Radiation substance automatic selection system based on electric rotary filtering disc

Abstract

The invention relates to the technical field of automatic switching and calibration of filters, and discloses an automatic radiation substance selection system based on an electric rotary filter disc, which utilizes a servo motor to drive an electric filter rotary disc to rotate, so that an industrial bar code scanner gradually scans bar codes on the filter disc until a required filter is found, and the automatic switching and calibration of the filter is realized. And then the position of the filter is corrected in a mode that light emitted by the cross laser transmitter penetrates through the light through hole to irradiate the spot-shaped laser receiving module, so that the center of the filter is overlapped with the beam center of the X-ray source, the effects of full-automatic switching and high-precision positioning of the filter are realized, the working efficiency is greatly improved, and the production cost is reduced. And errors caused by manual selection of radiation substances are reduced.

Description

Technical Field

[0001] This invention relates to the field of automatic filter switching and positioning technology, and in particular to an automatic radiation quality selection system based on an electrically driven rotating filter disk. Background Technology

[0002] A standard X-ray radiation field refers to a radiation environment constructed by precisely controlling X-ray source parameters (energy, dose rate, homogeneity, etc.) to have known radiation values ​​(such as air kerma, absorbed dose, dose equivalent, etc.) and meet specific metrological requirements. Its core objective is to provide a unified calibration benchmark for radiation measurement equipment (such as ionization chambers, dosimeters, and detectors), ensuring the consistency and traceability of measurement results across different equipment and scenarios. Since the mid-20th century, the rapid development of fields such as medical diagnostics (X-ray imaging), industrial non-destructive testing, and radiation protection has placed greater demands on standardized X-ray radiation fields. According to the international standard ISO-4037, the X-ray source tube voltage (kV) and filter (radiative material) thickness need to be adjusted to regulate the radiation spectrum and match the requirements of different application scenarios (such as soft X-rays for mammography and high-energy X-rays for industrial CT). In other words, each tube voltage requires a corresponding filter (radiative material) for filtering.

[0003] Currently, the existing process for switching filter (radiative material) is as follows: 1. Set the tube voltage; 2. Pre-irradiate the required filter (radiative material) on the electric filter turntable, and rotate the filter turntable to select the appropriate filter (radiative material); 3. Confirm that the center of the selected filter (radiative material) coincides with the beam center of the radiation and corresponds to the set voltage.

[0004] While the above-mentioned methods for selecting and switching filters can solve the problem of filter (radiative material) switching to some extent, human intervention and determination are still required in the actual implementation process. The entire process of switching the radioactive material takes at least 5-6 minutes, which greatly increases the time cost of instrument detection. In addition, X-rays are generated during the exposure process, and people cannot enter the room to monitor in real time. If the position of the electric filter turntable changes during the filter switching process, it will affect the detection accuracy of the instrument, and thus affect the accuracy of subsequent detection results. Summary of the Invention

[0005] The purpose of this invention is to solve the above problems by designing an automatic radiometric selection system based on an electrically driven rotating filter disk.

[0006] To achieve the above objectives, the technical solution of this invention is an automatic radiation quality selection system based on an electrically driven rotating filter disk, comprising an electrically driven filter disk and a control system. The electrically driven filter disk is arranged with several filter plates, four of which are distributed at 90° intervals around the center of the disk. The positions of the four light-transmitting holes on each filter plate match their positions on the electrically driven filter disk. The extension line connecting the centers of the two light-transmitting holes at both ends of the filter plate must be on the center of the electrically driven filter disk to ensure accurate stopping when it reaches a position coinciding with the center of the radiation beam. The control system can control the electrically driven filter disk to rotate or stop via a servo motor. The control system selects the required filter plate through a filter plate identification unit. The filter plates on the electrically driven filter disk can be disassembled and replaced, but each barcode corresponds to one filter plate, maintaining a one-to-one correspondence. The control system positions the selected filter plate using a laser positioning unit.

[0007] The filter identification mechanism includes an industrial barcode scanner and several barcodes, each barcode corresponding one-to-one with a specific filter. The barcodes are positioned on one side of the electric filter turntable, close to their corresponding filter. The industrial barcode scanner is located on one side of the electric filter turntable. When the electric filter turntable rotates, the industrial barcode scanner scans each barcode sequentially. The control system uses a barcode recognition and matching algorithm to identify and match the barcodes scanned by the industrial barcode scanner and select the desired filter.

