An adaptive coded imaging positioning method
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- INST OF HIGH ENERGY PHYSICS CHINESE ACAD OF SCI
- Filing Date
- 2023-04-27
- Publication Date
- 2026-06-26
Smart Images
Figure CN117075176B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of gamma ray detection and imaging applications, and in particular relates to a coding imaging positioning method that can adaptively adjust the coded plate processing shape and the working state of the imaging detector. Background Technology
[0002] In the field of gamma-ray detection and imaging, gamma-ray imaging technology can intuitively indicate the azimuth and intensity of a radiation source through hotspot imaging, and has broad application prospects in nuclear radiation monitoring, nuclear facility decommissioning, nuclear emergency response, nuclear security, nuclear medicine imaging, and astronomical gamma-ray burst observation. Currently, the mainstream gamma-ray imaging techniques are coded aperture imaging and Compton imaging. For gamma-ray sources with energies less than 200 keV, coded aperture imaging is more efficient and is therefore widely used. However, limited by the machining capabilities of the coded plate and the actual operating conditions of the imaging detector, coded imaging systems often struggle to achieve ideal design conditions. For example, the coded plate requires an additional cross-shaped frame for fixation, affecting the actual coded projection pattern. Similarly, the imaging detector is a pixel-type position-sensitive detector, which often suffers from dead pixels or missing pixels. Furthermore, when splicing large-area detector arrays, gaps can form between modules, creating dead zones. These factors prevent the detector projection obtained by the imaging system from reaching ideal values, affecting the imaging quality and positioning accuracy of the radiation source.
[0003] Research on coded imaging technology, currently reported in domestic and international literature, mainly focuses on decoding imaging methods. It doesn't address methods for adjusting and adaptively optimizing imaging methods based on the actual operating conditions of the imaging system. All existing research on coded-decoding imaging techniques and methods is based on the premise that the encoding board and detector are operating stably. For example, in terms of algorithms, early convolutional direct decoding image reconstruction algorithms were applicable to coded patterns such as URA, MURA, PNP, Singer, and RANDOM. The core of this algorithm lies in designing a corresponding decoding matrix based on the characteristics of the coded pattern, ensuring that the system point spread function is an ideal delta function. When the projection is ideal, the decoded image has no inherent artifacts and a high signal-to-noise ratio. Building upon direct decoding imaging, compressed sensing imaging algorithms and neural network algorithms were later developed, which can improve imaging sensitivity to some extent when the number of samples obtained by the coded imaging system is low. Unlike the aforementioned single-shot computational imaging algorithms, iterative imaging algorithms based on the maximum likelihood method are also widely used. These significantly improve the situation where direct decoding imaging has large inherent system artifacts, but they suffer from image distortion due to excessive iterations. Besides imaging algorithms, research on coded imaging technology also focuses on novel designs for encoding plates and detectors, such as coded patterns with different aperture ratios, detection of multiple scattering events by detectors, and depth effect studies. However, current coded imaging localization methods rely on the coded imaging system operating normally or meeting expected geometric conditions. In some special scenarios, such as when the encoding plate area is large and additional support structures are needed, the coded pattern changes, leading to system artifacts during image reconstruction using existing imaging algorithms. With prolonged use, pixel-type imaging detectors may develop bad pixels or slit dead zones during module stitching, preventing the acquisition of accurate coded projection patterns and thus introducing deviations in decoding imaging localization. These system instabilities were not addressed in previous coded imaging localization methods. Therefore, achieving accurate coded imaging and X-ray source azimuth localization is a pressing technical problem to be solved when the geometry of the encoding plate and the performance of the detector are not ideally designed in actual operation. Summary of the Invention
[0004] As a supplement and methodological innovation to existing coding imaging positioning methods, the purpose of this invention is to propose a coding imaging positioning method that can adaptively adjust the processing shape of the coding plate and the working state of the imaging detector, providing a new technical means for coding imaging and positioning methods. This invention can correct the coding process according to the actual opening and closing shape of the coding plate and the bad pixels and slit characteristics of the imaging detector, thereby eliminating image artifacts caused by unstable factors of the coding plate and detector, and improving the imaging quality and positioning accuracy of the X-ray source.
[0005] This invention proposes an adaptive coding imaging localization method, the flowchart of which is as follows: Figure 1 As shown, the specific implementation steps include:
[0006] 1. Based on the mechanical processing shape of the encoding board, the actual encoding pattern is given, and a decoding matrix is generated based on the encoding pattern, the dimension of the imaging detector projection, and the expected dimension of the image. The dimension of the imaging detector projection can be obtained from the actual coverage area of the imaging detector.
