average thickness, distance, and initial scattering signal as input and real scattering signal as output for training. After training, the model is used to correct the scattering signal of online scanning. This scheme can fit more complex nonlinear mapping relationships, further improving correction accuracy, and is suitable for radiotherapy scenarios with extremely high imaging quality requirements. The specific form of the physical mapping correction function is not limited in this disclosure. The average thickness, distance, and initial scattering signal corresponding to each shadow region are substituted into the physical mapping correction function to calculate the corrected scattering signal. Two-dimensional interpolation is then performed on the corrected scattering signals of all shadow regions to obtain a target BSA scattering image with the same size as the original projected image. The interpolation method can be cubic spline interpolation, bilinear interpolation, etc., consistent with the interpolation method used in generating the initial BSA scattering image. A pre-calibrated polynomial function accurately describes the mapping relationship between the initial scattering measurement value and the true scattering value, taking into account both the penumbra error caused by thickness and the boundary error caused by distance.

[0088] This disclosure corrects the initial scattering signal of each shadow region by introducing average thickness and boundary distance, which can effectively eliminate measurement errors caused by penumbra and boundary effects. Different correction coefficients are used for shadow regions with different thicknesses and boundary distances, so that the corrected scattering signal is closer to the true scattering value.

[0089] In some feasible approaches, the specific process for determining the optimized scattering kernel parameter set is as follows: First, a two-dimensional weighted image is constructed based on the first and second regions. This image distinguishes the contribution of different regions during the optimization process, with each pixel's weight value representing the weight coefficient of the scattering data at that location in the objective function. The weights in the first region are set to 1, and those in the second region are set to 0. This configuration ensures that only the scattering data from the first region participates in the optimization iteration process, preventing the low-precision data from the second region from affecting the optimization. This guarantees that the optimization of the scattering kernel parameter set is entirely guided by the data with the highest accuracy measured by the BSA method, avoiding interference from low-precision data in the optimization results.

[0090] Then, the original projected image is pre-corrected using the target BSA scattering image to obtain the pre-corrected projected image. Specifically, the pixel value of the corresponding position in the target BSA scattering image is subtracted from each pixel value in the original projected image to obtain the pre-corrected projected image. Since most of the scattered signal has been removed from the pre-corrected projected image, it is closer to the real ray signal. The basic assumption of the SKS method is that the scattered signal is proportional to the ray intensity. Using the pre-corrected projected image, which is closer to the real ray, as the convolution input can significantly improve the accuracy of the scattering kernel superposition calculation and reduce the calculation error caused by the large number of scattered components in the input signal.

[0091] Next, a target optimization function is constructed to measure the degree of difference between the intermediate model scattering image and the target BSA scattering image. The intermediate model scattering image is obtained by convolving the pre-corrected projection image with the kernel function corresponding to the scattering kernel parameter set of the current iteration. The kernel function usually adopts the form of a double Gaussian function, and its shape and amplitude are completely defined by the scattering kernel parameter set of the current iteration.

[0092] The objective optimization function includes a mean squared error term and a total variation regularization term:

[0093] in, and These are the width and height of the original projected image, respectively. For a two-dimensional weighted image in coordinates The pixel value at that location is 1 in the first region and 0 in the second region. For the target BSA scattering image in coordinates The pixel value at that location is used as a reference ground value for optimization. For the scattering kernel parameter set based on the current iteration The calculated intermediate model scattering image in coordinates Pixel value at that location, The total number of pixels in the first region is used to normalize the error so that the error value is independent of the image size. ,in, and These represent the gradients of the intermediate model's scattering image in the x and y directions, respectively. is the regularization penalty coefficient, a pre-defined hyperparameter used to balance the weights of the data fidelity term and the regularization term. The expression before the plus sign is the mean squared error term, used to measure the difference in pixel values ​​between the two images in the first region. The expression after the plus sign is the total variation regularization term, used to ensure that the generated scattering image has smooth characteristics. The smaller the total variation value corresponding to the total variation regularization term, the smoother the image, avoiding noise and artifacts.

[0094] After constructing the target optimization function, the initial scattering kernel parameter set is used as the initial iteration value. The kernel function corresponding to the pre-corrected projection image and the current iteration scattering kernel parameter set is convolved to obtain the intermediate model scattering image. Under the constraint of the two-dimensional weight image, the initial scattering kernel parameter set is iteratively updated by minimizing the target optimization function until the difference between the intermediate model scattering image and the target beam blocking array scattering image meets the convergence condition, and the optimized scattering kernel parameter set is obtained.

[0095] The iterative process is as follows: using the initial scattering kernel parameter set as the input parameter for the first round of iteration, each round of iteration performs the following operations in sequence: The first step is to calculate the intermediate model scattering image for the current iteration. A two-dimensional convolution is performed between the pre-corrected projection image and the kernel function corresponding to the scattering kernel parameter set for the current iteration to obtain the intermediate model scattering image for that iteration.

[0096] The second step is to calculate the objective optimization function value for the current iteration. Based on the difference between the intermediate model scattering image and the target BSA scattering image in this iteration, and combined with the constraints of the two-dimensional weight image, the objective optimization function value for the current iteration is calculated.

