A method, system, and medium for energy calibration of a photon counting detector
Patent Information
- Application Number
- CN202611005094.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-07
- Publication Date
- 2026-08-04
AI Technical Summary
然而，此种方法的缺点是无法单独衡量出探测器晶体的能量响应非线性，在实际应用中往往受到计数率低、时间成本高、对环境变化适应性差以及难以实时补偿ASIC漂移等关键问题的限制
本发明首先消除专用集成电路（ASIC）对能量标定的影响，将像素S曲线调节到一致再进行能量标定，可以减小各像素间标定差异，使能量标定更加稳定。本发明方法直接使用X射线进行标定，无需滤过，标定过程更简化。
Smart Images
Figure CN122498869A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of medical imaging technology, specifically to a method, system, and medium for energy calibration of a photon counting detector. Background Technology
[0002] Photon-counting CT detectors identify and count incident photon energy by converting it into electrical charge. Several factors can cause the identified energy and the incident photon energy to be nonlinear, exhibiting a non-linear relationship. Two of the most significant factors include differences between the ASIC chip channels of the photon-counting detector, and potential minute defects in the detector crystal, such as impurities and deep energy levels.
[0003] Energy calibration of photon-counting CT detectors is fundamental to ensuring their performance. Insufficient energy calibration can lead to inaccurate detector threshold settings. Setting the threshold too low for low energy levels may cause background noise to be mistakenly counted as a signal, while setting it too high will miss true low-energy photons, resulting in a decrease in the count rate. Uncalibrated detectors will exhibit nonlinear energy responses, leading to distortion of the measured X-ray energy spectrum, especially in the high-energy range where a tailing effect may occur, affecting the accuracy of spectral decomposition. Lack of energy calibration can also amplify systematic errors. If the differences between channels of the detector's ASIC chip are not uniformly calibrated, these differences will be amplified over time, causing systematic deviations in image reconstruction results and making it difficult to guarantee repeatability across different time points or different detectors.
[0004] Conventional energy calibration methods involve acquiring data after X-ray exposure and then calibrating the energy differences between pixels using methods such as K-edge absorption or irradiation with various natural radioactive elements. Pixels with significantly different calibrated coefficients are considered bad pixels or pixels with poor performance. However, this method has drawbacks, including the inability to independently measure the nonlinearity of the detector crystal's energy response. In practical applications, it is often limited by key issues such as low count rate, high time cost, poor adaptability to environmental changes, and difficulty in real-time compensation for ASIC drift. Therefore, a simple method for directly calibrating the energy response curve of the detector crystal is urgently needed. Summary of the Invention
[0005] To address the problems of existing technologies, this invention provides a method, system, and medium for calibrating the energy of a photon counting detector.
[0006] This invention provides a method for energy calibration of a photon counting detector, characterized by comprising the following steps: Step 1: Perform steps 1.1 to 1.2 on the X-ray source under two different tube voltages kV1 and kV2 respectively to obtain the first calibration coefficient and the calibrated S-curve for each pixel except the reference pixel. Step 1.1: Acquire the S-curve of the count values of each pixel of the detector under different energy thresholds x under X-ray exposure conditions; Step 1.2: Select a reference pixel and calibrate the S-curves of the remaining pixels under two X-ray exposure conditions to obtain the first calibration coefficients k and b and the calibrated S-curve for each pixel. The calibrated S-curve is denoted as F(m,y), where m is the pixel number, y is the calibrated energy threshold, k represents k1 or k2, and b represents b1 or b2. Step 2: Calculate the logarithmic count value of the S-curve for each pixel and average it to obtain the average curve G(y). Fit the tail of the average curve G(y) to obtain the fitting threshold cutoff point th. The fitting threshold cutoff point th represents th1 or th2, where th1 is the threshold cutoff point obtained by fitting G(y) under the kV1 condition and th2 is the threshold cutoff point obtained by fitting G(y) under the kV2 condition. Step 3: Fit the data points (th1, kV1) and (th2, kV2) to obtain the relationship between the energy threshold x and the actual voltage kV and the second calibration coefficients c1 and c0; the second calibration coefficients for all pixels are c1 and c0. Step 4: Average the first calibration coefficients k1 and k2 for the same pixel under different tube voltages kV1 and kV2, and average b1 and b2 to obtain the average first calibration coefficient for this pixel. and According to the average first calibration coefficient and And the second calibration coefficients c1 and c0, to calculate the final energy calibration formula for each pixel and the third calibration coefficients p1 and p0.