[0007] 2. Give the pixel index positions of the projection matrix corresponding to the bad pixels and slit dead zones of the imaging detector.
[0008] 3. Correct the pixel values corresponding to the matrix elements in the decoding matrix by setting them to 0 according to the index positions in step 2. When the ray beam emitted from the ray source is modulated by the encoding plate and incident on the imaging detector to form an encoded pattern projection, the corrected decoding matrix is used to perform decoding calculations on the encoded pattern projection to generate a decoded image. A pixel value of 0 represents that the ray does not pass through, that is, the detection efficiency of bad pixels or dead zone pixels of the imaging detector is 0, which is equivalent to the encoding plate pixel at the corresponding pixel position being a closed aperture, thereby achieving normalization of detection efficiency.
[0009] 4. Proceed to the X-ray source localization process, locate the position of the X-ray source from the decoded image, and provide the azimuth angle of the corresponding hot spot of the X-ray source.
[0010] Furthermore, the pixel array size of the encoding board is Px×Py, where Px is the number of pixels in the X direction and Py is the number of pixels in the Y direction; the expected X-direction dimension of the imaging is Nx and the Y-direction dimension is Ny, and the X-direction dimension of the imaging detector projection is Mx and the Y-direction dimension is My. Then, the size of the decoding matrix is Lx×Ly; where Lx = Mx + Nx - 1 and Ly = My + Ny - 1. The Px×Py region at the center of the decoding matrix is the encoded pattern region, and the values of the peripheral pixels are all 0.
[0011] Furthermore, Px = Py, Nx = Ny, Mx = My, Lx = Ly.
[0012] Furthermore, the method for locating the position and azimuth of the hot spot corresponding to the radiation source is as follows:
[0013] 41) Obtain the maximum pixel value ima_max in the decoded image O and its corresponding pixel position row and column coordinate indices [ima_max_row, ima_max_column];
[0014] 42) Taking the pixel corresponding to the maximum pixel value in the decoded image O as the center, obtain a heat value matrix region peak_area;
[0015] 43) Based on the pixel position at the maximum pixel value ima_max, obtain the second maximum value x_ima_max_2 in the X direction and its corresponding row and column coordinate indices [x_ima_max_2_row, x_ima_max_2_column], and the second maximum value y_ima_max_2 in the Y direction and its corresponding row and column coordinate indices [y_ima_max_2_row, y_ima_max_2_column] in the peak area of the heat value matrix; then calculate and obtain the X / Y coordinate index weight values peak_X / peak_Y of the hot spot peak in the real heat value area;
[0016] 44) The X / Y azimuth angles peak_phi / peak_theta of the hotspot are calculated based on the expected imaging dimension, the distance S between the encoding plate and the imaging detector, and the X / Y pixel index weight values.
[0017] Furthermore, the X-coordinate index weight value Y-coordinate index weight value Azimuth in the X direction Azimuth in the Y direction
[0018] The dimension of the imaging detector projection is determined based on the actual coverage area of the imaging detector.
[0019] Compared with the prior art, the positive effects of the present invention are as follows:
[0020] 1) When the encoding pattern of the encoding board changes due to machining, the decoding matrix can be adaptively corrected.
[0021] 2) When there are bad pixels and slit dead zones in the imaging detector, the decoding matrix elements can be corrected according to their corresponding position indices to reduce the noise of the decoded and reconstructed image.
[0022] 3) When the geometry of the encoder board and the performance of the detector are not ideal under actual working conditions, adaptive imaging adjustment is performed by modifying the decoding matrix, which improves the accuracy of encoded imaging and X-ray source azimuth positioning.
[0023] This invention improves the adaptive dynamic adjustment capability of the coding imaging system. Even when the geometry of the coding plate and the performance of the detector are not ideal under actual operating conditions, it can achieve precise coding imaging and X-ray source azimuth positioning. This invention provides a supplement and methodological innovation to the field of nuclear radiation detection imaging, eliminating image artifacts caused by uncertainties in the coding plate and detector in the coding imaging system, improving the imaging quality and positioning accuracy of the X-ray source, and has broad application prospects in nuclear radiation monitoring, nuclear facility decommissioning, nuclear emergency response, and nuclear security. Attached Figure Description
[0024] Figure 1 This is a flowchart of the method of the present invention.
