CN116125517B

CN116125517B - Single-pixel radiation imaging method and system based on rotation measurement

Info

Publication number: CN116125517B
Application number: CN202310120827.5A
Authority: CN
Inventors: 李三刚; 廖山; 周智; 杨明; 谭清山; 苏蓉蓉; 赵楚湘; 成毅
Original assignee: Chengdu Univeristy of Technology
Current assignee: Chengdu Univeristy of Technology
Priority date: 2023-02-13
Filing date: 2023-02-13
Publication date: 2025-06-17
Anticipated expiration: 2043-02-13
Also published as: CN116125517A

Abstract

The present invention discloses a single-pixel radiation imaging method and system based on rotation measurement, which determines the beam area and boundary line in the image; calculates the area of each pixel of the object passed by the beam at different angles according to the situation that the boundary line passes through the pixel; obtains the ray field after the sub-coding plate is rotated and modulated according to the preset angle; loops through all pixels to obtain the system matrix, and uses the reconstruction algorithm to combine the system matrix to obtain the reconstructed image of the object. The new recovery scheme provided by the method increases the dimension of measurement form change, improves the translation measurement to rotation measurement, uses the spatial change of a single coding plate to achieve equivalent modulation, and constructs an algorithm to obtain the system matrix at different angles. It is achieved by multiplexing a coding plate through rotation, and a very small number of coding plates are used to complete the full modulation of the ray field to obtain high-quality imaging results. Multiple modulations of the ray field are achieved, and the utilization rate of a single coding plate is improved, so as to achieve the purpose of reducing costs and manufacturing difficulty.

Description

Single-pixel radiation imaging method and system based on rotation measurement

Technical Field

The invention relates to the technical field of single-pixel radiation imaging, in particular to a single-pixel radiation imaging method and system based on rotation measurement.

Background

Single pixel imaging classified into radiation imaging is different from general single pixel imaging. Since the ray penetration power is far stronger than that of light, the encoding of the incident ray can only be achieved at present by the physical attenuation of the ray in the encoding plate.

Considering single-pixel radiation imaging with a general pixel number k x k, the different code plates required for each measurement need to be all fabricated, and the modulation of the radiation is accomplished by mechanical movement. The number of pixels of the coding plates required by the imaging is as high as k-4, the manufacturing difficulty and the manufacturing cost of the huge number of coding plates are high, and the mechanical movement of the coding plates greatly prolongs the measurement time.

One solution for achieving a fast measurement is to consider that it is not necessary to move or replace one part code plate completely at a time, and only one column of the plate is moved after each measurement, so that the modulation effect is regarded as being changed. Experiments prove that the scheme ensures the imaging quality while the miniaturization of the coding plate is considered. Overall, such a solution allows for a rapid measurement while at the same time greatly reducing the size of the code plate (manufacturing difficulty).

Disclosure of Invention

In view of the above, it is an object of the present invention to provide a novel single-pixel radiation imaging method based on rotation measurement, which uses a rotary encoder plate for single-pixel radiation imaging.

In order to achieve the above purpose, the present invention provides the following technical solutions:

the invention provides a single-pixel radiation imaging method based on rotation measurement, which comprises the following steps:

Determining a boundary line of the beam region;

Calculating the areas of the beams passing through pixels of the object under different angles according to the passing condition of upper and lower boundary lines of the system matrix;

obtaining a ray field after the sub-coding plate is rotationally modulated according to a preset angle;

obtaining a system matrix by using the ray fields of all the sub-coding plates;

An object image is reconstructed using a reconstruction algorithm and a system matrix.

The coding plate comprises a plurality of sub-coding plates designed according to a preset mode, and is used for partially resisting rays emitted to the sub-coding plates, so that the rays are modulated by the sub-coding plates to obtain ray fields distributed in a beam state;

the code division plate is a stripe code plate, and the stripe code plate is provided with a region allowing rays to pass through.

