Scattering correction method and electronic equipment

Abstract

The present application provides a scatter correction method and an electronic device. In this embodiment, a scattering physical model is constructed with the help of a calibration body to determine an optimal first scatter correction coefficient and an optimal second scatter correction coefficient. Subsequently, the scattering components in the image can be corrected with the help of the optimal first scatter correction coefficient and the optimal second scatter correction coefficient, thereby achieving scatter correction for the image generated by the X-ray imaging device, effectively increasing the contrast of the image, improving the signal-to-noise ratio, etc., and facilitating more accurate material identification based on the image.

Description

Technical Field

[0001] The present application relates to image processing technology, and in particular to a scatter correction method and electronic equipment. Background Art

[0002] When X-rays pass through matter, they interact with the electrons inside the matter, and this interaction changes the propagation direction of the X-rays, which is considered to be scattered. The scattering of X-rays will reduce the quality of the images generated by the X-ray imaging device, such as reduced image contrast, reduced spatial resolution, and reduced signal-to-noise ratio, which will in turn affect subsequent image processing based on the images generated by the X-ray imaging device, such as target recognition. Summary of the invention

[0003] The present application provides a scatter correction method and electronic equipment to implement scatter correction for images generated by an X-ray imaging device.

[0004] The present application provides a scatter correction method, which includes:

[0005] Generate a scattering physical model for describing scattering; the scattering physical model indicates that: the pixel feature value P_total of a pixel point at any pixel position in a training image corresponding to any calibration body is related to the contribution value P_direct of a direct X-ray directly irradiated to the pixel position, the contribution value P_scatter of any scattered X-ray scattered to the pixel position, and the pixel feature value distribution P of the scattered X-ray; the P_scatter is related to a first scattering correction coefficient to be optimized, and the P is related to a second scattering correction coefficient to be optimized; wherein, under the premise that the P_total at any pixel position is determined, the scattering physical model represents the relationship between the P_direct at the pixel position and the first scattering correction coefficient and the second scattering correction coefficient;

[0006] On the premise of determining the P_total at each different pixel position in each region of interest in the training image, when the first scattering correction coefficient and the second scattering correction coefficient are set to different values, the P_direct at each different pixel position in each region of interest is determined based on the scattering physical model; based on the set first loss function about the P_direct at each pixel position in each region of interest, the optimal first scattering correction coefficient and the optimal second scattering correction coefficient corresponding to the calibration body are determined; the signal strength at any two different pixel positions in the same region of interest meets the set similarity requirement;

[0007] The optimal first scatter correction coefficient and the optimal second scatter correction coefficient are used to perform scatter correction on any image.

[0008] The present application also provides a scatter correction method, which includes:

[0009] Obtain a target image to be corrected; the target image is an image obtained by performing image acquisition on the target object;

[0010] Based on the grayscale value of each pixel position in the target image and the grayscale value function corresponding to each calibration body in the obtained scatter correction information, the weight coefficient corresponding to each calibration body is determined; the scatter correction information includes at least one optimal scatter coefficient corresponding to the calibration body, and the optimal scatter coefficient includes at least: an optimal first scatter correction coefficient and an optimal second scatter correction coefficient;

[0011] Determine the target atomic number of the target object based on the weight coefficient corresponding to each calibration body and the calibration atomic number of each calibration body; and check whether there is a calibration body that meets the conditions in the obtained scattering correction information, and the calibration atomic number of the calibration body that meets the conditions meets the first set similarity requirement with the target atomic number;

[0012] If there is a qualified calibration body in the obtained scatter correction information, scatter correction is performed on the target image based on the optimal scatter coefficient corresponding to the qualified calibration body; and / or

[0013] If there is no qualified calibration body in the obtained scatter correction information, based on the fact that a reference calibration body is found in the obtained scatter correction information, and the calibration atomic number of any reference calibration body meets the second set similarity requirement with the target atomic number; scatter correction is performed on the target image based on the optimal scatter coefficient corresponding to each reference calibration body.

[0014] The embodiment of the present application also provides an electronic device. The electronic device includes: a processor and a machine-readable storage medium;

[0015] The machine-readable storage medium stores machine-executable instructions that can be executed by the processor;

[0016] The processor is used to execute machine executable instructions to implement the steps of the above disclosed method.

[0017] It can be seen from the above technical solution that this embodiment constructs a scattering physical model with the help of a calibration body to determine the optimal first scattering correction coefficient and the optimal second scattering correction coefficient. Subsequently, the scattering components in the image can be corrected with the help of the optimal first scattering correction coefficient and the optimal second scattering correction coefficient, thereby realizing scattering correction of the image generated by the X-ray imaging device, effectively increasing the contrast of the image, improving the signal-to-noise ratio, etc., and facilitating more accurate material identification based on the image.

[0018] Furthermore, in this embodiment, a scattering physical model is constructed with the aid of a calibration body to determine an optimal first scattering correction coefficient and an optimal second scattering correction coefficient for subsequent scattering correction of an image, without the need for hardware or Monte Carlo simulation support. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0020] Figure 1 A flow chart of the method provided in the embodiment of the present application;

[0021] Figure 2 A scattering schematic diagram provided for an embodiment of the present application;

[0022] Figure 3 Another scattering schematic diagram provided for an embodiment of the present application;

[0023] Figure 4 Another method flow chart provided for an embodiment of the present application;

[0024] Figure 5 A structural diagram of a device provided in an embodiment of the present application;

[0025] Figure 6 Another device structure diagram provided for an embodiment of the present application;

[0026] Figure 7 A structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the technical solutions provided by the embodiments of the present application and to make the above-mentioned purposes, features and advantages of the embodiments of the present application more obvious and understandable, the technical solutions in the embodiments of the present application are further described in detail below in conjunction with the accompanying drawings.

[0028] See also Figure 1 , Figure 1 The method flow chart provided in the embodiment of the present application. The method can be applied to electronic devices. The electronic devices here are, for example, terminals, servers, etc., which are not specifically limited in this embodiment.

[0029] like Figure 1 As shown, the process may include the following steps:

[0030] Step 101: Generate a scattering physical model for describing scattering.