[0007] Preferably, in step 1, the S-curve is denoted as S(m, x), where m is the pixel number, m ranges from 1 to M, and x is the energy threshold, x ranges from 1 to n.
[0008] Preferably, in step 1, the reference pixel is selected by averaging the horizontal coordinate values corresponding to half of the maximum value of the S-curve count of all pixels other than bad pixels, and finding the curve that is closest to the average value; the bad pixel is defined as: the count value of its S-curve at the energy threshold x deviates by more than 50% compared with the median count value of the S-curve of all pixels at the energy threshold x.
[0009] Preferably, in step 1, the calibration formula is: y = k∙x + b; the first calibration coefficients are the values of k and b that minimize the error function, and the error function is: Where S(ref, x) is the S-curve of the reference pixel, ref represents the reference pixel, x is the energy threshold, F(m,y) is the calibrated S-curve, m is the pixel number, M is the total number of pixels, and y is the calibrated energy threshold.
[0010] Preferably, in step 2, the formula for the average curve G(y) is: Wherein, F(m, y) is the calibrated S-curve, m is the pixel label, m ranges from 1 to M, M is the total number of pixels, and y is the value of x after calibration; the fitting threshold cutoff point th is the calibrated energy threshold y corresponding to the intersection of G(y) and the x-axis, the tail of the average curve G(y) is fitted using the least squares method, and the range of the tail of the average curve G(y) is 5% to 50% of the maximum value of G(y).
[0011] Preferably, in step 3, the relationship between the energy threshold x and the actual voltage kV is as follows: Where c1 and c0 are the second calibration coefficients, k and b are the first calibration coefficients, and th is the cutoff point for the fitting threshold.
[0012] Preferably, in step 4, the energy calibration formula is: Where z is the photon energy, x is the energy threshold, and p1 and p0 are the third calibration coefficients. c1 and c0 are the second calibration coefficients. and This is the average first calibration coefficient.
[0013] The present invention also provides a photon counting detector energy calibration system for implementing the above method, comprising: The input module is configured to irradiate the detector with an X-ray source under two different tube voltages kV1 and kV2, and to acquire the count value S-curve of each pixel of the detector under different energy thresholds x under the X-ray exposure conditions. The construction module is configured to: select a reference pixel; calibrate the S-curve of each pixel to obtain first calibration coefficients k and b and the calibrated S-curve; average the logarithmic count values of the calibrated S-curves of each pixel to obtain an average curve G(y); fit the tail of the average curve G(y) to obtain a fitting threshold cutoff point th, where the fitting threshold cutoff point th represents th1 or th2; fit the data points (th1, kV1) and (th2, kV2) to obtain the relationship between the energy threshold x and the actual voltage kV, as well as the second calibration coefficients c1 and c0; and average the first calibration coefficients of the same pixel under different tube voltages kV1 and kV2 to obtain the average first calibration coefficient of this pixel. and The energy calibration formula for each pixel and the third calibration coefficients p1 and p0 are obtained based on the average first and second calibration coefficients. The output module is configured to output the curve showing the correspondence between the calibrated energy threshold x and the photon energy z.
[0014] The present invention also provides a computer-readable storage medium having stored thereon a computer program for implementing the above-described photon counting detector energy calibration system.
[0015] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above.