[0097] The third step is to update the scattering kernel parameter set. Numerical optimization algorithms such as gradient descent and L-BFGS are used to adjust the scattering kernel parameter set for the current iteration based on the gradient direction of the objective function. The adjusted parameter set is then used as the input parameters for the next iteration. Each parameter adjustment causes the objective function value to decrease.

[0098] The fourth step is convergence assessment. After each iteration, it is determined whether the preset convergence condition is met. If not, the process returns to step one and continues to the next iteration; if the condition is met, the iteration process terminates, and the scattering kernel parameter set of the current iteration is used as the final optimized scattering kernel parameter set. The convergence condition can be that the difference between the intermediate model scattering image and the target BSA scattering image is less than a difference threshold, or that the rate of change of the difference between the intermediate model scattering image and the target BSA scattering image is less than a change threshold over a preset number of iterations.

[0099] Through iterative optimization, the originally fixed initial set of scattering kernel parameters was adjusted to a set of parameters that accurately reflects the actual scattering characteristics of the current object being detected. The optimization process is guided by high-precision measurement data of the current object being detected, thus it can adapt to objects of different body sizes, positions, and tissue compositions, fundamentally eliminating the systematic bias caused by fixed parameters in the scattering kernel superposition method.

[0100] After obtaining the optimized scattering kernel parameter set, the pre-corrected projection image can be convolved using the optimized scattering kernel parameter set to obtain the SKS scattering image. In some implementations, the pre-corrected projection image can be convolved with the kernel function corresponding to the optimized scattering kernel parameter set in two dimensions to obtain an SKS scattering image with the same size as the original projection image.

[0101] The final generated SKS scattering image in the first region shows that, due to the optimization of the scattering kernel parameter set, the scattering estimation result is highly consistent with the target BSA scattering image, retaining the high accuracy advantage of the BSA method. In the second region, SKS obtains a smooth scattering distribution through continuous convolution calculation, eliminating the penumbra error and sparse sampling error of the BSA method, effectively compensating for the shortcomings of the beam blocking array method.

[0102] In some embodiments, the specific implementation of scattering correction of the original projection image based on the SKS scattering image includes: generating a target scattering image based on the target BSA scattering image and the SKS scattering image, and using the target scattering image to perform scattering correction on the original projection image.

[0103] Specifically, the pixel values ​​of the first region in the target BSA scattering image can be used as the pixel values ​​of the corresponding region in the target scattering image, and the pixel values ​​of the second region in the SKS scattering image can be used as the pixel values ​​of the corresponding region in the target scattering image to generate a complete target scattering image. Then, the pixel values ​​corresponding to the target scattering image are subtracted pixel by pixel from the original projection image to obtain the corrected ray projection image.

[0104] In the first region, the BSA method can directly measure the scattered signal, and the measurement results are not limited by the assumptions of the scattering kernel model. Correction using the target BSA scattering image can retain the high accuracy advantage of the BSA method in regions with high total X-ray attenuation. In the second region, the optimized scattering kernel parameter set can accurately reflect the actual scattering characteristics of the current object being detected. The SKS scattering image calculated based on the optimized scattering kernel parameter set does not have the penumbra error and sparse sampling error of the BSA method. It can obtain continuous and accurate scattering distributions in regions with low total X-ray attenuation and tissue boundary regions. Correction using the SKS scattering image can effectively avoid the accuracy defects of the BSA method in low attenuation regions.

[0105] It is understandable that a target scattering image is generated based on the target BSA scattering image and the SKS scattering image. Specifically, it could also involve stitching together a first region from the target BSA scattering image and a second region from the SKS scattering image to generate the target scattering image.

[0106] The stitching process is performed row-by-row, position-by-position (Z-axis). The Z-position corresponds to the detector's row index and to the same Z-coordinate plane in the scan space. In cone-beam CT systems, X-rays received by detectors in the same row all originate from the same Z-height plane. The scattered signal at the same Z-position changes relatively smoothly, ensuring accurate boundary intensity matching. The stitching process is as follows: Starting from the first Z position, the corresponding rows of data in the target BSA scattering image and SKS scattering image are stitched together sequentially.

[0107] Based on the pre-defined first and second regions, determine the coordinates of the boundary point between the first and second regions in the current row. In some implementations, the boundary point can be configured as the coordinates of the last pixel in the row belonging to the first region, or the coordinates of the first pixel belonging to the second region.

[0108] The intensity scaling factor is calculated based on the scattering values ​​in the target BSA scattering image at the boundary of the first region. In some implementations, the intensity scaling factor can be determined by taking the pixel values ​​of the target BSA scattering image at the boundary point and the pixel values ​​of the SKS scattering image at the boundary point, and determining the ratio between the two.

[0109] The pixel values ​​of all SKS scattering images belonging to the second region in the current row are multiplied by the calculated intensity scaling factor to complete intensity matching. After scaling, the scattering intensity at the boundary of the second region is consistent with the scattering intensity at the boundary of the first region, eliminating the intensity step at the boundary. For pixels in the first region in the current row, the pixel values ​​of the target BSA scattering image are used; for pixels in the second region, the pixel values ​​of the intensity-scaled SKS scattering image are used.