[0031] In this embodiment, the scattering physical model indicates that: the pixel feature value P_total of a pixel point at any pixel position in the training image corresponding to any calibration body is related to the contribution value P_direct of the direct X-ray directly irradiated to the pixel position, the contribution value P_scatter of any scattered X-ray scattered to the pixel position, and the pixel feature value distribution P of the scattered X-ray; the P_scatter is related to the first scattering correction coefficient to be optimized, and P is related to the second scattering correction coefficient to be optimized; wherein, under the premise that P_total at any pixel position is determined, the scattering physical model represents the relationship between P_direct at the pixel position and the first scattering correction coefficient and the second scattering correction coefficient.

[0032] It should be noted that the calibration body can be an object made of uniform material and easily processed into a stepped block. The calibration material is not limited to carbon, aluminum, iron, etc.

[0033] The following describes how to generate a scattering physics model:

[0034] Taking the above pixel characteristic value as signal intensity as an example, the gray value or other characteristic values ​​have similar principles and will not be described one by one.

[0035] Taking the above pixel characteristic value as signal strength as an example, the above P_total is I_total. Figure 2 As shown, for any training image corresponding to the calibration object, the signal intensity I_total of the pixel point at any pixel position in the training image is composed of the signal intensity I_direct contributed by the X-ray directly irradiated to the pixel position (referred to as direct X-ray) to the pixel position, and the signal intensity I_scatter contributed by each X-ray scattered to the pixel position (referred to as scattered X-ray) to the pixel position.

[0036] That is, the signal intensity I_total of a pixel at any pixel position in the training image is specifically: I_total = I_direct + I_scatter. Optionally, I_direct approximately obeys the Lambert-Beer law, for example .in, Indicates that direct X-rays are incident on Figure 2 The signal strength of the calibration body shown, represents the attenuation coefficient of direct X-rays, and t represents the attenuation coefficient of direct X-rays in Figure 2 The distance that the calibration body shown in the figure passes through (also called the distance that the direct X-ray is incident on Figure 2 The thickness of the calibration body shown.

[0037] from Figure 2As can be seen from the schematic diagram, the scattering signal intensity of a pixel point at any pixel position can be specifically the sum of the signal intensities contributed to the pixel position by all scattered X-rays scattered to the pixel position. In specific implementation, the signal intensities contributed to the pixel position by different scattered X-rays scattered to the pixel position can be approximately considered independent. Based on this principle, in the process of generating the scattering physical model, this embodiment needs to consider the incident X-rays such as Figure 2 The scattering components of X-rays at any position of the calibration body at different pixel positions are shown. Figure 3 An example is given of the pixel position range (referred to as the scattering area) to which X-rays incident on a position on the calibration object are scattered.

[0038] like Figure 3 As shown in FIG. 1 , any X-ray incident on the calibration body will contribute to the signal intensity of the pixel points at different pixel positions in the scattering area when it is scattered. In a specific implementation, the scattering contribution of the X-ray to a pixel position in the scattering area is the same as the scattering contribution of the X-ray incident on the calibration body. Figure 3 The thickness of the calibration body at the position (incident position) shown, the material of the calibration body at the incident position, and the distance between other pixel positions in the scattering area and the pixel position. As an example, the scattering contribution of X-rays to the pixel position can be expressed by a point spread function such as Indicates. represents the total signal intensity of X-ray scattering, is a point spread function, for example, represented by: the scattering signal intensity distribution of X-ray scattering (also called the normalized scattering function spatial distribution). Optionally, It may be the scattered signal intensity distribution of the X-ray in a reference coordinate system formed with the above pixel position as the center.

[0039] Optionally, the total signal intensity S of the X-ray scattering is related to the material of the X-ray at the incident position of the calibration body. , the pixel position to which the X-ray directly hits Signal strength I_total For example, S can be written as If the material of the calibration body is a single material, S can be written as , directly expressed as about I_total Function .

[0040] As an embodiment, S can be constructed by fitting. It should be noted that S is not limited to being implemented by fitting.

[0041] Taking the fitting method to construct S as an example, if the above S is fitted by a polynomial. For simple representation, if the material of the calibration body is a single material, the above S can be expressed as:

[0042] ;

[0043] Where i represents the order of the polynomial, is the first scatter correction coefficient to be optimized, Contains all Based on S It can be seen that if Known, then S It can be confirmed.

[0044] Optionally, the point spread function It can be represented by a two-dimensional function. For example, the point spread function pass express. , Indicates the pixel position where the X-ray directly hits The position information of any pixel position to which the X-ray is scattered relative to the above center in the reference coordinate system formed by the center. is the parameter to be optimized. In, if Known, then It can be confirmed.

[0045] In the specific implementation, It can be composed of multiple components, each of which is represented by a two-dimensional product of marginal distributions such as Gaussian functions and exponential functions. Taking the two-dimensional Gaussian distribution as an example, Specifically: .

[0046] In the above formula, is a 2*2 covariance matrix, ; Parameters to be optimized include The parameters to be optimized are is related to the construction form of. i represents the number of Gaussian function components, , As described above.

[0047] The point spread function It is applied to the case of linear translation invariance. Here, the linear translation invariance means that the calibration body is translated under the premise that the X-ray source and detector remain unchanged, and the change of the signal intensity at the same pixel position before and after the translation meets the set minimum change requirement.

[0048] The linear translation variable case is applied to the case where the calibration body is translated under the premise that the X-ray source and the detector remain unchanged. If the change in the signal intensity at the same pixel position before and after the translation does not meet the set minimum change requirement, such as a large change, then the above point spread function It can be further expressed as The four-dimensional function of .

[0049] As described above, ( , ) represents the pixel position directly hit by X-rays The position information of any pixel position scattered by the X-ray relative to the above center in the reference coordinate system formed by the center. Based on this, as an embodiment, the above parameters to be optimized are It can be further expressed as For example, It can be expressed as: .

[0050] In this embodiment, the above parameters to be optimized The parameters included in It can be constructed according to actual needs, for example, by two-dimensional polynomial fitting. For example, It can be expressed as: .in, The corresponding polynomial order. Others are similar.