[0025] Figure 2 This is a schematic diagram of the encoder board.
[0026] Figure 3 This is a schematic diagram of the coding pattern.
[0027] Figure 4 This is a schematic diagram of the decoding matrix.
[0028] Figure 5 This is a schematic diagram of the imaging detector.
[0029] Figure 6 This is a schematic diagram of the decoded image.
[0030] Figure 7 This is a region map of the calorific value matrix. Detailed Implementation
[0031] The present invention will now be described in further detail with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.
[0032] like Figure 1 As shown, the steps of the present invention include:
[0033] 1. Based on the mechanical processing shape of the encoding board, the actual encoding pattern is given, and the decoding matrix is generated based on the encoding pattern, the dimension of the detector projection, and the dimension of the expected imaging.
[0034] 1.1 The key components of an coded aperture imaging system are the encoding plate and the detector. A position-resolved imaging detector is commonly used. When a radiation source appears in the imaging field of view, the radiation beam is modulated by the encoding plate and projected onto the detector plane to form an coded pattern. The position of the radiation source is then reconstructed through a decoding process, i.e., decoding imaging. The pixel unit size of the encoding plate and the detector should be identical, while the decoding matrix needs to be generated based on the pattern characteristics of the encoding plate and the size of the detector's projection matrix.
[0035] 1.2 Assuming the pixel array size of the encoder board is P×P, for the sake of convenience, P=92 is set as an example. To effectively block modulated gamma rays, such as Figure 2 As shown, black pixels represent tungsten-copper alloy (value is 0), white pixels represent hollow areas (value is 1), and the mathematical form of the encoding board is a two-dimensional matrix with values of 0 and 1.
[0036] 1.3 When machining the encoding board, an additional frame is often needed to provide good support and fixation. Taking a cross-shaped frame as an example, assuming the frame runs through the entire encoding board and occupies a pixel width of 2, the actual shape of the encoding pattern is as follows: Figure 3 As shown.
[0037] 1.4 The size of the projection matrix is given according to the coverage area of the imaging detector. The pixel array covered by the imaging detector is set to M×M. M=28 is set as an example. The projection matrix of the imaging detector is represented by the symbol D.
[0038] 1.5 The expected imaging dimension is set to N, which means that the decoded image can resolve N×N azimuth angles in the imaging field of view. N = 75 is set, and the decoded image is represented by the symbol O, which is also a two-dimensional matrix.
[0039] 1.6 The size of the decoding matrix is L×L, L=M+N-1. According to the derivation above, L=28+75-1=102, where the 92×92 region at the center of the decoding matrix is... Figure 3 The coded pattern shown has all outer pixels with values of 0. The decoding matrix, denoted by the symbol A, is also a two-dimensional matrix, as shown below. Figure 4 As shown.
[0040] 2. Give the pixel index positions of the projection matrix corresponding to the bad pixels and slit dead zones of the imaging detector.
[0041] like Figure 5 As shown, the matrix size of the imaging detector is 28×28. The projection count of the pixels corresponding to bad pixels and slit dead zones is 0. The corresponding pixel index positions of bad pixels and slit dead zones are given. There are a total of S pixels, and the corresponding index positions are S in total according to row and column coordinates. The index position list is represented by the symbol T, which has S elements, and each element corresponds to a row and column coordinate.
[0042] 3. Correct the matrix elements corresponding to the decoding matrix A according to the index position in step 2, that is, set the corresponding pixel value to 0. When the ray beam emitted by the ray source is modulated by the encoding plate and incident on the imaging detector to form an encoded pattern projection, use the corrected decoding matrix A to perform decoding calculation on the encoded pattern projection to generate a decoded image.
[0043] 3.1 The decoding calculation formula is as follows:
[0044]
[0045] Among them, O i,j A represents the value of the element with row and column indices (i,j) in the decoded image. m,n This represents the value of the element with row and column indices (m, n) in the decoding matrix.
[0046] 3.2 Based on the index position list T in step 2, correct the pixels corresponding to the decoding matrix elements. The decoding matrix elements are:
[0047]
[0048] Among them, A m,n The range is M×M, which is 28×28. Iterate through each element of the index position list T, and set the pixel value corresponding to the decoding matrix element to 0 according to the row and column coordinates of the element in T.
[0049] 3.3 Combining steps 3.1 and 3.2, complete the calculation of the decoded image O, as follows: Figure 6 As shown.