[0110] like Figure 6 The diagram shows a Z-position stitching of the BSA and SKS scattering images. The X-axis represents the detector column direction, i.e., the second direction, corresponding to the horizontal direction (left-right direction) in the scan space, and indicates the column index of the two-dimensional detector array. As can be seen in the diagram, the dashed lines in the X-direction divide the region into the first and second regions. The Z-axis represents the detector row direction, indicating the row index of the two-dimensional detector array, corresponding to the Z-coordinate plane (head-to-foot direction) in the scan space. The sparse dotted area in the diagram represents the first region, and the dense dotted area represents the second region. The solid curve in the first region represents the intensity distribution curve of the target BSA scattering image, the dashed curve in the second region represents the intensity distribution curve of the SKS scattering image, and the dotted-dash line represents the intensity distribution curve of the scaled SKS scattering image. The final target scattering image shows a continuous intensity distribution at the boundaries.

[0111] In other embodiments, a transition band with a width of N pixels can be set on both sides of the boundary point (the value of N can be from 15 to 20, and the target BSA scattering image and the intensity-scaled SKS scattering image are fused in a linear weighted manner).

[0112] For pixels in the current row located in the first region but outside the transition zone, the pixel values ​​of the target BSA scattering image are used; for pixels located in the second region but outside the transition zone, the pixel values ​​of the scattering image are superimposed using an intensity-scaled scattering kernel; pixels within the transition zone use weighted fused pixel values. After processing all Z positions, the target scattering image is stitched together.

[0113] Combination Figure 6 ,like Figure 7 As shown, in Figure 7 The blank area between the first and second regions is the transition zone. Weighted fusion is performed within the transition zone, and the resulting target scattering image is connected and smoothly transitioned at the stitching point.

[0114] The final generated target scattering image has the advantages of high accuracy of target BSA scattering image in the high total ray attenuation region and continuous estimation advantage of SKS scattering image in the low total ray attenuation region, and there is no boundary intensity step problem.

[0115] Understandably, after obtaining the target scattering image, the difference between the original projection image and the target scattering image is used to determine the ray projection image. Then, based on the ray projection image, an attenuation line integral image is determined. Finally, interpolation is used to complete the data in the attenuation line integral image corresponding to the shaded areas of the original projection image. The completed attenuation line integral image reflects the total attenuation of X-rays after passing through the detected object and is the direct input for 3D reconstruction.

[0116] Specifically, the ray projection image can be obtained by subtracting the corresponding pixel value of the target scattering image from each pixel value of the original projection image. Then, the attenuation line integral image is calculated based on the Beer-Lambert law, using the following formula: ,in, Indicates the empty field image in coordinates Pixel value at that location, Indicates the ray projection image in coordinates The pixel values ​​at each location are then analyzed. The scattering signal in the original projected image is completely removed, resulting in an attenuation line integral image containing only ray information, providing accurate input data for subsequent 3D reconstruction.

[0117] Each blocking element in the BSA creates a circular shadow region on the detector, which blocks the ray signal, resulting in data loss in the attenuation line integral image. In some feasible methods, grayscale gradual interpolation based on the structure tensor can be used to complete the shadow region in the attenuation line integral image corresponding to the original projected image. The specific implementation process is as follows: For each shaded region, a valid pixel neighborhood (e.g., 15×15 or 17×17) is determined, and the gray-level gradients of each pixel within the neighborhood in the X and Y directions are calculated. A two-dimensional covariance matrix is ​​constructed, and eigenvalue decomposition is performed on the covariance matrix to obtain the eigenvectors of the two principal directions. , and corresponding eigenvalues , . The maximum eigenvalue represents the magnitude of change in the direction of the most drastic grayscale change. correspond The feature vector points in the direction of the most dramatic grayscale change, that is, the direction perpendicular to the tissue edge; The minimum eigenvalue represents the magnitude of change in the direction where the grayscale change is most gradual. correspond The feature vector points in the direction where the grayscale change is the most gradual, that is, along the edge of the tissue.

[0118] For each interpolation point in the shaded region, calculate the direction vector from that point to each valid pixel in its neighborhood. Based on the angle between the direction vector and the two principal directions, calculate the weight of each neighboring pixel. Pixels along directions with gentle grayscale changes are assigned higher weights, while pixels along directions with drastic grayscale changes are assigned lower weights. The pixel value of the interpolation point is obtained by weighted averaging of the neighboring pixels.

[0119] Specifically, for the interpolation point within the shaded area, the direction vector is... A certain pixel in the neighborhood The weights are: ,in, , , This indicates the distance from the interpolation point (center of the shaded area) to the neighboring pixels. directional vector, This represents the projection length of the direction vector along the direction of drastic grayscale change. This represents the projection length of the direction vector along the direction where the grayscale change is gradual. When... When the eigenvector tends to be parallel to the eigenvalue corresponding to the larger eigenvalue, the penalty term in the denominator of the formula will increase sharply, thus making the interpolation weight assigned to the neighboring pixels smaller.