[0051] It can be found that the parameters to be optimized Further expansion to include fitting coefficients (such as ) of the second scatter correction coefficient The second scatter correction factor here is is the scatter correction coefficient to be optimized.

[0052] As described above, in actual image processing, it is impossible to obtain the signal intensity of X-ray scattering to each pixel position, and the scattering component in the total signal intensity I_total of the pixel point at each pixel position is the superposition of the signal intensity of multiple X-ray scattering to the pixel position. Based on this, combined with the first scattering correction coefficient to be optimized and the second scatter correction factor , a scattering physics model can be generated.

[0053] As an embodiment, the scattering physical model can be expressed by the following formula:

[0054] .

[0055] In the above scattering physical model, I_total (x, y) can be measured. Under this premise, if the parameters are given , , then we can determine That is, in the scattering physics model, if I_total (x, y) is measured, I_direct (x, y) is determined by the given first scattering correction coefficient and the second scattering correction coefficient.

[0056] The above describes how to generate a scattering physical model for describing scattering by taking signal intensity as an example. Similar scattering physical models for describing scattering can also be obtained by using other pixel feature values ​​such as grayscale values, which will not be described in detail here.

[0057] Step 102, on the premise of determining P_total at different pixel positions in each region of interest in the training image, respectively set the first scatter correction coefficient and the second scatter correction coefficient to different values, and determine P_direct at different pixel positions in each region of interest based on the scattering physical model.

[0058] Optionally, before executing step 102, the first region of interest may be extracted from the training image. Optionally, this embodiment uses information such as grayscale and gradient to exclude areas such as air where no calibration body exists based on a reasonable threshold value to obtain the area where the calibration body is located (i.e., the first region of interest).

[0059] Optionally, before extracting the first region of interest from the training image, the training image may be first subjected to image correction. The image correction here is, for example, correcting the signal intensity I at each pixel position in the training image one by one: .

[0060] in, Represents dark field data, specifically the average signal intensity in the image collected by the detector when the ray source is not activated. This can be represented by a column of signal strength values. represents bright field data, which is obtained by averaging the signal intensities of at least the first few columns containing only air in the training image, such as the first 10 columns. This can be represented by a column of signal strength values. and are equal in size. Represents the signal intensity at the pixel position in the training image.

[0061] Taking the correction of the signal intensity at any pixel position in any column of the training image (pixel position on the i-th row) as an example, .in, represents the signal intensity of the pixel position on the i-th row in the column, Indicates that The signal strength on the i-th row in , express The signal strength on the i-th row in .

[0062] In this embodiment, after obtaining the first region of interest, regions with approximately equal grayscale in the connected domain (recorded as second regions of interest) can be output one by one from the first region of interest. For example, based on a clustering algorithm, information such as grayscale, standard deviation, and pixel position can be used to output all second regions of interest that meet the screening conditions.

[0063] In specific applications, the same first region of interest can obtain multiple second regions of interest, and the signal strength at any two different pixel positions in the same second region of interest meets the set similarity requirements (for example, the signal strength is the same, or the grayscale is approximately equal). If the signal strength at any two different pixel positions in the same second region of interest meets the set similarity requirements, it means that the same second region of interest is an image of an area on the calibration body whose thickness meets the set similarity requirements. Taking the calibration body as a step block as an example, a step block of a specific thickness corresponds to a second region of interest. Of course, if there is no second region of interest that meets the conditions in the first region of interest, it is deemed that there is no second region of interest. If the calibration body is a fluid, there may be no second region of interest.

[0064] Based on the above description, in this embodiment, the region of interest in step 102 may be the above-mentioned second region of interest. The scattering model described above can obtain the P_total of each pixel position in the second region of interest, such as the above-mentioned I_total, by measuring for each second region of interest. Afterwards, when the first scattering correction coefficient and the second scattering correction coefficient are set to different values, the P_direct at each different pixel position in each second region of interest is determined based on the above-mentioned scattering physical model. For example, for a pixel position in the second region of interest, when the first scattering correction coefficient is set to the first value and the second scattering correction coefficient is set to the second value, the I_total of the pixel position is substituted into the above-mentioned scattering physical model, and the P_direct at the pixel position is obtained based on the first value and the second value; and when the first scattering correction coefficient is set to the third value and the second scattering correction coefficient is set to the fourth value, the I_total of the pixel position is substituted into the above-mentioned scattering physical model, and the P_direct at the pixel position is obtained based on the third value and the fourth value. By analogy, when the first scattering correction coefficient and the second scattering correction coefficient are set to different values, if the P_total at different pixel positions in each second region of interest is measured, such as the above-mentioned I_total, then the above-mentioned scattering physical model is substituted, and finally the P_direct at different pixel positions in each region of interest can be determined.

[0065] Step 103 : Based on the set first loss function of P_direct at each pixel position in each region of interest, determine the optimal first scatter correction coefficient and the optimal second scatter correction coefficient corresponding to the calibration body.

[0066] The first loss function is a function of P_direct at each pixel position in each second region of interest.

[0067] Optionally, as described above, P_direct at any pixel position in any second region of interest is determined by substituting P_total at the pixel position, such as the above-mentioned I_total, into the above-mentioned scattering physical model, given the first scattering correction coefficient and the second scattering correction coefficient to be optimized, and based on the given first scattering correction coefficient and the second scattering correction coefficient.

[0068] For example, still taking the pixel feature value as signal intensity, combined with the above scattering physical model, for a pixel position (x, y) in the training image, the pixel position is the pixel position in the nth region of interest, then , The specific value of It should be noted that given , Different values ​​of can get different .

[0069] Given the first scatter correction coefficient and the second scatter correction coefficient, we can obtain After that, the above-mentioned first loss function can be substituted. The optimal first scatter correction coefficient and the optimal second scatter correction coefficient are determined by minimizing the first loss function. The minimizing principle of the first loss function is similar to the existing principle of minimizing the loss function, and this embodiment does not specifically limit it.