[0050] 4. Proceed to the X-ray source localization process, locate the X-ray source position from the decoded image, and provide the azimuth angle of the corresponding hot spot.
[0051] 4.1 Give the maximum pixel value ima_max of the decoded image and the row and column coordinates of the pixel position of the maximum value [ima_max_row,ima_max_column].
[0052] 4.2 Using the pixel corresponding to the maximum value in the decoded image as the center, obtain a 3×3 heat value matrix region, peak_area, with a total of 9 elements, as shown below. Figure 7 As shown.
[0053] 4.3 Using the pixel position of the maximum value ima_max as the reference, obtain the second maximum value x_ima_max_2 in the X direction and its corresponding row and column coordinate indices [x_ima_max_2_row, x_ima_max_2_column], and the second maximum value y_ima_max_2 in the Y direction and its corresponding row and column coordinate indices [y_ima_max_2_row, y_ima_max_2_column] in the peak area, and calculate the X / Y coordinate index weight values peak_X / peak_Y of the hotspot peak in the real hotspot region:
[0054]
[0055]
[0056] 4.4 Based on the distance S between the encoder and the detector, and the imaging dimension N (in this example, the dimensions in both the X and Y directions are N; this also applies when the dimensions in the X and Y directions are different), the X / Y azimuth angles peak_phi / peak_theta of the hotspot can be calculated based on the X / Y pixel index weight values:
[0057]
[0058]
[0059] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. An adaptive coded imaging localization method, comprising the following steps: 1) Obtain the actual encoding pattern according to the mechanical processing shape of the encoding board, and generate the decoding matrix A according to the encoding pattern, the dimension of the imaging detector projection and the dimension of the expected imaging; 2) Generate the projection matrix D of the imaging detector, and obtain the pixel index positions of the bad pixels and the pixels corresponding to the slit dead zone of the imaging detector in the projection matrix D. 3) Set the pixel value corresponding to the corresponding matrix element in the decoding matrix A to 0 according to the pixel index position. A pixel value of 0 means that the ray does not pass through. When the ray beam emitted by the ray source is modulated by the encoding plate and incident on the imaging detector to form an encoded pattern projection, the encoded pattern projection is decoded and calculated using the modified decoding matrix A to generate a decoded image O. 4) Locate the position and azimuth of the hot spot corresponding to the ray source from the decoded image O.
2. The method according to claim 1, characterized in that, The pixel array size of the encoding board is Px×Py, where Px is the number of pixels in the X direction and Py is the number of pixels in the Y direction; the expected X-direction dimension of the imaging is Nx and the Y-direction dimension is Ny, and the X-direction dimension of the imaging detector projection is Mx and the Y-direction dimension is My. Then the size of the decoding matrix is Lx×Ly; where Lx = Mx + Nx - 1 and Ly = My + Ny - 1. The Px×Py region at the center of the decoding matrix is the encoded pattern region, and the values of the peripheral pixels are all 0.
3. The method according to claim 2, characterized in that, Px=Py, Nx=Ny, Mx=My, Lx=Ly.
4. The method according to claim 1, 2, or 3, characterized in that, The method for locating the position and azimuth of the hot spot corresponding to the radiation source is as follows: 41) Obtain the maximum pixel value ima_max in the decoded image O and its corresponding pixel position row and column coordinate indices [ima_max_row, ima_max_column]; 42) Taking the pixel corresponding to the maximum pixel value in the decoded image O as the center, obtain a heat value matrix region peak_area; 43) Based on the pixel position at the maximum pixel value ima_max, obtain the second maximum value x_ima_max_2 in the X direction and its corresponding row and column coordinate indices [x_ima_max_2_row, x_ima_max_2_column], and the second maximum value y_ima_max_2 in the Y direction and its corresponding row and column coordinate indices [y_ima_max_2_row, y_ima_max_2_column] in the peak area of the heat value matrix; then calculate and obtain the X / Y coordinate index weight values peak_X / peak_Y of the hot spot peak in the real heat value area; 44) The X / Y azimuth angles peak_phi / peak_theta of the hotspot are calculated based on the expected imaging dimension, the distance S between the encoding plate and the imaging detector, and the X / Y pixel index weight values.
5. The method according to claim 4, characterized in that, X-coordinate index weight value Y-coordinate index weight value Azimuth in the X direction Azimuth in the Y direction 6. The method according to claim 1, characterized in that, The dimension of the imaging detector projection is determined based on the actual coverage area of the imaging detector.