Further, the beam region and boundary line are calculated as follows:

Dividing an image area into a plurality of pixel grid areas;

respectively calculating an upper boundary ray and a lower boundary ray according to a beam boundary equation;

determining a beam region from the upper boundary ray and the lower boundary ray;

Further, the system matrix is calculated as follows:

circularly traversing along an upper boundary line and a lower boundary line to calculate the areas of the beams (ray fields) passing through each pixel of the object under different angles;

correcting the area of the ray passing through the pixel according to the condition that the two boundary lines pass through the pixel;

The system matrix is obtained using the ray fields of all the subcode plates.

Further, the beam boundaries are calculated as follows:

Let the beam boundary equation be y=mx+b, its longitudinal intercept intersecting a pixel is d ₁ and d ₂ in succession, where the pixel's lower left corner vertex is (x _a,y_j), then the two intercept expressions are:

d₁＝mx_a+b-y_j

d₂＝m(x_a+δ)+b-y_j

determining the ray condition and the passing pixel number according to d ₁、d₂;

Wherein d ₁ represents the intercept of the upper boundary line to the pixel, d ₂ represents the intercept of the lower boundary line to the pixel, delta represents the side length of the unit pixel, m represents the slope of the current beam, b represents the intercept of the current ray to the longitudinal axis of the coordinate system;

further, the calculating the area of the upper boundary ray and the area of the lower boundary ray passing through the pixel according to the intersection condition of the beam and the passing pixel specifically comprises the following four cases:

Case I:

The lower boundary:

Upper boundary:

case II:

The lower boundary:

Upper boundary:

Case III:

The lower boundary:

Upper boundary:

Case VI:

The lower boundary:

Upper boundary:

Wherein S _△、S_□ represents the area of the overlapping portion of the beam region and the pixel.

And establishing a matrix P as a system matrix for recording the modulation effect of all the coding plates, and recording the modulation effect of a coding plate after one rotation by P _i as one line of P.

Further, the system matrix is updated as follows:

Classifying according to the condition that two rays, namely an upper boundary ray and a lower boundary ray, pass through a pixel;

traversing pixel numbers passing along boundary lines, and updating the areas of corresponding pixel positions in a system matrix according to classification conditions for different passing conditions of upper and lower boundary rays;

Further, the classification is performed according to the situation that two rays of the upper and lower boundary rays pass through pixels, and specifically includes three situations, namely 1) that two rays pass through the same pixel grid, 2) that two different pixels respectively pass through, and that no pixel passing through between the two different pixels is contained, and 3) that two different pixels respectively pass through, and that other pixels not passing through are contained between the two different pixels;

the area of the corresponding pixel in the system matrix is updated in the following manner:

(1) Two rays pass through the same pixel grid;

W_i＝W_1-i+W_2-i-δ²;

Wherein W _i represents the modulation effect on the ith pixel, W _1-i represents the area of the lower boundary on the ith pixel, W _2-i represents the area of the upper boundary on the ith pixel;

(2) The modulation effect of the pixels which respectively pass through two different pixels and do not pass through the pixels is unchanged;

(3) The spatial element area among the two different pixels is recorded as W _i＝δ² if the two different pixels are penetrated respectively and other pixels which are not penetrated are contained between the two different pixels.

The single-pixel radiation imaging system based on rotation measurement provided by the invention comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, wherein the processor realizes the method when executing the program.

The invention has the beneficial effects that:

The invention provides a novel single-pixel radiation imaging method and system based on rotation measurement, which are used for determining a beam area and a boundary line in an image, calculating the areas of the beam passing through each pixel of an object under different angles according to the condition that the boundary line passes through the pixels, obtaining a ray field after rotation modulation of a sub-coding plate according to a preset angle, circularly traversing all the pixels to obtain a system matrix, and combining a reconstruction algorithm with the system matrix to obtain a reconstructed image of the object. The method provides a novel imaging scheme, increases the dimension of the change of a measurement form, improves translational measurement to rotation measurement, realizes ray modulation by using the spatial change of a single sub-coding plate, and constructs an algorithm to obtain a system matrix under different angles. The method has the advantages that the rotation multiplexing of one sub-coding plate is realized, the utilization rate of a single sub-coding plate is improved, the modulation of rays is completed by using few sub-coding plates, the purposes of reducing cost and manufacturing difficulty are achieved, and meanwhile, the imaging result with high quality is obtained.

Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objects and other advantages of the invention may be realized and obtained by means of the instrumentalities and combinations particularly pointed out in the specification.