[0120] The final grayscale value of the interpolation point in the shadow area is obtained by weighted averaging of the grayscale values ​​of all valid reference pixels. ,in, For neighboring pixels The attenuation line integral value is used, and the denominator is the weight normalization factor to ensure that the gray range of the interpolation result is consistent with that of the original image.

[0121] like Figure 8 As shown, this is a schematic diagram of the weight allocation for the shaded area. The central circle of the shaded area to be filled in has no valid data for any of its pixels, with the direction vector as... The origin is defined by circles of varying densities surrounding it, representing pixels with different weights within their neighborhood. Higher density pixels have higher weights. High-weight pixels are located in the direction with the gentlest grayscale change (along the tissue edge). These pixels' grayscale values ​​are closest to the true grayscale values ​​of the shadow area, thus contributing the most to the interpolation calculation. Medium-weight pixels are located between the two main directions; their grayscale values ​​have some reference value. Low-weight pixels are located in the direction with the most drastic grayscale change (perpendicular to the tissue edge). These pixels' grayscale values ​​differ significantly from the true grayscale values ​​of the shadow area, therefore contributing very little to the interpolation calculation. The long arrow in the diagram points to the direction with the gentlest grayscale change. Direction: the longer the arrow, the slower the weight decay in that direction; short arrows point in the direction of the most drastic grayscale change. The shorter the arrow, the faster the weight decays in that direction.

[0122] Traditional interpolation methods (such as bilinear interpolation and cubic spline interpolation) do not consider the local structure of the image, and are prone to blurring in areas with drastic gray-level changes, such as tissue boundaries. Gray-level gradual interpolation based on structure tensors can adaptively interpolate along the direction of gradual gray-level changes, effectively preserving the edge information of the image, reducing interpolation artifacts, and improving the accuracy of data in shadow areas.

[0123] While structural tensor interpolation can adapt well to image texture, in certain high-frequency areas of fine bone or contrast-enhanced blood vessels, any two-dimensional image interpolation algorithm struggles to recover physically occluded three-dimensional perspective details. Therefore, in some implementations, forward projection can be used to complete the shadow areas.

[0124] The original projection image includes projection images at multiple projection angles, all formed by rays projected from the BSA onto the object being detected. First, a reconstructed image of the object is acquired. This reconstructed image is generated by performing three-dimensional reconstruction on the attenuation line integral images completed at multiple projection angles.

[0125] For example, a faster bilinear interpolation method (or a gray-scale gradual interpolation method based on structure tensors, etc.) can be used to complete the attenuation line integral image corresponding to the projection image at each projection angle. The completed attenuation line integral image can then be used for 3D reconstruction to obtain the reconstructed image.

[0126] Then, the reconstructed image is forward-projected to generate forward projection data for the corresponding shadow area. The forward projection data is then used to complete the data in the shadow area of ​​the attenuation line integral image, which corresponds to the shadow area of ​​the original projected image, to obtain the final completed attenuation line integral image.

[0127] In some implementations, the Filtered Back Projection (FDK) algorithm can be used to perform 3D reconstruction on the attenuation line integral image completed at multiple projection angles to obtain the 3D volume data of the object being detected. Then, the 3D volume data is processed by slicing, resampling, and adjusting window width and window level to generate the 2D tomographic image required for clinical diagnosis.

[0128] By using the completed attenuation line integral image for reconstruction, scattering artifacts and shadow interpolation artifacts can be effectively eliminated, resulting in a reconstructed image with high contrast, accurate CT values, and clear edges.

[0129] Forward projection completion utilizes global structural information in the 3D reconstructed image, rather than relying solely on local neighborhood pixel values. The reconstructed image contains complete anatomical information of the detected object; therefore, the shadow region data obtained by forward projection can accurately reflect the true ray distribution, essentially preserving all high-frequency structural information and further reducing artifacts in the reconstructed image.

[0130] Furthermore, to reduce information loss caused by shadow interpolation, in some embodiments, when BSA interval occlusion occurs, the original projection image also includes a projection image of the shadowless region formed by rays that are not occluded by the BSA and are projected onto the detection object. The method further includes: For a target projection angle not obscured by the BSA, the target BSA scattering image at the target projection angle is obtained by interpolation using two adjacent BSA scattering images corresponding to the BSA-occluded projection angles. The optimized scattering kernel parameter set at the target projection angle is then obtained by interpolation using two adjacent optimized scattering kernel parameter sets corresponding to the BSA-occluded projection angles. Based on the optimized scattering kernel parameter set and the original projection image at the target projection angle, the SKS scattering image at the target projection angle is determined. Finally, scattering correction is performed on the projection image at the target projection angle based on the target BSA scattering image and the SKS scattering image at the target projection angle.

[0131] In specific interval scanning scenarios, the BSA can be configured to operate in a dynamically moving mode. For example, at fixed angles (e.g., every 5 degrees), the BSA can be inserted into the ray path for an occluded scan, while at other intermediate projection angles, it can be removed from the ray path for an unobstructed, transparent scan. This mechanism preserves complete ray information at most projection angles, reducing the burden of shadow completion.