[0070] Still taking the pixel feature value as signal strength as an example, as an embodiment, the first loss function (denoted as ) can be expressed as:

[0071] ;

[0072] Here, n represents the index or number of the second region of interest, for example, n is 1, 2, etc. This embodiment does not specifically limit this. I_direct (specifically, the signal intensity contributed by direct X-rays to the pixel position (x, y)) at any pixel position (x, y) in the nth second region of interest; It indicates the average value of I_direct of all pixel positions in the nth second region of interest. It should be noted that the first loss function is just an example of the average value, not for limitation, and the representation of L1 is not for limitation, but only for a clearer understanding of the present solution.

[0073] Based on the above description, the optimal scattering coefficient will eventually be obtained. After that, the image can be scatter corrected based on the optimal scattering coefficient to improve image contrast, signal-to-noise ratio, spatial resolution, and material recognition capabilities. The following will give an example to describe how to perform scatter correction, which will not be repeated here.

[0074] So far, completed Figure 1 Description of the process shown.

[0075] pass Figure 1 As shown in the process, this embodiment constructs a scattering physical model with the help of a calibration body to determine the optimal first scattering correction coefficient and the optimal second scattering correction coefficient. Subsequently, the scattering components in the image can be corrected with the help of the optimal first scattering correction coefficient and the optimal second scattering correction coefficient, thereby realizing scattering correction of the image generated by the X-ray imaging device, effectively increasing the contrast of the image, improving the signal-to-noise ratio, etc., and facilitating more accurate material identification based on the image.

[0076] Furthermore, in this embodiment, a scattering physical model is constructed with the aid of a calibration body to determine an optimal first scattering correction coefficient and an optimal second scattering correction coefficient for subsequent scattering correction of an image, without the need for hardware or Monte Carlo simulation support.

[0077] Optionally, the scattering physics model further corresponds to a second loss function. The second loss function is used to determine the optimal mass attenuation coefficient.

[0078] Specifically, the second loss function may be a function of the quality of the calibration volume in the training image. Still taking the pixel feature value as the signal intensity as an example, based on the above description, the second loss function may be expressed as: .

[0079] in, is the quality of the calibration volume obtained based on the kth training image of the same calibration volume, It should be noted that the representation of the second loss function is only an example and is not intended to be limiting.

[0080] Optionally, the quality of the calibration body is obtained based on any training image of the calibration body by:

[0081] .

[0082] Where m represents the quality of the calibration body obtained based on a training image of the calibration body, Represents the signal intensity of any pixel position (x, y) in the nth second region of interest in the training image. represents the mass attenuation coefficient of the nth second region of interest. As an example, It can be expressed by the following formula: .

[0083] here, : represents the average signal intensity of all pixel positions in the nth second region of interest. Indicates the thickness of the calibration volume corresponding to the nth second region of interest (for details, see the description of the second region of interest).

[0084] By minimizing the second loss function, the optimal mass attenuation coefficient can be obtained.

[0085] In other words, the optimal mass attenuation coefficient can only be achieved by measuring different training images of the same calibration object.

[0086] As an embodiment, this embodiment will pre-set the ratio of the first loss function and the second loss function, for example, the first weight parameter corresponding to the first loss function, and the second weight parameter corresponding to the second loss function. Based on this, when determining the optimal first scatter correction coefficient and the optimal second scatter correction coefficient, the following steps can be performed: obtaining the first weight parameter corresponding to the first loss function; based on the first weight parameter, substituting P_direct, such as I_direct, at each pixel position in each region of interest into the first loss function, and determining the optimal first scatter correction coefficient and the optimal second scatter correction coefficient corresponding to the calibration body according to the principle of minimizing the value of the first loss function.

[0087] Similarly, when determining the optimal mass attenuation coefficient corresponding to the calibration object, the above may include: obtaining a second weight parameter corresponding to the second loss function; based on the second weight parameter and the second loss function and in combination with the quality of the calibration object obtained from different training images and the actual mass of the calibration object, determining the optimal mass attenuation coefficient corresponding to the calibration object.

[0088] The following describes how to perform scatter correction:

[0089] See also Figure 4 , Figure 4 Another method flow chart provided in the embodiment of the present application. Figure 4 As shown, the method may include the following steps:

[0090] Step 401, obtaining a target image to be corrected; the target image is an image obtained by capturing an image of a target object.

[0091] Step 402 : Based on the grayscale value of each pixel position in the target image and the grayscale value function corresponding to each calibration body in the obtained scatter correction information, the weight coefficient corresponding to each calibration body is determined.

[0092] In this embodiment, after obtaining the optimal scattering coefficient, such as the above-mentioned optimal first scattering correction coefficient and the optimal second scattering correction coefficient, or the optimal first scattering correction coefficient, the optimal second scattering correction coefficient, and the optimal mass attenuation coefficient, the material corresponding to the calibration body and the optimal scattering coefficient can be recorded in scattering correction information, such as a scattering correction table. By analogy, the scattering correction information, such as a scattering correction table, can record the materials corresponding to multiple different calibration bodies and the optimal scattering coefficients.

[0093] In this embodiment, step 402 can be implemented as follows:

[0094] Determine the weight coefficient corresponding to each calibration body according to the following formula:

[0095] .

[0096] in, Represents the gray value function obtained based on the gray value of each pixel position in the target image, represents the weight coefficient of the jth calibration volume in the obtained scatter correction information, Represents the gray value function of the j-th calibration volume.

[0097] In this embodiment, The gray value function of the jth calibration volume can be obtained by observing the target image. The calibration can be performed in advance by methods including but not limited to fitting, interpolation, etc. According to the above formula, the weight coefficient corresponding to each calibration body can be obtained in the end.

[0098] Step 403, based on the weight coefficients corresponding to the calibration bodies and the calibration atomic numbers of the calibration bodies, determine the target atomic number of the target object; and check whether there is a qualified calibration body in the obtained scatter correction information, and the calibration atomic number of the qualified calibration body meets the first set similarity requirement with the target atomic number.

[0099] As an example, the target atomic number of the target object can be determined by the following formula:

[0100] . .

[0101] in, represents the target atomic number of the target object, As described above, Represents the calibrated atomic number of the jth calibration body.

[0102] After obtaining the target atomic number of the target object, it is possible to check whether there is a calibration body that meets the conditions in the obtained scatter correction information, as described in step 402. Specifically, it is possible to check whether there is a calibration body in the scatter correction information that meets the first set similarity requirement between the calibration atomic number and the target atomic number. The first set similarity requirement here can be set according to actual conditions, such as being equal, or the absolute value of the difference is less than or equal to a first set threshold value such as 2, etc.