Drawings

In order to make the objects, technical solutions and advantageous effects of the present invention more clear, the present invention provides the following drawings for description:

Fig. 1 is a schematic view of parallel beam CT in a rotation/translation mode.

Fig. 2 is a flow chart of a novel single-pixel radiation imaging method based on rotation measurement.

Fig. 3 is a representation of a single column striped code plate with a resolution of 8 x 8.

Fig. 4 is a diagram of beam model versus object (pixel).

FIG. 5 is a ray versus pixel graph.

Fig. 6 is a diagram showing the result of single-pixel radiation imaging under full sampling.

Fig. 7 is a single-pixel radiation imaging quality evaluation table.

Fig. 8 is a diagram showing the result of undersampled single pixel radiation imaging.

Fig. 9 is a single-pixel radiation imaging quality evaluation value for each scheme.

Detailed Description

The present invention will be further described with reference to the accompanying drawings and specific examples, which are not intended to limit the invention, so that those skilled in the art may better understand the invention and practice it.

Example 1

As shown in fig. 1, a parallel beam CT schematic diagram of the rotation/translation mode of fig. 1 shows the general concept of this embodiment, in which each ray path follows lambert-beer's law:

Wherein d is the thickness, mu _j is the attenuation coefficient, and I ₀ is the incident intensity of the rays;

In parallel beam CT imaging in a rotation/translation mode, the detector receives the sum of the intensities of rays after the beam passes through an object during each detection, i.e. the detector obtains a value during each detection, and simultaneously, the detector obtains a total intensity value during each sub-coding plate measurement in single-pixel imaging.

Therefore, using the existing concept of rotational measurement in CT imaging, the striped code plates are designed to mimic the form of the acquired data in CT imaging. After a cluster of light beams passing through the code plate stripes acts on an object to be detected, a detector in the single-pixel imaging system obtains a total intensity value.

As shown in fig. 2, the novel single-pixel radiation imaging method based on rotation measurement provided in this embodiment includes the following steps:

Determining a boundary line of the beam region;

Circularly traversing along the upper and lower boundary lines of the shot to calculate the areas of the beam areas passing through each pixel of the object under different angles;

Obtaining a modulation effect of the sub-coding plate after rotating according to a preset angle;

obtaining a system matrix by using the ray fields of all the sub-coding plates;

and using a reconstruction algorithm and a system matrix to perform reconstruction recovery.

The beam model building process and the calculation method in the embodiment are as follows:

the code plate in the embodiment comprises a plurality of sub-code plates designed according to a preset mode and used for partially resisting rays emitted to the sub-code plates, so that the rays are modulated by the sub-code plates to obtain a pattern of a beam shape;

The code dividing plate in the embodiment is a stripe code plate, wherein a row of areas allowing rays to pass through are arranged in the stripe code plate;

The code plate in this embodiment adopts a striped code plate, as shown in fig. 3, fig. 3 is a diagram showing all striped code plates with a resolution of 8×8, and a single-column striped code plate designed in a form of simulating CT acquisition data is used as the code plate for rotation measurement. For an imaged object with a resolution of 8 x 8, a total of 8 single column striped code plates are required, with the specific composition shown in fig. 3:

at a certain angle, the intensity distribution of the ray field is changed drastically after being modulated by the striped code plates, as shown in the following figure. Because of the striped code plates, only one or more columns of areas are not filled with material (only one column in this embodiment), radiation passes intact, while other areas passing through the split code plates are blocked by material. This results in that the modulated radiation field will have only one or more columns of regions where the radiation particles are present and will overall assume the beam shape.

As shown in fig. 4, fig. 4 is a graph of beam model versus object, the image area is divided into n=n×n pixel areas, and assuming that each pixel has a side length δ, the shadow area is a radiation field distribution area modulated by the encoding plate, i.e. a beam area, τ represents a distance between an upper boundary and a lower boundary of the beam, where it is noted that the beam passes through the pixels at different angles. To obtain high quality imaging results, it is most important to accurately calculate the areas of the beam passing through the pixels of the object at different angles, describing the effect of the code division plate on the modulation of the radiation.