[0132] However, unobstructed angles also face strong scattering interference, requiring calculation of their scattering distribution. For the target projection angle... The angles of two adjacent BSAs are respectively The corresponding BSA scattering images are as follows: , The target projection angle between the two angles The corresponding BSA scattering image can be obtained through the formula: Interpolation is obtained.

[0133] Simultaneously, the optimized scattering kernel parameter set obtained under these two adjacent occlusion angles is utilized. and Similarly, interpolation is performed according to the angular distance ratio to obtain the angle suitable for the target projection. transition kernel parameter set Subsequently, based on the target projection angle The transition kernel parameter set and the original projected image at the target projection angle are used for convolution calculation to determine the SKS scattering image at the target projection angle. .

[0134] It is conceivable that, in integrating the results derived through interpolation... and When these values ​​change non-linearly with the angle shift, their confidence levels can be determined based on the target projection angle and the two adjacent projection angles under BSA occlusion. Based on the assigned weights, the target BSA scattering image and the target SKS scattering image at the target projection angle are weighted and summed to determine the target scattering image at the target projection angle. The scattering image at the target projection angle is then used to perform scattering correction on the projection image at the target projection angle.

[0135] Specifically, this could involve introducing a dynamic weight allocation mechanism based on a Gaussian function. The central bisector angle of two occlusion angles is defined as... Target projection angle Assigned weights It can be represented as: ,in, The standard deviation of the Gaussian kernel in the angle domain is an adjustable parameter used to control... and The degree of contribution balance. The target scattering image at the final synthesized target projection angle. When using this formula: .

[0136] Through this weight allocation, the closer the target projection angle is to the occlusion angle (i.e., closer to the occlusion angle), the better. , When ), weight The smaller the value, the more the target scattering image depends on ; and the closer the target projection angle is to the center bisector angle At that time, weight The larger the value, the more the target scattering image depends on... By cross-integrating the temporal and spatial dimensions, the accuracy of scattering estimation can be maintained across all scanning angles while reducing the frequency of physical occlusion.

[0137] Figure 9 This is a structural block diagram of a scattering correction device disclosed herein, such as... Figure 9 As shown, it includes: The determining part 901 is configured to determine a first region and a target beam blocking array scattering image based on the original projection image, wherein the total ray attenuation in the first region is greater than or equal to an attenuation threshold, and the original projection image includes a projection image with a shadowed area formed after the rays projected onto the detection object are blocked by the beam blocking array; the optimizing part 902 is configured to optimize the initial scattering kernel parameter set using pixels corresponding to the first region in the target beam blocking array scattering image to obtain an optimized scattering kernel parameter set; the determining part 901 is also configured to determine a scattering kernel superimposed scattering image based on the optimized scattering kernel parameter set and the original projection image; the correcting part 903 is configured to perform scattering correction on the original projection image based on the scattering kernel superimposed scattering image.

[0138] In some embodiments, determining portion 901 is configured to determine a second region based on the original projected image, wherein the total ray attenuation of the second region is less than an attenuation threshold.

[0139] In some embodiments, the determining portion 901 is configured to determine a thickness image based on the original projected image; to determine the region in the thickness image whose thickness is greater than a thickness threshold as the first region; and to determine the region in the thickness image whose thickness is less than or equal to the thickness threshold as the second region.

[0140] In some embodiments, the determining portion 901 is configured to acquire an empty field image after the ray is blocked by a beam blocking array; interpolate the initial scattering signals of each shadow region in the original projection image to obtain an initial beam blocking array scattering image; homogenize the initial beam blocking array scattering image to obtain a uniform scattering image; and determine a thickness image based on the original projection image, the empty field image, and the uniform scattering image.

[0141] In some embodiments, determining portion 901 is configured to determine a target beam blocking array scattering image based on the original projection image and the thickness image.

[0142] In some embodiments, the determining portion 901 is configured to, for each shadow region in the original projected image, determine the average thickness of each shadow region in the neighborhood in the thickness image; determine the distance of each shadow region from the boundary of the detected object; and, based on the average thickness value, distance, and initial scattering signal corresponding to the shadow region, determine the corrected scattering signal corresponding to each shadow region through a physical mapping correction function to obtain a target beam blocking array scattering image.

[0143] In some embodiments, the apparatus further includes: an acquisition section and a generation section; the acquisition section is configured to, for each phantom of different thicknesses, acquire a calibration beam blocking array measurement signal corresponding to an original calibration image formed by rays being blocked by the beam blocking array and projected onto each phantom in a horizontal direction at different positions of the beam blocking array relative to each phantom; acquire a conical beam projection image with the beam beam in a fully open state, and a fan-shaped beam projection image with the beam beam in a head-to-feet direction minimized state; a determination section 901 is configured to determine the true scattering signal of the calibration beam blocking array based on the difference between the conical beam projection image and the fan-shaped beam projection image; and the generation section is configured to fit the calibration beam blocking array measurement signal and the corresponding true scattering signal of the calibration beam blocking array to generate a physical mapping correction function.