[0016] By adopting the technical solution of the present invention, the following beneficial effects can be obtained: This invention first eliminates the influence of application-specific integrated circuits (ASICs) on energy calibration by adjusting the pixel S-curves to be consistent before energy calibration, which reduces calibration differences between pixels and makes energy calibration more stable. The method of this invention directly uses X-rays for calibration, eliminating the need for filtering and simplifying the calibration process.
[0017] Obviously, based on the above description of the present invention, and according to common technical knowledge and conventional methods in the field, various other modifications, substitutions or alterations can be made without departing from the basic technical concept of the present invention.
[0018] The following detailed embodiments further illustrate the above-described content of the present invention. However, this should not be construed as limiting the scope of the present invention to the following examples. All technologies implemented based on the above-described content of the present invention fall within the scope of the present invention. Attached Figure Description
[0019] Figure 1 This is a flowchart of the method of the present invention.
[0020] Figure 2 This is a schematic diagram of the S-curve for a single pixel.
[0021] Figure 3 S-curves for two different pixels.
[0022] Figure 4 This is the S-curve of all pixels that have been calibrated and overlapped after the first calibration coefficient.
[0023] Figure 5 An example diagram showing the threshold cutoff point th obtained by fitting the tail of the G(y) curve. Detailed Implementation
[0024] The algorithms for data acquisition, transmission, storage, and processing steps not specifically described in the following embodiments, as well as the hardware structures and circuit connections not specifically described, can all be implemented using publicly available information in the prior art.
[0025] Example 1: Energy Calibration System and Method for Photon Counting Detector The system in this embodiment includes: The input module is configured to irradiate the detector with an X-ray source under two different tube voltages kV1 and kV2, and to acquire the count value S-curve of each pixel of the detector under different energy thresholds x under the X-ray exposure conditions. The module is configured to: select a reference pixel; calibrate the S-curve of each pixel to obtain first calibration coefficients k and b, and the calibrated S-curve, denoted as F(m,y), where m is the pixel label and y is the calibrated energy threshold; average the logarithmic counts of the calibrated S-curves of each pixel to obtain an average curve base G(y); fit the tail of the G(y) curve to obtain a fitting threshold cutoff point th, where th represents th1 or th2; fit the data points (th1, kV1) and (th2, kV2) to obtain the relationship between the energy threshold x and the actual voltage kV, and the second calibration coefficients c1 and c0; and average the first calibration coefficients of the same pixel under different tube voltages kV1 and kV2 to obtain the average first calibration coefficient of that pixel. , The energy calibration formula for each pixel and the third calibration coefficients p1 and p0 are obtained based on the average first and second calibration coefficients. The output module is configured to output the curve showing the correspondence between the calibrated energy threshold x and the photon energy z.
[0026] The following practical application example illustrates the method for energy calibration of the photon counting detector using the above system. The flowchart is shown below. Figure 1 The specific steps are as follows: Irradiate with two different X-ray tube voltages kV1 and kV2 (e.g., 80kVp and 140kVp), and obtain the fitting threshold cutoff points of the two curves according to steps 1 to 4 below, corresponding to th1 and th2 respectively.
[0027] 1. Perform X-ray exposure, and scan each pixel (the pixel number is denoted as m, where m ranges from 1 to M) by changing different energy thresholds (the threshold value is denoted as x, where x ranges from 1 to n) to obtain the value of each pixel. Plot the count value curve of each pixel under different energy thresholds, which is called the S-curve, denoted as S(m, x). Figure 2 The example shows a schematic diagram of the S-curve of a single pixel. Before calibration, the S-curves of two different pixels do not overlap. Figure 3 ).
[0028] 2. Take the average of the x-coordinate values corresponding to the half-maximum S-curve count values of all pixels (after removing bad pixels; a bad pixel is defined as one whose S-curve count value at energy threshold x deviates by more than 50% compared to the median of the S-curve count values of all pixels at energy threshold x). Take the curve closest to the average as the reference pixel curve, denoted as ref. Subsequent algorithms will use this pixel as the reference to calibrate the S-curves of the remaining pixels.