Therefore, a system matrix in a form of rotation measurement needs to be accurately calculated by establishing a beam model, and the beam model calculation method comprises boundary analysis and calculation, cyclic calculation and quick traversal;

The beam region and boundary line in this embodiment are calculated as follows:

Dividing an image area into a plurality of pixel grid areas;

the system matrix in this embodiment is calculated as follows:

circularly traversing along the upper and lower boundary lines of the shot to calculate the areas of the beams (ray fields) passing through each pixel of the object under different angles;

The system matrix is obtained using the ray fields of all the subcode plates.

The beam boundaries in this embodiment are calculated as follows:

The intersection area of each angle beam and the pixel is the modulation effect of the stripe (beam) shape coding plate, and each beam with width delta is surrounded by an upper boundary line and a lower boundary line. Wherein fig. 4 is a part of an n×n system matrix, a red font (such as J-n, etc.) is the number of a pixel, J is the pixel to be discussed (i.e., a 'GJB'), and there are four cases where a ray intersects the pixel J (rays I, ii..v). For pixel J, the sitting at the lower left corner B ' is marked (x _a,y_j), the intercept of the boundary line to the pixel is calculated with this as the origin, assuming that the beam boundary equation y=mx+b, d ₁ is the intercept of the upper boundary line pair B ' a ', d ₂ is the intercept of the lower boundary line pair B ' a ', and the intercepts with a pixel are d ₁ and d ₂ in order, the two intercept expressions are:

d₁＝mx_a+b-y_j (2)

d₂＝m(x_a+δ)+b-y_j (3)

The ray condition and the passing pixel number are determined according to d ₁、d₂, namely, the ray condition and the next passing pixel number (traversing sequence) can be determined according to the positive and negative of d ₁、d₂, which is shown in table 1, wherein delta represents the pixel side length.

As shown in fig. 5, fig. 5 is a ray-pixel relative relationship diagram, in which the areas of the upper boundary ray and the lower boundary ray passing through the pixel are calculated according to the intersection condition of the beam and the pixel passing through the pixel;

in this embodiment, according to d ₁、d₂, the ray condition and the pixel number that will pass through next can be determined, and specific table 1 shows:

TABLE 1

Where J represents the number of the pixel and n represents the size (n×n) of the system matrix;

in this embodiment, the area of the upper boundary ray or the lower boundary ray passing through the pixel is calculated according to the intersection area of the beam and the pixel, and the areas of the beam and the pixel are discussed as follows for the rays (cases) I, ii..v, according to which the upper boundary and the lower boundary of the beam:

Case I:

The lower boundary:

Upper boundary:

case II:

The lower boundary:

Upper boundary:

Case III:

The lower boundary:

Upper boundary:

Case VI:

The lower boundary:

Upper boundary:

wherein S, S, etc. are represented as areas of overlapping portions of beam regions and pixels in fig. 5;

△□

The area of the pixel penetrated by the boundary of the beam area after the stripe coding plate rotates a certain angle and the calculation of the classification of the boundary can be realized through the formula, and the pixel to be calculated next is determined to recursion according to the ray classification until the beam exceeds the object space. The number of pixels through which the beam passes and the area of the beam area through the pixels are then obtained.

The system matrix provided in this embodiment is updated in the following manner:

Classifying according to the condition that two rays of the upper and lower boundary rays pass through pixels, wherein the three conditions specifically comprise 1) that the two rays pass through the same pixel grid, 2) that the two rays respectively pass through two different pixels, no pixel which does not pass through the two different pixels exists between the two different pixels, and 3) that the two different pixels respectively pass through the two different pixels, and the two different pixels contain other pixels which do not pass through the two different pixels;

traversing pixel numbers passing along boundary lines, and updating the area of corresponding pixel positions in a system matrix according to different passing conditions of upper and lower boundary rays:

(1) Two rays pass through the same pixel grid;

when two rays pass through the same pixel grid, namely, in the case of the I, calculating a system matrix, namely, integrating an upper boundary line and a lower boundary line to consider the total area of the beam to the pixels;

W_i＝W_1-i+W_2-i-δ² (14)

Wherein, pixel W _i represents the modulation effect on the ith pixel, W _1-i represents the area of the lower boundary on the ith pixel, W _2-i represents the area of the upper boundary on the ith pixel;

(3) Passing through two different pixels respectively, and other pixels which do not pass through are contained between the two different pixels, wherein for the space domain element in the space domain element, the area record is W _i＝δ²;

Example 2

The novel single-pixel radiation imaging method based on rotation measurement provided by the embodiment is illustrated by the following specific procedures:

The feasibility of a single pixel radiation imaging scheme under rotation measurement was verified. For this purpose, 32 striped code plates were used, measuring once per code plate revolution of 5.625 ° (180/32), for a total of 1024 measurements.