[0008] The barcode recognition and matching algorithm used in the control system includes the following steps:

[0009] Step 1, Preprocessing: Median filtering is used to remove noise from the barcode image, and adaptive threshold binarization is used to convert the image into a black and white pixel matrix to enhance the contrast of the barcode edges;

[0010] Step 2, Feature Extraction: Based on Hough transform, straight lines are detected at the barcode edges to locate the barcode region. A Convolutional Neural Network (CNN) is then used to identify the barcode type. Its core layer structure includes: 1. Input Layer: A pre-processed binary barcode image of uniform size. 2. Convolutional Layer: A sliding convolutional kernel extracts local features from the input image, such as edges, width variations of bars and spaces, and texture. 3. Activation Layer: A non-linear transformation is introduced after the convolutional layer. The ReLU activation function is applied to solve the gradient vanishing problem, accelerating training. 4. Pooling Layer: The binary image is reduced in dimensionality (size and parameters) while retaining key features, enhancing the network's robustness to small image displacements. 5. Fully Connected Layer: The extracted high-dimensional binary feature map is "flattened" into a one-dimensional vector and mapped to the class space through a weight matrix, achieving feature-to-class conversion. The extracted feature mapping is then transformed into the network's final output, the classification label. The process of identifying barcode types is as follows: "convolution to extract local features → pooling to reduce dimensionality and enhance robustness → fully connected layer to integrate abstract features → output layer classification". Combined with the barcode's bar and space rules (start character, encoding unit, etc.), it can accurately identify different barcode types.

[0011] Step 3, Decoding and Verification: The data is parsed according to the barcode type, and the accuracy of the data is verified using a CRC checksum to ensure that the filter information matches the instructions. The verification principle is as follows: CRC achieves efficient verification of data accuracy with low computational cost through the process of "generating a polynomial → calculating the checksum → verifying the remainder". Its core is modulo-2 division, and the key is that the sending and receiving ends use the same generating polynomial.

[0012] Step 4, Optimization Strategy: When X-ray interference exists in the scanning environment, automatically activate multi-frame image fusion technology to reduce noise impact and improve recognition success rate (≥99.5%). The core principle of image fusion for noise reduction is to utilize the "signal consistency" and "noise randomness" in multiple noisy images of the same scene: the same target has stable signal characteristics in multiple images, while noise (such as Gaussian noise, salt-and-pepper noise, etc.) is randomly distributed in different images. By fusing these images, common effective signals can be preserved, and random noise can be canceled, thereby achieving noise reduction. The process can be divided into the following key steps: acquiring relevant noisy images → registration and alignment → designing fusion rules (spatial domain / transform domain) → post-processing optimization.

[0013] After selecting the required filter, the control system activates the laser positioning unit to position and brake the electric filter turntable, ensuring that the center of the required filter coincides with the center of the X-ray.

[0014] The laser positioning unit includes a cross-shaped laser emitter, a dot-shaped laser receiving module, and several light-transmitting holes. The cross-shaped laser emitter is located on one side of the electric filter turntable, and the dot-shaped laser receiving module is located on the other side of the electric filter turntable. The several light-transmitting holes are arranged on the same circumference around the center of the electric filter turntable. Under the control of the control system, the cross-shaped laser emitter can emit a laser beam towards the electric filter turntable. When the laser beam shines on the dot-shaped laser receiving module through the several light-transmitting holes, the dot-shaped laser receiving module sends a signal to the control system. The control system determines the positioning of the filter according to the laser positioning and alignment algorithm and controls the servo motor to stop.

[0015] The laser positioning and alignment algorithm includes the following steps:

[0016] Step a, threshold judgment: When all four point laser receiving modules output high level at the same time (i.e., all four point laser receiving modules are irradiated by laser), it is judged as "center aligned" state.

[0017] Step b, time window filtering: To avoid instantaneous signal fluctuations caused by mechanical vibration, a 20ms time window is set. The braking command is only triggered if the four signals remain stable within the time window to prevent malfunctions.

[0018] Step c, establish the deviation compensation model: When some point-shaped laser receiving modules cannot receive laser irradiation, the filter offset Δx is calculated using a triangulation algorithm, as shown in the following formula:

[0019]

[0020] Wherein, I1-I4 are the light intensity values ​​of the four modules, k is the calibration coefficient, and D is the preset reference distance. The calculation results are used to drive the servo motor to fine-tune the position (accuracy up to 0.05mm).

[0021] The control system employs a three-stage speed control scheme for the servo motor, which is as follows:

[0022] Acceleration Phase: An S-curve acceleration is employed to avoid mechanical shock during startup. The formula is as follows:

[0023]

[0024] Constant speed phase: Maintain maximum speed V max =120° / s, with position error fed back in real time via encoder;

[0025] Deceleration Phase: Predictive braking model is used to calculate braking advance, based on PID control (proportional coefficient K). p =0.6, integral coefficient K i =0.02, differential coefficient K d =0.1), dynamically adjust the braking force to ensure that the stopping position error is ≤0.1mm.