[0144] In some embodiments, the optimization section 902 is configured to construct a two-dimensional weighted image based on a first region and a second region, wherein the weights in the first region are configured as 1 and the weights in the second region are configured as 0; pre-correct the original projection image using the scattering image of the target beam blocking array to obtain a pre-corrected projection image; use an initial scattering kernel parameter set as the initial iteration value to perform convolution calculation on the kernel function corresponding to the pre-corrected projection image and the scattering kernel parameter set of the current iteration to obtain an intermediate model scattering image; construct a target optimization function; and under the constraint of the two-dimensional weighted image, iteratively update the initial scattering kernel parameter set by minimizing the target optimization function until the difference between the intermediate model scattering image and the scattering image of the target beam blocking array satisfies the convergence condition to obtain an optimized scattering kernel parameter set; the determination section 901 is configured to perform a convolution operation on the pre-corrected projection image using the optimized scattering kernel parameter set to obtain a scattering kernel superimposed scattering image.

[0145] In some embodiments, the generating part is configured to stitch together the target beam blocking array scattering image and the scattering kernel superimposed scattering image to generate a target scattering image.

[0146] In some embodiments, the correction section 903 is configured to determine the difference between the original projected image and the target scattering image as a ray projection image; determine an attenuation line integral image based on the ray projection image; and complete the data in the attenuation line integral image corresponding to the shadow area of ​​the original projected image by interpolation.

[0147] In some embodiments, the correction section 903 is configured to use grayscale gradual interpolation based on the structure tensor to interpolate and complete the shadow region in the attenuation line integral image corresponding to the original projected image.

[0148] In some embodiments, the correction section 903 is configured to acquire a reconstructed image of the detected object, the original projection image including projection images at multiple projection angles, the projection images corresponding to the multiple projection angles being formed by blocking the rays projected onto the detected object by a beam blocking array, the reconstructed image being an image generated by three-dimensional reconstruction of the attenuation line integral image after completion at multiple projection angles; forward projection of the reconstructed image to generate forward projection data corresponding to the shadow area; and using the forward projection data to complete the data in the attenuation line integral image corresponding to the shadow area of ​​the original projection image.

[0149] In some embodiments, when the beam blocking array is obstructed by the projection angle, the original projection image also includes a projection image of a shadowless region formed by rays that are not obstructed by the beam blocking array and are projected onto the detection object; the apparatus further includes: an interpolation section configured to, for a target projection angle not obstructed by the beam blocking array, interpolate a target beam blocking array scattering image at the target projection angle using two target beam blocking array scattering images corresponding to the projection angle under beam blocking array obstruction adjacent to the target projection angle; interpolate an optimized scattering kernel parameter set at the target projection angle using two optimized scattering kernel parameter sets corresponding to the projection angle under beam blocking array obstruction adjacent to the target projection angle; a determination section 901 configured to determine a scattering kernel superimposed scattering image at the target projection angle based on the optimized scattering kernel parameter set at the target projection angle and the original projection image at the target projection angle; and a correction section 903 configured to perform scattering correction on the projection image at the target projection angle based on the target beam blocking array scattering image at the target projection angle and the scattering kernel superimposed scattering image at the target projection angle.

[0150] In some embodiments, the correction section 903 is configured to determine the allocation weight based on the target projection angle and two projection angles adjacent to the target projection angle under beam blocking array occlusion; based on the allocation weight, perform a weighted summation of the target beam blocking array scattering image at the target projection angle and the scattering kernel superimposed scattering image at the target projection angle to determine the target scattering image at the target projection angle; and use the target scattering image at the target projection angle to perform scattering correction on the projection image at the target projection angle.

[0151] In this embodiment, each part can implement the scattering correction method provided in the above method embodiments and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0152] Please refer to Figure 10 This illustration shows a schematic diagram of the hardware structure of an electronic device provided in an exemplary embodiment of this disclosure. In some examples, the electronic device may be at least one of devices such as a smartphone, smartwatch, desktop computer, laptop, virtual reality terminal, augmented reality terminal, wireless terminal, and laptop computer. The electronic device has communication capabilities and can access wired or wireless networks. The term "electronic device" can refer to one of multiple terminals; those skilled in the art will understand that the number of such terminals may be more or less.

[0153] like Figure 10 The diagram illustrates a structural block diagram of an electronic device provided in an exemplary embodiment of this disclosure. In some examples, Figure 10 The electronic device shown can be implemented as Figure 1The electronic device 40 in the scattering correction system 1 shown. It is understood that the electronic device 40 undertakes the calculation and processing work of the technical solution of this disclosure, and this disclosure does not limit it.

[0154] Optionally, the processor 1010 connects to various parts within the electronic device using various interfaces and lines. It executes various functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 1020, and by calling data stored in the memory 1020. Optionally, the processor 1010 can be implemented using at least one hardware form of Digital Signal Processing (DSP), Field Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 1010 can integrate one or more of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), Neural-network Processing Unit (NPU), and baseband chip. Specifically, the CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content displayed on the touchscreen; the NPU implements Artificial Intelligence (AI) functions; and the baseband chip handles wireless communication. It is understandable that the aforementioned baseband chip may not be integrated into the processor 1010, but may be implemented using a separate chip.