Fig. 6 shows the results of single-pixel radiation imaging under full sampling, wherein fig. (a) is an original image, (b) is a comparison, and shows the recovery effect of a general imaging system (recovery by using a Hadamard coding plate for full sampling and a compressed sensing algorithm), (c) is the recovery effect of a traditional second-order correlation algorithm under rotation measurement, and (d) is the recovery result of the compressed sensing algorithm under rotation measurement.

FIG. 7 is a single-pixel radiation imaging quality evaluation chart, from which it can be seen that a novel single-pixel radiation imaging method of rotating striped code plate measurement and computing a system matrix using a beam model is possible. The introduction of the compressed sensing algorithm can also greatly improve the image quality, experiments under undersampling are carried out in consideration of practical application, and in the traditional Hadamard single-pixel scheme, better imaging quality can be obtained by using an RR sequence at a sampling rate of about 20%. Each striped code plate is measured only 7 times in the corresponding rotation measuring scheme. The imaging results of the two schemes are shown in the following figures, wherein fig. 8 is a diagram showing the undersampled single-pixel radiation imaging result, in fig. 8, the right diagram is a traditional Hadamard single-pixel scheme, a recovery result of a compressed sensing algorithm is adopted, and the left diagram is a recovery result of a compressed sensing algorithm adopted by a rotary coding plate measuring scheme (in principle, the imaging effect of the latter is superior to that of the former).

Fig. 9 shows the evaluation value of single-pixel radiation imaging quality of each scheme, and it can be seen that when the sampling rate is 20%, the Hadamard single-pixel imaging result of the RR sequence has a certain degree of edge blurring compared with the original image, and compared with the condition under full sampling, the image PSNR is reduced from 24.3 to 18.9, and the reduction amplitude is 22.2%. The novel single-pixel imaging result has little visual perception change, and the PSNR of the image is objectively reduced from the original 37.8 to 32.6, and only reduced by 13%. The latter (rotation measurement scheme) is more advantageous in that the number of separate code plates required for imaging still only requires 32 blocks, while the number of separate code plates required in the Hadamard scheme still is up to 205, which is 6.5 times the rotation measurement scheme. It follows that in the case of undersampling, the novel single-pixel imaging scheme using rotation measurement not only requires a small number of separate code plates, but also ensures very high imaging quality.

In the embodiment, the system matrix corresponding to the rotation measurement calculated by using the Geant4 software simulation data and the proposed beam model is utilized to realize high-quality single-pixel radiation imaging under the rotation measurement by using a very small number of encoding plates. The imaging quality PSNR is as high as 37.8 at adequate sampling. Then, a related undersampling experiment is carried out, the reduction of imaging quality is not visually perceived at the sampling rate of 20%, and the PSNR of the image is slightly reduced to 32.6 but far exceeds the Hadamard imaging effect under the same condition. The imaging method under the new measurement form can greatly reduce the requirement of an imaging system on the number of split coding plates, simplify the manufacturing and operation difficulties of related modulation devices, and be beneficial to the development of single-pixel radiation imaging technology.

The above-described embodiments are merely preferred embodiments for fully explaining the present invention, and the scope of the present invention is not limited thereto. Equivalent substitutions and modifications will occur to those skilled in the art based on the present invention, and are intended to be within the scope of the present invention. The protection scope of the invention is subject to the claims.