[0026] Its control system flowchart is as follows Figure 3 As shown.

[0027] The predictive braking model calculates the braking advance based on motor speed, load inertia, and laser alignment signal transmission delay (approximately 5ms). The formula is as follows:

[0028] θ pre =v current ×(t delay +t brake )

[0029] Where t brake The optimal parameters for the braking system response time are obtained by fitting historical data.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. This invention can effectively save radiometric selection time. Through a fully automatic control system, the automatic selection and alignment of radiometrics can be completed within 1 minute, which is 5 times more efficient than the previous radiometric selection system.

[0032] 2. This invention can perform high-precision positioning of the filter during the switching process. By utilizing the principle that light travels in a straight line, the error in the concentricity between the filter (radiative material) and the X-ray source outlet can be reduced to less than 0.1 mm.

[0033] 3. This invention can effectively control human error. Through a standardized and fully automated switching process, it can effectively reduce manual debugging time and minimize human operation, thereby greatly improving the accuracy of automatic selection of radiation quality and detection by ionizing radiation instruments. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the structure of an automatic radiation quality selection system based on an electric rotating filter disk according to the present invention;

[0035] Figure 2 This is a flowchart of the automatic radiation quality selection system based on an electric rotating filter disk according to the present invention.

[0036] Figure 3 This is a flowchart of the control system described in this invention;

[0037] In the diagram, 1. Electric filter turntable; 2. Filter sheet; 3. Industrial barcode scanner; 4. Cross laser emitter; 5. Dot laser receiver module; 6. Light transmission hole; 7. Barcode. Detailed Implementation

[0038] The present invention will now be described in detail with reference to the accompanying drawings, such as... Figure 1-3 As shown;

[0039] An automatic radiometric selection system based on an electrically driven rotating filter disk includes an electrically driven filter disk 1 (servo motor driven), an industrial barcode scanner 3, a filter 2 (radiative material) with a barcode 7 and four 1mm light-transmitting holes 6 at its edge, a cross-shaped laser emitter 4, a dot-shaped laser receiver module 5, and a matching control system. Its layout diagram is shown below. Figure 1 As shown, on the far left are an industrial barcode scanner and a cross-shaped laser emitter. The industrial barcode scanner is placed above the cross-shaped laser emitter, so that the motorized filter turntable needs to scan the barcode to identify the filter (radiative material) information before aligning the filter (radiative material) with the center of the X-ray beam exit. When the motorized filter rotates to the appropriate position, the laser emitted by the cross-shaped laser can pass through the 1m light-transmitting hole on the filter (radiative material) to reach the point-shaped laser receiving module.

[0040] The specific principle is as follows: 1. Power on the entire system and send the required filter (radiative material) information to the control system. The system will convert the sent command into the corresponding barcode on the filter. 2. The electric filter turntable driven by the servo motor starts to rotate. The industrial barcode scanner begins to scan the barcodes on the filter (radiative material) installed on the filter turntable until it finds the required filter (radiative material) sent to the control system in step 1. 3. After finding the filter (radiative material), the electric filter turntable will not stop rotating, but will continue to rotate until the light emitted by the cross laser emitter passes through the small holes on the edge of the filter (radiative material) and reaches the point laser receiving module. 4. Since the center of the line connecting the four small holes is the center of the filter (radiative material), when all four point laser receiving modules receive the laser, it means that the center of the filter (radiative material) and the X-ray has coincided. At this time, the point laser receiving module will output a high level. 5. The high-level signal output by the point-shaped laser receiver module is transmitted to the control system. The control system then activates the braking system of the electric filter turntable servo motor, causing the electric filter turntable to stop rotating. The stopping position is precisely where the filter (radiative material) is concentric with the center of the X-ray source beam. If a different radioactive material needs to be selected, the system will repeat steps 1-5. The system operation principle diagram is shown below. Figure 2 As shown.

[0041] Core algorithm logic of the control system:

[0042] 1. Barcode recognition and matching algorithm

[0043] Image processing workflow: After the industrial barcode scanner acquires the barcode image of the filter, it processes it through the following steps:

[0044] Preprocessing: Median filtering is used to remove noise, and adaptive threshold binarization is used to convert the image into a black and white pixel matrix to enhance the edge contrast of the barcode.

[0045] Feature extraction: Detect straight lines at barcode edges using Hough transform to locate barcode regions; identify barcode type using a convolutional neural network (CNN).