[0103] Step 404: if there is a qualified calibration body in the obtained scatter correction information, then based on the optimal scatter coefficient corresponding to the qualified calibration body, scatter correction is performed on the target image; and / or, if there is no qualified calibration body in the obtained scatter correction information, then based on finding a reference calibration body in the obtained scatter correction information, if the calibration atomic number of any reference calibration body meets the target atomic number and satisfies the second set similarity requirement; then scatter correction is performed on the target image based on the optimal scatter coefficient corresponding to each reference calibration body.

[0104] In this embodiment, the scatter correction of the target image based on the optimal scatter coefficient may be performed as follows:

[0105] Still taking the signal strength as an example, first substitute the optimal first scattering correction coefficient and the optimal second scattering correction coefficient in the optimal scattering coefficient into the above scattering physical model to obtain the following scattering components: ; Then, the target image can be corrected for scattering based on the scattering component. For example, the signal intensity at each pixel position can be corrected for scattering according to the following formula:

[0106] .in, represents the above-mentioned scattered component.

[0107] In addition, in this embodiment, the second similarity requirement is different from the first similarity requirement, for example, the atomic number difference may be greater than the first threshold and less than the second threshold, such as 4, and so on.

[0108] In step 404, scatter correction is performed on the target image based on the optimal scatter coefficient corresponding to each reference calibration body. There are many implementation methods in specific implementation, such as:

[0109] As an embodiment, based on the optimal scattering coefficients corresponding to each reference calibration body, the optimal scattering coefficient corresponding to the target object is determined; based on the optimal scattering coefficient corresponding to the target object, scattering correction is performed on the target image.

[0110] Optionally, determining the optimal scattering coefficient corresponding to the target object based on the optimal scattering coefficient corresponding to each reference calibration body includes: if the number of reference calibration bodies is greater than 1, determining the optimal scattering coefficient corresponding to the target object based on the reference weight of each reference calibration body and the optimal scattering coefficient corresponding to each reference calibration body; if the number of reference calibration bodies is 1, determining the optimal scattering coefficient corresponding to the target object based on the optimal scattering coefficient corresponding to the reference calibration body.

[0111] Optionally, there are many ways to determine the reference weights of the reference calibration bodies, for example, the weight coefficients of the reference calibration bodies determined in step 402 can be directly used. Alternatively, the weight coefficients of the reference calibration bodies can be determined according to the following formula: .

[0112] in, represents the target atomic number mentioned above, is the reference weight of the rth reference calibration body, which is unknown at this time. is the calibrated atomic number of the rth reference calibration body.

[0113] After the reference weights of the reference calibration bodies are obtained, the optimal scattering coefficient corresponding to the target object can be determined based on the optimal scattering coefficient corresponding to each reference calibration body.

[0114] The optimal scattering coefficient corresponding to the target object is ,in, .

[0115] in, represents an optimal scattering coefficient, such as a first optimal scattering correction coefficient, or an optimal second scattering correction coefficient, It is the optimal scattering coefficient of the rth reference calibration volume, such as the optimal first scattering correction coefficient or the optimal second scattering correction coefficient.

[0116] Finally, the optimal scattering coefficient corresponding to the target object can be determined according to the above. After determining the optimal scattering coefficient corresponding to the target object, the optimal scattering coefficient corresponding to the target object can be substituted into the above scattering physical model to obtain the scattering component at any pixel position in the target image. Still taking the signal intensity as an example, the scattering component is:

[0117] .

[0118] As another embodiment, this embodiment can also substitute the optimal scattering coefficient corresponding to each reference calibration body into the generated scattering physical model to obtain the target image reference scattering component associated with each reference calibration body; based on the target image reference scattering component associated with each reference calibration body, determine the scattering component corresponding to the target image, and perform scattering correction on the target image based on the scattering component corresponding to the target image.

[0119] For example, the optimal scattering coefficient corresponding to any reference calibration body is substituted into the generated scattering physical model to obtain the reference scattering component of the target image associated with the reference calibration body as shown below:

[0120] .

[0121] in, , They respectively represent the optimal scattering coefficient corresponding to any reference calibration body.

[0122] Afterwards, based on the reference weights of each reference calibration body and the reference scattering components of the target image associated with each reference calibration body, the scattering components corresponding to the target image are determined. For example, the scattering components corresponding to the target image are:

[0123] .

[0124] After obtaining the scattering component corresponding to the target image, the target image can be corrected for scattering. For example, the direct component at each pixel position in the target image can be obtained according to the following formula:

[0125] .

[0126] If the optimal scattering coefficient of the calibration body also includes the optimal mass attenuation coefficient, the optimal mass attenuation coefficient of the target object can also be determined in the above manner. When performing scattering correction on the target object, the optimal mass attenuation coefficient can also be considered, which is similar to the above-mentioned optimal first scattering correction coefficient or the optimal second scattering correction coefficient.

[0127] As an example, if the target object is a known material, A more accurate picture may be obtained.

[0128] As an embodiment, the result after scattering correction can also be used as the target image; based on the grayscale value of each pixel position in the target image and the grayscale value function corresponding to each calibration body in the obtained scattering correction information, the weight coefficient corresponding to each calibration body is determined; based on the weight coefficient corresponding to each calibration body and the calibrated atomic number of each calibration body, the target atomic number of the target object is determined; if the currently determined target atomic number and the target atomic number of the target object determined last time meet the third set similarity requirement, the current target image is output; if the currently determined target atomic number and the target atomic number of the target object determined last time do not meet the third set similarity requirement, return to the step of checking whether there is a qualified calibration body in the obtained scattering correction information.

[0129] So far, completed Figure 4 Description of the process shown.

[0130] pass Figure 4 As can be seen from the process shown, this embodiment constructs a scattering physics model with the help of a calibration body to determine the optimal first scattering correction coefficient and the optimal second scattering correction coefficient. Subsequently, the scattering components in the image can be corrected with the help of the optimal first scattering correction coefficient and the optimal second scattering correction coefficient, thereby realizing scattering correction of the image generated by the X-ray imaging device, effectively increasing the contrast of the image, improving the signal-to-noise ratio, etc., and facilitating more accurate material identification based on the image.