[0155] The memory 1020 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory 1020 may include a non-transitory computer-readable storage medium. The memory 1020 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 1020 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the various method embodiments described above, etc.; the data storage area may store data created based on the use of the electronic device, etc.

[0156] In addition, those skilled in the art will understand that the structure of the electronic device shown in the above figures does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements. For example, the electronic device may also include a display screen, camera assembly, microphone, speaker, radio frequency circuit, input unit, sensors (such as accelerometer, angular velocity sensor, light sensor, etc.), audio circuit, WiFi module, power supply, Bluetooth module, etc., which will not be described in detail here.

[0157] This disclosure also provides a computer-readable storage medium storing at least one instruction that is executed by a processor to implement the scattering correction method as described in the various embodiments above.

[0158] This disclosure also provides a computer program product including computer instructions stored in a computer-readable storage medium; a processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions to cause the electronic device to perform the scattering correction method described in the above embodiments.

[0159] This disclosure also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described scattering correction method embodiments and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0160] It should be understood that the chip mentioned in the embodiments of this disclosure may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0161] In the several embodiments provided in this disclosure, it should be understood that the disclosed systems, apparatuses, servers, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0162] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0163] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0164] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or all or part 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 disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0165] Those skilled in the art will recognize that the functions described in this disclosure in one or more of the examples above can be implemented using hardware, software, firmware, or any combination thereof. When implemented in software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transfer of a computer program from one place to another. Storage media can be any available medium accessible to a general-purpose or special-purpose computer.

[0166] It should be noted that the technical solutions described in this disclosure can be combined arbitrarily as long as they do not conflict.

[0167] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A scattering correction method, characterized in that, The method includes: Based on the original projection image, a first region and a target beam blocking array scattering image are determined, wherein the total ray attenuation in the first region is greater than or equal to an attenuation threshold, and the original projection image includes a projection image with a shadowed area formed after the beam blocking array blocks the rays projected toward the detection object. Using the pixels corresponding to the first region in the scattering image of the target beam blocking array, the initial scattering kernel parameter set is optimized to obtain an optimized scattering kernel parameter set; Based on the optimized scattering kernel parameter set and the original projection image, a superimposed scattering image of the scattering kernels is determined; and The original projection image is then corrected for scattering based on the superimposed scattering image of the scattering kernel.

2. The scattering correction method according to claim 1, characterized in that, The method further includes: Based on the original projection image, a second region is determined, wherein the total ray attenuation in the second region is less than the attenuation threshold.

3. The scattering correction method according to claim 2, characterized in that, Determining the first region based on the original projected image includes: Based on the original projected image, determine the thickness image; The region in the thickness image whose thickness is greater than a thickness threshold is defined as the first region; Accordingly, determining the second region based on the original projected image includes: The region in the thickness image whose thickness is less than or equal to the thickness threshold is defined as the second region.

4. The scattering correction method according to claim 3, characterized in that, Determining the thickness image based on the original projected image includes: Acquire an image of the empty field after the ray is blocked by a beam blocking array; Interpolate the initial scattering signals of each shadow region in the original projection image to obtain the initial beam blocking array scattering image; The initial beam blocking array scattering image is homogenized to obtain a uniform scattering image; and The thickness image is determined based on the original projection image, the empty field image, and the uniform scattering image.

5. The scattering correction method according to claim 3, characterized in that, The process of determining the target beam blocking array scattering image based on the original projection image includes: Based on the original projection image and the thickness image, the scattering image of the target beam blocking array is determined.

6. The scattering correction method according to claim 5, characterized in that, Determining the scattering image of the target beam blocking array based on the original projection image and the thickness image includes: For each shadow region in the original projected image, the average thickness of each shadow region in the neighborhood is determined in the thickness image; Determine the distance of each shadow region from the boundary of the detected object; and Based on the average thickness value corresponding to the shadow region, the distance, and the initial scattering signal, the corrected scattering signal corresponding to each shadow region is determined by the physical mapping correction function, thereby obtaining the scattering image of the target beam blocking array.

7. The scattering correction method according to claim 6, characterized in that, The method further includes: For each phantom with different thicknesses, at different positions of the beam blocking array relative to each phantom in the horizontal direction, the calibration beam blocking array measurement signal corresponding to the original calibration image formed by the beam blocking array blocking the beam onto each phantom is acquired. Acquire a cone-shaped beam projection image with the snoopor fully open and a fan-shaped beam projection image with the snoopor minimized in the head-to-feet direction; Based on the difference between the cone-beam projection image and the fan-beam projection image, the true scattering signal of the calibration beam blocking array is determined; and The physical mapping correction function is generated by fitting the measured signal of the calibrated beam blocking array with the corresponding real scattering signal of the calibrated beam blocking array.