[0046] Decoding and Verification: Parse the data according to the barcode type, verify the accuracy of the data through CRC check code, and ensure that the filter information is consistent with the instructions.

[0047] Optimization strategy: When there is X-ray interference in the scanning environment, the algorithm automatically enables multi-frame image fusion technology to reduce the impact of noise and improve the recognition success rate (≥99.5%).

[0048] 2. Laser positioning and alignment algorithm:

[0049] Signal synchronization processing: The electrical signal output by the point-shaped laser receiver module is processed through the following logic:

[0050] Threshold discrimination: Set a light intensity threshold (e.g., 0.8V). When all four modules output a high level (≥ threshold) at the same time, it is determined to be in the "center aligned" state.

[0051] Time window filtering: To avoid instantaneous signal fluctuations caused by mechanical vibration, a 20ms time window is set. The braking command is only triggered if the four signals remain stable within the time window to prevent malfunctions.

[0052] Deviation Compensation Model: If only some modules receive laser light, the system calculates the filter offset using a triangulation algorithm, as shown in the following formula:

[0053]

[0054] Wherein, I1-I4 are the light intensity values ​​of the four modules, k is the calibration coefficient, and D is the preset reference distance. The calculation results are used to drive the servo motor to fine-tune the position (accuracy up to 0.05mm).

[0055] 3. Servo motor motion control algorithm:

[0056] Three-stage speed planning:

[0057] Acceleration Phase: An S-curve acceleration is employed to avoid mechanical shock during startup. The formula is as follows:

[0058]

[0059] Uniform speed phase: Maintain maximum speed Vmax = 120° / s, and use the encoder to provide real-time feedback on position error.

[0060] Deceleration Phase: Predictive braking model is used to calculate braking advance, based on PID control (proportional coefficient K). p =0.6, integral coefficient K i =0.02, differential coefficient K d =0.1), dynamically adjust the braking force to ensure that the stopping position error is ≤0.1mm.

[0061] Predictive braking model: Based on motor speed, load inertia, and laser alignment signal transmission delay (approximately 5ms), the braking advance is calculated in advance using the following formula:

[0062] θ pre =v current ×(t delay +t brake )

[0063] Where t brake To determine the braking system response time, optimal parameters were obtained by fitting historical data. The control system flowchart is shown below. Figure 3 As shown.

[0064] It should be noted that the main inventive points of this invention include:

[0065] 1. A multimodal control algorithm integrating barcode recognition, laser synchronous detection, and predictive braking is used to achieve a positioning accuracy of ≤0.1mm;

[0066] 2. By utilizing the real-time coupling of servo motor PID control parameters and laser positioning signals, millisecond-level synchronization between mechanical actions and electrical signal processing is ensured;

[0067] 3. An anti-interference mechanism based on time window filtering and deviation compensation is used to address signal stability issues under X-ray environments. To address the potential impact of scattered light on the stability of laser positioning signals, a median filtering method is employed to sort signals within a time window, taking the median value as the output. A corresponding threshold range is set, and a sliding time window is used for real-time filtering. Values ​​exceeding or falling below the threshold range are considered outliers and dynamically removed, while valid signals are retained.

[0068] 4. An LSTM neural network deep learning model was introduced. This model was trained using historical operating data (filter weight, motor current, alignment time, etc.) to continuously and dynamically optimize the system. This includes:

[0069] a. Scanning path planning: Based on the probability distribution of filter elements, the optimal scanning sequence is predicted. Specifically, historical data can be used to quickly locate the position where the filter element needs to be selected, thereby determining the rotation direction of the electric filter turntable. Furthermore, by adjusting the rotation speed, the scanning time can be reduced by an average of 30%.

[0070] b. Adaptive braking parameters: Since the weight of each filter element is different, the moment of inertia of the electric filter disc varies at different positions. The PID coefficient is adjusted in real time to adapt to the difference in moment of inertia caused by the weight of different filter elements. The braking model is continuously optimized through historical data to improve the repeatability accuracy to 0.08mm.

[0071] The above technical solutions only embody the preferred technical solutions of the present invention. Any modifications that may be made by those skilled in the art to certain parts thereof embody the principles of the present invention and fall within the protection scope of the present invention.

Claims

1. An automatic radiation quality selection system based on electric rotating filter disc, comprising an electric filter rotating disc (1) and a control system, a plurality of filter pieces (2) are arranged on the electric filter rotating disc, the plurality of filter pieces are distributed at equal angles around the center of the electric filter rotating disc, the control system selects the required filter piece through a filter piece identification unit during the rotation of the electric filter rotating disc, and then positions the selected filter piece through a laser positioning unit.