[0131] Furthermore, in this embodiment, a scattering physical model is constructed with the aid of a calibration body to determine an optimal first scattering correction coefficient and an optimal second scattering correction coefficient for subsequent scattering correction of an image, without the need for hardware or Monte Carlo simulation support.

[0132] Based on the same inventive concept as the above method embodiment, the device provided in the embodiment of the present application is described below:

[0133] See also Figure 5 , Figure 5 This is a diagram of the structure of the device provided in the embodiment of the present application. Figure 1 The process shown in FIG. 1 may include:

[0134] A model generation unit, used to generate a scattering physical model for describing scattering; the scattering physical model indicates: a pixel feature value P_total of a pixel point at any pixel position in a training image corresponding to any calibration body is related to a contribution value P_direct of a direct X-ray directly irradiated to the pixel position, a contribution value P_scatter of any scattered X-ray scattered to the pixel position, and a pixel feature value distribution P of the scattered X-ray; the P_scatter is related to a first scattering correction coefficient to be optimized, and the P is related to a second scattering correction coefficient to be optimized; wherein, under the premise that the P_total at any pixel position is determined, the scattering physical model represents the relationship between the P_direct at the pixel position and the first scattering correction coefficient and the second scattering correction coefficient;

[0135] The optimization unit is used to determine the P_total at each different pixel position in each region of interest in the training image, and to set the first scattering correction coefficient and the second scattering correction coefficient to different values ​​respectively, and to determine the P_direct at each different pixel position in each region of interest based on the scattering physical model; based on the set first loss function about the P_direct at each pixel position in each region of interest, determine the optimal first scattering correction coefficient and the optimal second scattering correction coefficient corresponding to the calibration body; the signal strength at any two different pixel positions in the same region of interest meets the set similarity requirement;

[0136] The optimal first scatter correction coefficient and the optimal second scatter correction coefficient are used to perform scatter correction on any image.

[0137] Optionally, the physical shape of the calibration body is a stepped block shape.

[0138] Optionally, the first loss function is expressed by the following formula: ;

[0139] Where n represents the index or number of the region of interest, P_direct represents any pixel position (x, y) in the nth region of interest; Represents the average value of P_direct for all pixel positions in the nth region of interest.

[0140] Optionally, the scattering physics model further corresponds to a second loss function, and the second loss function is a function of the mass of the calibration body;

[0141] The optimization unit further obtains the mass of the calibration body based on each training image corresponding to the calibration body; determines the optimal mass attenuation coefficient corresponding to the calibration body based on the mass of the calibration body obtained based on each training image and the actual mass of the calibration body, and in combination with the second loss function; wherein the optimal mass attenuation coefficient is used together with the optimal first scatter correction coefficient and the optimal second scatter correction coefficient to perform scatter correction on any image.

[0142] Optionally, the first loss function and the second loss function are set with corresponding weight parameters;

[0143] The determining of the optimal first scatter correction coefficient and the optimal second scatter correction coefficient corresponding to the calibration body based on the set first loss function of P_direct at each pixel position in each region of interest comprises: obtaining a first weight parameter corresponding to the first loss function; based on the first weight parameter and substituting P_direct at each pixel position in each region of interest into the first loss function, determining the optimal first scatter correction coefficient and the optimal second scatter correction coefficient corresponding to the calibration body according to the principle of minimizing the value of the first loss function;

[0144] The quality of the calibration object obtained for each training image and the actual quality of the calibration object, combined with the second loss function, determine the optimal quality attenuation coefficient corresponding to the calibration object, including: obtaining a second weight parameter corresponding to the second loss function; based on the second weight parameter and the second loss function and combined with the quality of the calibration object obtained from different training images and the actual quality of the calibration object, determine the optimal quality attenuation coefficient corresponding to the calibration object.

[0145] Optionally, obtaining the quality of the calibration body based on each training image corresponding to the calibration body includes:

[0146] For any training image, the quality of the calibration volume is obtained according to the following formula:

[0147] ;

[0148] m represents the mass of the calibration body, represents the signal intensity of any pixel position (x, y) within the nth region of interest in the training image. It represents the mass attenuation coefficient of the nth region of interest. The signal intensities at any two different pixel positions in the same region of interest meet the set similarity requirements.

[0149] Optionally, the second loss function is expressed by the following formula: ;

[0150] Among them, L2 represents the second loss function, represents the quality of the calibration body determined based on the kth training image of the calibration body, is the actual mass of the calibration body.

[0151] So far, completed Figure 5 Structural description of the device shown.

[0152] See also Figure 6 , Figure 6 Another device structure diagram provided in the embodiment of the present application. The device corresponds to Figure 4 The process shown in FIG. 1 may include:

[0153] An acquisition unit, used for acquiring a target image to be corrected; the target image is an image acquired by performing image acquisition on a target object;

[0154] a correction unit, configured to determine weight coefficients corresponding to respective calibration bodies based on the grayscale value of each pixel position in the target image and the grayscale value function corresponding to each calibration body in the obtained scatter correction information; the scatter correction information includes an optimal scatter coefficient corresponding to at least one calibration body, the optimal scatter coefficient including at least an optimal first scatter correction coefficient and an optimal second scatter correction coefficient; and

[0155] Determine the target atomic number of the target object based on the weight coefficient corresponding to each calibration body and the calibration atomic number of each calibration body; and check whether there is a calibration body that meets the conditions in the obtained scatter correction information, and the calibration atomic number of the calibration body that meets the conditions meets the first set similarity requirement with the target atomic number; and

[0156] If there is a qualified calibration body in the obtained scatter correction information, scatter correction is performed on the target image based on the optimal scatter coefficient corresponding to the qualified calibration body; and / or

[0157] If there is no qualified calibration body in the obtained scatter correction information, based on the fact that a reference calibration body is found in the obtained scatter correction information, and the calibration atomic number of any reference calibration body meets the second set similarity requirement with the target atomic number; scatter correction is performed on the target image based on the optimal scatter coefficient corresponding to each reference calibration body.