[0029] 3. Perform a linear transformation on x, denoted as k and b as the intercept. Set initial values for k and b (e.g., k=0.5, b=0). The calibrated threshold is y=k∙x+b, and the calibrated S-curve is denoted as F(m, y), where y is the calibrated value of x. To obtain the calibration coefficients k and b, define an error function. Where S(ref,x) is the S-curve of the reference pixel, ref represents the reference pixel, x is the energy threshold, F(m,y) is the calibrated S-curve, m is the pixel label, and M is the total number of pixels. The calibration coefficients are the values of k and b that minimize the error function L(k,b). Each pixel obtains a pair of k and b values through the above calculations.
[0030] 4. After the above steps, the S-curves of all pixels will almost completely overlap. Figure 4 The average curve base G(y) is obtained by taking the logarithmic value of the S-curve count for each pixel and averaging the results. The tail portion of curve G(y) (the region from 20% to 50% of the maximum value of G(y)) is fitted with a quadratic polynomial, and the calibrated energy threshold y corresponding to the intersection of the fitted polynomial and the x-axis is calculated and denoted as the fitting threshold cutoff point th. An example of obtaining the threshold cutoff point th by fitting the tail portion of the G(y) curve is shown below. Figure 5th represents th1 or th2, where th1 is the threshold cutoff point obtained by fitting G(y) under the kV1 condition, and th2 is the threshold cutoff point obtained by fitting G(y) under the kV2 condition.
[0031] 5. Using the linear fitting formula kV=c1∙th+c0, fit the data points (th1, kV1) and (th2, kV2) to finally obtain the correlation coefficients c1 and c0.
[0032] 6. Since the calibration coefficients k and b for a pixel corresponding to kV1 and kV2 obtained in the above steps may have some fitting bias, the calibration coefficients k and b for the same pixel under kV1 and kV2 are averaged to obtain the average calibration coefficient for this pixel. and Perform the above operations on all pixels. Obtain the energy calibration formula for each pixel based on the average calibration coefficient and correlation coefficient, defining p1 and p0 as energy calibration coefficients, and letting... The energy calibration formula for each pixel is: .
[0033] This invention first eliminates the influence of ASIC on energy calibration by adjusting the pixel S-curves to be consistent before energy calibration, which reduces calibration differences between pixels and makes energy calibration more stable. The method of this invention directly uses X-rays for calibration, eliminating the need for filtering and simplifying the calibration process. In summary, the method of this invention simplifies the energy calibration process of photon counting detectors, makes the response changes of each pixel more consistent, and makes energy calibration more stable, showing great application potential.
Claims
1. A method for calibrating the energy of a photon counting detector, characterized in that, Includes the following steps: Step 1: Perform steps 1.1 to 1.2 on the X-ray source under two different tube voltages kV1 and kV2 respectively to obtain the first calibration coefficient and the calibrated S-curve for each pixel except the reference pixel. Step 1.1: Acquire the S-curve of the count values of each pixel of the detector under different energy thresholds x under X-ray exposure conditions; Step 1.2: Select a reference pixel and calibrate the S-curves of the remaining pixels under two X-ray exposure conditions to obtain the first calibration coefficients k and b and the calibrated S-curve for each pixel. The calibrated S-curve is denoted as F(m,y), where m is the pixel number, y is the calibrated energy threshold, k represents k1 or k2, and b represents b1 or b2. Step 2: Calculate the logarithmic count value of the S-curve for each pixel and average it to obtain the average curve G(y). Fit the tail of the average curve G(y) to obtain the fitting threshold cutoff point th. The fitting threshold cutoff point th represents th1 or th2, where th1 is the threshold cutoff point obtained by fitting G(y) under the kV1 condition and th2 is the threshold cutoff point obtained by fitting G(y) under the kV2 condition. Step 3: Fit the data points (th1, kV1) and (th2, kV2) to obtain the relationship between the energy threshold x and the actual voltage kV and the second calibration coefficients c1 and c0; the second calibration coefficients for all pixels are c1 and c0. Step 4: Average the first calibration coefficients k1 and k2 for the same pixel under different tube voltages kV1 and kV2, and average b1 and b2 to obtain the average first calibration coefficient for this pixel. and According to the average first calibration coefficient and And the second calibration coefficients c1 and c0, to calculate the final energy calibration formula for each pixel and the third calibration coefficients p1 and p0.