Claims

1. A single pixel radiation imaging method based on rotation measurement, characterized in that it comprises the following steps:

Determining the boundary lines of the beam area;

Loop along the upper and lower boundary lines to calculate the area of each pixel of the object where the beam passes at different angles;

Obtaining the ray field after the sub-encoding plate is rotated and modulated according to a preset angle;

Use the ray fields of all sub-encoding plates to obtain the system matrix;

Use reconstruction algorithm and system matrix to perform reconstruction and recovery;

The beam area and boundary lines are calculated as follows:

Divide the image area into a number of pixel grid areas;

Calculate the upper boundary ray and the lower boundary ray respectively according to the beam boundary equation;

Determine a beam region according to an upper boundary ray and a lower boundary ray;

The area of the pixel passed by the upper boundary ray and the lower boundary ray is calculated respectively according to the intersection of the beam and the passed pixel, specifically including the following four cases:

Case I:

Lower Boundary:

Upper Boundary:

Case II:

Lower Boundary:

Upper Boundary:

Case III:

Lower Boundary:

Upper Boundary:

Case VI:

Lower Boundary:

Upper Boundary:

Among them, S _△ and S _□ represent the area of the overlap between the beam region and the pixel;

The system matrix is updated as follows:

Classify based on the situation where the upper and lower boundary rays pass through the pixels;

The pixel numbers passing through the boundary line are traversed, and for different passing situations of the upper and lower boundary rays, the area of the corresponding pixel position in the system matrix is updated according to the classification situation;

The classification is performed based on the situation that the upper and lower boundary rays pass through pixels, specifically including three situations: 1) two rays pass through the same pixel, 2) two rays pass through two different pixels respectively and there is no pixel between the two different pixels that neither ray passes through, 3) two rays pass through two different pixels respectively and there is other pixel between the two different pixels that neither ray passes through;

The area of the corresponding pixel in the system matrix is updated as follows:

(1) Two rays pass through the same pixel grid;

W _i =W _1-i +W _2-i -δ ² ;

Wherein, _Wi represents the modulation effect (area) on the i-th pixel; _W1-i represents the area of the lower boundary on the i-th pixel; W2 _-i represents the area of the upper boundary on the i-th pixel;

(2) Passing through two different pixels respectively and there is no pixel between the two different pixels that has not been passed through, the modulation effect of the pixels that meet the conditions remains unchanged;

(3) Passing through two different pixels respectively and there are other pixels between the two different pixels that have not been passed through, the spatial element modulation effect is recorded as W _i =δ ² ;

Wherein, d1 represents the intercept of the upper boundary line to the pixel; d2 represents the intercept of the lower boundary line to the pixel; δ represents the side length of the unit pixel; and m represents the slope of the current beam.

2. The single-pixel radiation imaging method based on rotation measurement as described in claim 1 is characterized in that: the sub-coding plate is designed in a preset manner to partially block the rays directed toward the sub-coding plate; so that the rays are modulated by each sub-coding plate to obtain a ray field distributed in a beam state.

3. The single-pixel radiation imaging method based on rotation measurement as described in claim 1 is characterized in that: the sub-coding plate is a stripe-shaped coding plate; and the stripe-shaped coding plate is provided with an area allowing rays to pass through.

4. The single-pixel radiation imaging method based on rotation measurement according to claim 1, characterized in that: the system matrix is calculated in the following manner:

The area of the pixel through which the ray passes is corrected according to the situation that the two boundary lines pass through the pixel;

Use the ray fields of all sub-encoding plates to obtain the system matrix;

The system matrix P is used to record the modulation effects of all encoding plates, and _Pi is a row of P, recording the modulation effect of a sub-encoding plate after one rotation.

5. The single-pixel radiation imaging method based on rotation measurement according to claim 1, characterized in that the boundary line of the beam area is calculated in the following manner:

Assume that the beam boundary equation is y=mx+b, and the longitudinal intercepts of the beam boundary at the intersection with a certain pixel are d1 and d2 respectively, where the vertex of the lower left corner of the pixel is ( _xa , _yj ), then the two intercept expressions are:

d ₁ = mx _a + by _j

d ₂ =m( _xa +δ)+by _j

Determine the ray status and the pixel number it passes through based on d1 and d2;

Among them, b represents the intercept of the current ray on the vertical axis of the coordinate system.

6. A single-pixel radiation imaging system based on rotation measurement, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any one of claims 1 to 5 when executing the program.