[0158] Optionally, determining the weight coefficients corresponding to the calibration bodies respectively based on the grayscale value of each pixel position in the target image and the grayscale value function corresponding to each calibration body in the obtained scatter correction information includes: determining the weight coefficients corresponding to the calibration bodies respectively according to the following formula:

[0159] .

[0160] in, represents a gray value function obtained based on the gray value of each pixel position in the target image, represents the weight coefficient of the jth calibration volume in the obtained scatter correction information, Represents the gray value function of the j-th calibration volume.

[0161] Optionally, performing scatter correction on the target image based on the optimal scatter coefficient corresponding to each reference calibration body includes:

[0162] Based on the optimal scattering coefficients corresponding to each reference calibration body, determining the optimal scattering coefficient corresponding to the target object; performing scattering correction on the target image based on the optimal scattering coefficient corresponding to the target object; or,

[0163] The optimal scattering coefficient corresponding to each reference calibration body is substituted into the generated scattering physical model to obtain the reference scattering component of the target image associated with each reference calibration body; based on the reference scattering component of the target image associated with each reference calibration body, the scattering component corresponding to the target image is determined, and scattering correction is performed on the target image based on the scattering component corresponding to the target image.

[0164] Optionally, determining the optimal scattering coefficient corresponding to the target object based on the optimal scattering coefficient corresponding to each reference calibration body includes:

[0165] If the number of reference calibration bodies is greater than 1, determining the optimal scattering coefficient corresponding to the target object based on the reference weight of each reference calibration body and the optimal scattering coefficient corresponding to each reference calibration body;

[0166] If the number of the reference calibration body is 1, the optimal scattering coefficient corresponding to the reference calibration body is used to determine the optimal scattering coefficient corresponding to the target object.

[0167] Optionally, determining the scattering component corresponding to the target image based on the target image reference scattering component associated with each reference calibration body comprises:

[0168] Based on the reference weights of the reference calibration volumes and the reference scattered components of the target image associated with the reference calibration volumes, the scattered components corresponding to the target image are determined.

[0169] Optionally, after performing scatter correction on the target image based on the optimal scatter coefficient corresponding to each reference calibration body, the method further includes:

[0170] The result after scatter correction is used as the target image;

[0171] Determine the weight coefficients corresponding to the calibration bodies based on the grayscale values ​​of the pixel positions in the target image and the grayscale value functions corresponding to the calibration bodies in the obtained scatter correction information; determine the target atomic number of the target object based on the weight coefficients corresponding to the calibration bodies and the calibration atomic number of the calibration bodies;

[0172] If the target atomic number currently determined and the target atomic number of the target object determined last time meet the third set similarity requirement, outputting the current target image;

[0173] If the currently determined target atomic number and the target atomic number of the target object determined last time do not meet the third set similarity requirement, then return to the step of checking whether there is a calibration object that meets the conditions in the obtained scatter correction information.

[0174] So far, completed Figure 6 Structural description of the device shown.

[0175] The present application also provides Figure 5 or Figure 6 The hardware structure of the device shown. Figure 7 , Figure 7 This is a structural diagram of an electronic device provided in an embodiment of the present application. Figure 7 As shown, the hardware structure may include: a processor and a machine-readable storage medium, the machine-readable storage medium storing machine-executable instructions that can be executed by the processor; the processor is used to execute the machine-executable instructions to implement the method disclosed in the above example of this application.

[0176] Based on the same application concept as the above method, an embodiment of the present application also provides a machine-readable storage medium, on which a number of computer instructions are stored. When the computer instructions are executed by a processor, the method disclosed in the above example of the present application can be implemented.

[0177] Exemplarily, the above-mentioned machine-readable storage medium can be any electronic, magnetic, optical or other physical storage device, which can contain or store information, such as executable instructions, data, etc. For example, the machine-readable storage medium can be: RAM (Radom Access Memory), volatile memory, non-volatile memory, flash memory, storage drive (such as hard disk drive), solid state drive, any type of storage disk (such as optical disk, DVD, etc.), or similar storage medium, or a combination thereof.

[0178] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A scatter correction method, characterized in that: The method includes: Generate a scattering physical model for describing scattering; the scattering physical model indicates that: the pixel feature value P_total of a pixel point at any pixel position in a training image corresponding to any calibration body is related to the contribution value P_direct of a direct X-ray directly irradiated to the pixel position, the contribution value P_scatter of any scattered X-ray scattered to the pixel position, and the pixel feature value distribution P of the scattered X-ray; the P_scatter is related to a first scattering correction coefficient to be optimized, and the P is related to a second scattering correction coefficient to be optimized; wherein, under the premise that the P_total at any pixel position is determined, the scattering physical model represents the relationship between the P_direct at the pixel position and the first scattering correction coefficient and the second scattering correction coefficient; On the premise of determining the P_total at each different pixel position in each region of interest in the training image, when the first scattering correction coefficient and the second scattering correction coefficient are set to different values, the P_direct at each different pixel position in each region of interest is determined based on the scattering physical model; based on the set first loss function about the P_direct at each pixel position in each region of interest, the optimal first scattering correction coefficient and the optimal second scattering correction coefficient corresponding to the calibration body are determined; the signal strength at any two different pixel positions in the same region of interest meets the set similarity requirement; The optimal first scatter correction coefficient and the optimal second scatter correction coefficient are used to perform scatter correction on any image.

2. The method according to claim 1, characterized in that The physical shape of the calibration body is a stepped block shape.

3. The method according to claim 1, characterized in that The first loss function is expressed by the following formula: ； Where n represents the index or number of the region of interest, P_direct represents any pixel position (x, y) in the nth region of interest; Represents the average value of P_direct for all pixel positions in the nth region of interest.

4. The method according to claim 1, characterized in that The scattering physics model also corresponds to a second loss function, and the second loss function is a function of the mass of the calibration body; The method further comprises: Based on each training image corresponding to the calibration body, respectively obtaining the quality of the calibration body; Determining an optimal mass attenuation coefficient corresponding to the calibration body based on the mass of the calibration body obtained based on each training image and the actual mass of the calibration body in combination with the second loss function; The optimal mass attenuation coefficient, together with the optimal first scatter correction coefficient and the optimal second scatter correction coefficient, are used to perform scatter correction on any image.