2. The method according to claim 1, characterized in that, In step 1, the S-curve is denoted as S(m, x), where m is the pixel number, ranging from 1 to M, and x is the energy threshold, ranging from 1 to n.
3. The method according to claim 1, characterized in that, In step 1, the reference pixel is selected by averaging the horizontal coordinate values corresponding to half of the maximum value of the S-curve count of all pixels other than bad pixels, and finding the curve that is closest to the average value. The bad pixel is defined as: the count value of its S-curve at the energy threshold x deviates by more than 50% compared with the median count value of the S-curve of all pixels at the energy threshold x.
4. The method according to claim 1, characterized in that, In step 1, the calibration formula is: y = k∙x + b; the first calibration coefficients are the values of k and b that minimize the error function, and the error function is: Where S(ref, x) is the S-curve of the reference pixel, ref represents the reference pixel, x is the energy threshold, F(m,y) is the calibrated S-curve, m is the pixel number, M is the total number of pixels, and y is the calibrated energy threshold.
5. The method according to claim 1, characterized in that, In step 2, the formula for the average curve G(y) is: Wherein, F(m, y) is the calibrated S-curve, m is the pixel label, m ranges from 1 to M, M is the total number of pixels, and y is the value of x after calibration; the fitting threshold cutoff point th is the calibrated energy threshold y corresponding to the intersection of G(y) and the x-axis, the tail of the average curve G(y) is fitted using the least squares method, and the range of the tail of the average curve G(y) is 5% to 50% of the maximum value of G(y).
6. The method according to claim 1, characterized in that, In step 3, the relationship between the energy threshold x and the actual voltage kV is as follows: Where c1 and c0 are the second calibration coefficients, k and b are the first calibration coefficients, and th is the cutoff point for the fitting threshold.
7. The method according to claim 1, characterized in that, In step 4, the energy calibration formula is: Where z is the photon energy, x is the energy threshold, and p1 and p0 are the third calibration coefficients. c1 and c0 are the second calibration coefficients. and This is the average first calibration coefficient.
8. A photon counting detector energy calibration system, characterized in that, To implement the method according to any one of claims 1-7, comprising: The input module is configured to irradiate the detector with an X-ray source under two different tube voltages kV1 and kV2, and to acquire the count value S-curve of each pixel of the detector under different energy thresholds x under the X-ray exposure conditions. The construction module is configured to: select a reference pixel; calibrate the S-curve of each pixel to obtain first calibration coefficients k and b and the calibrated S-curve; average the logarithmic count values of the calibrated S-curves of each pixel to obtain an average curve G(y); fit the tail of the average curve G(y) to obtain a fitting threshold cutoff point th, where the fitting threshold cutoff point th represents th1 or th2; fit the data points (th1, kV1) and (th2, kV2) to obtain the relationship between the energy threshold x and the actual voltage kV, as well as the second calibration coefficients c1 and c0; and average the first calibration coefficients of the same pixel under different tube voltages kV1 and kV2 to obtain the average first calibration coefficient of this pixel. and The energy calibration formula for each pixel and the third calibration coefficients p1 and p0 are obtained based on the average first and second calibration coefficients. The output module is configured to output the curve showing the correspondence between the calibrated energy threshold x and the photon energy z.
9. A computer-readable storage medium, characterized in that, It stores a computer program for implementing the photon counting detector energy calibration system of claim 8.
10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1-7.