2. A system for automatic selection of the quality of radiation based on an electrically driven rotating filter disc according to claim 1, characterized in that The filter piece identification mechanism comprises an industrial barcode scanner (3) and a plurality of barcodes (7), the plurality of barcodes are in one-to-one correspondence with the plurality of filter pieces, the barcodes are arranged on one side surface of the electric filter rotating disc, the barcodes are close to the filter pieces corresponding thereto, the industrial barcode scanner is located on one side of the electric filter rotating disc, when the electric filter rotating disc rotates, the industrial barcode scanner can scan the barcodes one by one, the control system can identify and match the barcodes scanned by the industrial barcode scanner through a barcode identification and matching algorithm, and select the required filter piece.

3. The automatic selection of the quality of radiation based on the electrically driven rotating filter disc system according to claim 1, characterized in that, The barcode identification and matching algorithm adopted by the control system comprises the following steps: Step one, preprocessing: removing noise of the barcode image by using median filtering, converting the image to a black and white pixel matrix by adaptive threshold binaryzation, and enhancing the edge contrast of the barcode; Step two, feature extraction: detecting barcode edge straight lines based on Hough transform, positioning the barcode area, and identifying the barcode type through a convolutional neural network; Step three, decoding and checking: analyzing data according to the barcode type, verifying data accuracy through CRC check code, and ensuring that the filter piece information is consistent with the instruction; Step four, optimization strategy: when there is X-ray interference in the scanning environment, automatically enabling a multi-frame image fusion technology to reduce the influence of noise points and improve the identification success rate.

4. The automatic selection of the quality of radiation based on the electrically driven rotating filter disc system according to claim 1, characterized in that, After selecting the required filter piece, the control system starts the laser positioning unit to brake the electric filter rotating disc, so as to ensure that the center of the required filter piece coincides with the center of the X-ray.

5. The automatic selection of the quality of radiation based on the electrically driven rotating filter disc system according to claim 1, characterized in that, The laser positioning unit comprises a cross laser emitter (4), a point-shaped laser receiving module (5) and a plurality of light transmission holes (6), the cross laser emitter is located on one side of the electric filter rotating disc, the point-shaped laser receiving module is located on the other side of the electric filter rotating disc, and the plurality of light transmission holes are arranged on the same circle around the center of the electric filter rotating disc, the cross laser emitter can emit a laser beam to the electric filter rotating disc under the control of the control system, when the laser beam irradiates to the point-shaped laser receiving module through the plurality of light transmission holes, the point-shaped laser receiving module sends a signal to the control system, the control system determines the positioning of the filter piece according to a laser positioning and alignment algorithm, and controls a servo motor to stop moving.

6. An automatic selection of radiation quality system based on electrically rotating filter disc according to claim 1, characterized in that, The laser positioning and alignment algorithm comprises the following steps: Step a, threshold judgment: when four point-shaped laser receiving modules output high level at the same time (that is, the four point-shaped laser receiving modules are irradiated by laser), it is judged as "center alignment" state; Step b, time window filtering: to avoid the instantaneous signal fluctuation caused by mechanical vibration, set a 20ms time window, and require four signals to be stable within the time window to trigger the brake command, to prevent false action; Step c, establish deviation compensation model: when part of the point-shaped laser receiving module cannot receive laser irradiation, calculate the filter sheet offset Δx through the triangular positioning algorithm, the formula is as follows: Where I1-I4 is the light intensity value of the four modules, k is the calibration coefficient, and D is the preset reference distance. The calculation result is used to drive the servo motor to fine-tune the position.

7. The automatic selection of the quality of radiation based on the electrically driven rotating filter disc system according to claim 1, characterized in that, The control system adopts a three-stage speed control plan for the servo motor, which is: Acceleration stage: S-shaped curve acceleration is adopted to avoid mechanical impact during startup, and the formula is: Constant speed stage: keep the maximum speed Vmax=120° / s, and feedback the position error in real time through the encoder; Deceleration stage: use the predictive brake model to calculate the brake advance, based on PID control, dynamically adjust the brake strength, and ensure that the stop position error is ≤0.1mm.

8. A system for automatic selection of the quality of radiation based on an electrically driven rotating filter disc according to claim 7, characterized in that The predictive brake model is to calculate the brake advance in advance according to the motor speed, load inertia and laser alignment signal transmission delay, and the formula is: θ pre = v current × (t delay + t brake ) where t brake is the brake system response time, the optimal parameters are obtained by fitting historical data.

Images