5. The method according to claim 4, characterized in that The first loss function and the second loss function are set with corresponding weight parameters; The determining of the optimal first scatter correction coefficient and the optimal second scatter correction coefficient corresponding to the calibration body based on the set first loss function of P_direct at each pixel position in each region of interest comprises: obtaining a first weight parameter corresponding to the first loss function; based on the first weight parameter and substituting P_direct at each pixel position in each region of interest into the first loss function, determining the optimal first scatter correction coefficient and the optimal second scatter correction coefficient corresponding to the calibration body according to the principle of minimizing the value of the first loss function; The quality of the calibration object obtained for each training image and the actual quality of the calibration object, combined with the second loss function, determine the optimal quality attenuation coefficient corresponding to the calibration object, including: obtaining a second weight parameter corresponding to the second loss function; based on the second weight parameter and the second loss function and combined with the quality of the calibration object obtained from different training images and the actual quality of the calibration object, determine the optimal quality attenuation coefficient corresponding to the calibration object.

6. The method according to claim 5, characterized in that The obtaining the quality of the calibration body based on each training image corresponding to the calibration body comprises: For any training image, the quality of the calibration volume is obtained according to the following formula: ； m represents the mass of the calibration body, represents the signal intensity of any pixel position (x, y) within the nth region of interest in the training image. It represents the mass attenuation coefficient of the nth region of interest. The signal intensities at any two different pixel positions in the same region of interest meet the set similarity requirements.

7. The method according to any one of claims 4 to 6, characterized in that: The second loss function is expressed by the following formula: ; Among them, L2 represents the second loss function, represents the quality of the calibration body determined based on the kth training image of the calibration body, is the actual mass of the calibration body.

8. A scatter correction method, characterized in that: The method includes: Obtain a target image to be corrected; the target image is an image obtained by performing image acquisition on the target object; Based on the grayscale value of each pixel position in the target image and the grayscale value function corresponding to each calibration body in the obtained scatter correction information, the weight coefficient corresponding to each calibration body is determined; the scatter correction information includes at least one optimal scatter coefficient corresponding to the calibration body, and the optimal scatter coefficient includes at least: an optimal first scatter correction coefficient and an optimal second scatter correction coefficient; Determine the target atomic number of the target object based on the weight coefficient corresponding to each calibration body and the calibration atomic number of each calibration body; and check whether there is a calibration body that meets the conditions in the obtained scattering correction information, and the calibration atomic number of the calibration body that meets the conditions meets the first set similarity requirement with the target atomic number; If there is a qualified calibration body in the obtained scatter correction information, scatter correction is performed on the target image based on the optimal scatter coefficient corresponding to the qualified calibration body; and / or If there is no qualified calibration body in the obtained scatter correction information, based on the fact that a reference calibration body is found in the obtained scatter correction information, and the calibration atomic number of any reference calibration body meets the second set similarity requirement with the target atomic number; scatter correction is performed on the target image based on the optimal scatter coefficient corresponding to each reference calibration body.

9. The method according to claim 8, characterized in that The determining of the weight coefficients corresponding to the calibration bodies based on the grayscale value of each pixel position in the target image and the grayscale value function corresponding to each calibration body in the obtained scatter correction information includes: determining the weight coefficients corresponding to the calibration bodies according to the following formula: in, represents a gray value function obtained based on the gray value of each pixel position in the target image, represents the weight coefficient of the jth calibration volume in the obtained scatter correction information, Represents the gray value function of the j-th calibration volume.

10. The method according to claim 8, characterized in that The performing scatter correction on the target image based on the optimal scatter coefficient corresponding to each reference calibration body comprises: Based on the optimal scattering coefficients corresponding to each reference calibration body, determining the optimal scattering coefficient corresponding to the target object; performing scattering correction on the target image based on the optimal scattering coefficient corresponding to the target object; or, The optimal scattering coefficient corresponding to each reference calibration body is substituted into the generated scattering physical model to obtain the reference scattering component of the target image associated with each reference calibration body; based on the reference scattering component of the target image associated with each reference calibration body, the scattering component corresponding to the target image is determined, and scattering correction is performed on the target image based on the scattering component corresponding to the target image.

11. The method according to claim 10, characterized in that The determining the optimal scattering coefficient corresponding to the target object based on the optimal scattering coefficient corresponding to each reference calibration body includes: If the number of reference calibration bodies is greater than 1, determining the optimal scattering coefficient corresponding to the target object based on the reference weight of each reference calibration body and the optimal scattering coefficient corresponding to each reference calibration body; If the number of the reference calibration body is 1, the optimal scattering coefficient corresponding to the reference calibration body is used to determine the optimal scattering coefficient corresponding to the target object.

12. The method according to claim 10, characterized in that The determining of the scattering component corresponding to the target image based on the target image reference scattering component associated with each reference calibration body comprises: Based on the reference weights of the reference calibration volumes and the reference scattered components of the target image associated with the reference calibration volumes, the scattered components corresponding to the target image are determined.

13. The method according to claim 8, characterized in that After performing scatter correction on the target image based on the optimal scatter coefficient corresponding to each reference calibration body, the method further includes: The result after scatter correction is used as the target image; Determine the weight coefficients corresponding to the calibration bodies based on the grayscale values ​​of the pixel positions in the target image and the grayscale value functions corresponding to the calibration bodies in the obtained scatter correction information; determine the target atomic number of the target object based on the weight coefficients corresponding to the calibration bodies and the calibration atomic number of the calibration bodies; If the target atomic number currently determined and the target atomic number of the target object determined last time meet the third set similarity requirement, outputting the current target image; If the currently determined target atomic number and the target atomic number of the target object determined last time do not meet the third set similarity requirement, then return to the step of checking whether there is a calibration object that meets the conditions in the obtained scatter correction information.

14. An electronic device, characterized in that: The electronic device includes: a processor and a machine-readable storage medium; The machine-readable storage medium stores machine-executable instructions that can be executed by the processor; The processor is used to execute machine executable instructions to implement the method steps of any one of claims 1-7 or 8 to 13.

Images