8. The scattering correction method according to claim 2, characterized in that, The step of optimizing the initial scattering kernel parameter set using pixels corresponding to the first region in the scattering image of the target beam blocking array to obtain an optimized scattering kernel parameter set includes: A two-dimensional weighted image is constructed based on the first region and the second region, wherein the weight in the first region is configured as 1 and the weight in the second region is configured as 0. The original projection image is pre-corrected using the scattering image of the target beam blocking array to obtain a pre-corrected projection image; Construct the objective optimization function; Using the initial scattering kernel parameter set as the initial iteration value, convolve the pre-corrected projection image with the kernel function corresponding to the current iteration's scattering kernel parameter set to obtain an intermediate model scattering image; and Under the constraint of the two-dimensional weighted image, the initial scattering kernel parameter set is iteratively updated by minimizing the objective optimization function until the difference between the intermediate model scattering image and the target beam blocking array scattering image satisfies the convergence condition, thus obtaining the optimized scattering kernel parameter set. Accordingly, determining the superimposed scattering image based on the optimized scattering kernel parameter set and the original projection image includes: The pre-corrected projection image is convolved using the optimized scattering kernel parameter set to obtain the scattering kernel superimposed scattering image.

9. The scattering correction method according to claim 2, characterized in that, The step of performing scattering correction on the original projection image based on the superimposed scattering image of the scattering kernel includes: A target scattering image is generated based on the scattering image of the target beam blocking array and the superimposed scattering image of the scattering kernel; and The original projected image is corrected for scattering using the target scattering image.

10. The scattering correction method according to claim 9, characterized in that, The process of generating a target scattering image based on the scattering image of the target beam blocking array and the superimposed scattering image of the scattering kernel includes: The first region in the scattering image of the target beam blocking array and the second region in the scattering kernel superimposed scattering image are stitched together to generate a target scattering image.

11. The scattering correction method according to claim 9, characterized in that, The step of using the target scattering image to perform scattering correction on the original projected image includes: The difference between the original projected image and the target scattering image is determined as the ray projection image; Based on the ray projection image, determine the attenuation line integral image; and Interpolation is used to complete the data in the attenuation line integral image corresponding to the shadow area of ​​the original projected image.

12. The scattering correction method according to claim 11, characterized in that, The step of interpolating to complete the data in the attenuation line integral image corresponding to the shadow area of ​​the original projected image includes: The grayscale gradual interpolation based on the structure tensor is used to interpolate and complete the shadow region in the attenuation line integral image that corresponds to the original projection image.

13. The scattering correction method according to claim 11, characterized in that, The step of interpolating to complete the data in the attenuation line integral image corresponding to the shadow area of ​​the original projected image includes: A reconstructed image of the detected object is obtained. The original projection image includes projection images at multiple projection angles. The projection images corresponding to the multiple projection angles are formed by blocking the rays projected onto the detected object by the beam blocking array. The reconstructed image is an image generated by three-dimensional reconstruction of the attenuation line integral image after completion at the multiple projection angles. The reconstructed image is forward-projected to generate forward projection data corresponding to the shadow region; and Using the positive projection data, complete the data in the attenuation line integral image corresponding to the shadow area of ​​the original projection image.

14. The scattering correction method according to claim 1, characterized in that, When the beam blocking array's projection angle is obstructed, the original projection image also includes a projection image of a shadowless region formed by rays projected onto the detected object that are not obstructed by the beam blocking array. The method further includes: For the target projection angle that is not blocked by the beam blocking array, the target beam blocking array scattering image under the target projection angle is obtained by interpolation using two target beam blocking array scattering images corresponding to the projection angle under beam blocking array occlusion, which are adjacent to the target projection angle. Using the optimized scattering kernel parameter sets corresponding to the projection angles under beam blocking arrays adjacent to the target projection angle, interpolation is performed to obtain the optimized scattering kernel parameter set at the target projection angle; and Based on the optimized scattering kernel parameter set under the target projection angle and the original projection image under the target projection angle, a superimposed scattering image of scattering kernels under the target projection angle is determined. Accordingly, the scattering correction of the original projection image based on the superimposed scattering image of the scattering kernel includes: Based on the target beam blocking array scattering image at the target projection angle and the scattering kernel superimposed scattering image at the target projection angle, scattering correction is performed on the projection image at the target projection angle.

15. The scattering correction method according to claim 14, characterized in that, The scattering correction of the projection image at the target projection angle based on the target beam blocking array scattering image at the target projection angle and the superimposed scattering image of the scattering kernel at the target projection angle includes: The weighting is determined based on the target projection angle and the two projection angles adjacent to the target projection angle under beam blocking array occlusion. Based on the assigned weights, a weighted sum is performed on the target beam blocking array scattering image and the scattering kernel superimposed scattering image at the target projection angle to determine the target scattering image at the target projection angle; and... The scattering image at the target projection angle is used to perform scattering correction on the projection image at the target projection angle.

16. A scattering correction system, characterized in that, include: The imaging source is configured to emit a beam of light toward the object being detected; A detector, positioned opposite the imaging source, is configured to receive a beam of light passing through the object being detected and to generate a raw projection image; A beam blocking array is disposed on the beam path between the imaging source and the detector and is configured to block a portion of the beams directed toward the object being detected. as well as An electronic device, communicatively connected to the detector, to acquire the original projected image from the detector and to perform the scattering correction method according to any one of claims 1 to 15.

17. The scattering correction system according to claim 16, characterized in that, The beam blocking array is movably disposed on the beam path and switches between a working position that cuts into the beam and a non-working position that cuts out of the beam.

Images