Calibration method for cargo density detection module

Abstract

The invention provides a calibration method for a cargo density detection module, and relates to the technical field of test calibration, and the method comprises the steps: 1, detecting the initial ray intensity, and carrying out the static calibration of a ray emitter based on a detection result; 2, testing the samples with different densities one by one based on the ray emitter subjected to static calibration, and dynamically calibrating sensor parameters according to a one-by-one test result; 3, after dynamic calibration is completed, historical working parameters of the cargo density detection module in the target time period are called; 4, analyzing the historical working parameters, judging whether the cargo density detection module has detection vulnerabilities or not, and recalibrating the cargo density detection module when the detection vulnerabilities exist; and the stable operation, the working efficiency and the working accuracy of the cargo density detection module are effectively guaranteed.

Description

Technical Field

[0001] This invention relates to the field of testing and calibration technology, and in particular to a calibration method for a cargo density detection module. Background Technology

[0002] Currently, in modern logistics, warehousing, and many industrial production and transportation processes, accurately understanding the density of goods is of paramount importance. Cargo density information plays a decisive role in rationally planning storage space, ensuring transportation safety (such as preventing damage to transport vehicles due to overweight), optimizing logistics costs, and selecting appropriate packaging and transportation methods based on density.

[0003] However, existing cargo density detection modules often suffer from insufficient calibration accuracy in practical applications, resulting in low accuracy of density detection and thus failing to guarantee the successful detection of cargo density. Furthermore, existing calibration is often based on pre-detection calibration rather than updating the calibration status in real time based on historical data, which also fails to guarantee the accuracy and effectiveness of cargo density detection.

[0004] Therefore, in order to overcome the above-mentioned technical problems, the present invention provides a calibration method for a cargo density detection module. Summary of the Invention

[0005] This invention provides a calibration method for a cargo density detection module. By detecting the initial radiation intensity, the method effectively performs static calibration of the radiation emitter based on the detection results, laying a solid foundation for cargo density detection and ensuring successful detection. After static calibration, by testing samples of different densities one by one, the sensor parameters can be dynamically calibrated based on the test results, ensuring the sensor's sensitivity and the accuracy of the detection results. This provides a good detection environment for subsequent cargo density detection, ensuring detection efficiency and accuracy. By collecting and analyzing historical operating parameters within a target time period, the cargo density detection module is recalibrated when detection loopholes are found, effectively ensuring the stable operation, efficiency, and accuracy of the cargo density detection module.

[0006] A calibration method for a cargo density detection module, comprising:

[0007] Step 1: Detect the initial radiation intensity and perform static calibration on the radiation emitter based on the detection results;

[0008] Step 2: Test samples of different densities one by one based on the statically calibrated X-ray emitter, and dynamically calibrate the sensor parameters according to the test results.

[0009] Step 3: After dynamic calibration is completed, retrieve the historical operating parameters of the cargo density detection module within the target time period;

[0010] Step 4: Analyze historical operating parameters to determine if there are any detection vulnerabilities in the cargo density detection module, and if so, recalibrate the cargo density detection module.

[0011] Preferably, a calibration method for a cargo density detection module includes, in step 1, detecting the initial radiation intensity and performing static calibration of the radiation emitter based on the detection results, comprising:

[0012] Record the initial radiation intensity of the radiation emitter multiple times under no-load conditions, analyze the initial radiation intensity, and determine whether static calibration of the radiation emitter is required.

[0013] When it is determined that static calibration of the X-ray emitter is required, the adjustment parameters of the X-ray emitter are analyzed, and static calibration of the X-ray emitter is performed based on the analysis results.

[0014] Preferably, a calibration method for a cargo density detection module involves recording the initial radiation intensity emitted by the radiation emitter multiple times under no-load conditions, analyzing the initial radiation intensity to determine whether static calibration of the radiation emitter is required, including:

[0015] Record the initial radiation intensity of the radiation emitter multiple times under no-load conditions, and calculate the average target intensity of the initial radiation intensity of the multiple emissions.

[0016] Based on the attribute information of the radiation emitter, the reference intensity value of the radiation emitter is matched in the preset management library; the average target intensity value is compared with the reference intensity value to determine whether static calibration of the radiation emitter is required;

[0017] If the average target intensity is consistent with the reference intensity value, then it is determined that no static calibration of the ray emitter is required; otherwise, it is determined that static calibration of the ray emitter is required.

[0018] Preferably, a calibration method for a cargo density detection module, when it is determined that static calibration of the X-ray emitter is required, analyzes the adjustment parameters of the X-ray emitter, and performs static calibration of the X-ray emitter based on the analysis results, including:

[0019] When it is determined that static calibration of the X-ray emitter is required, the adjustable parameters in the X-ray emitter are obtained, and a parameter group corresponding to each adjustable parameter is constructed.

[0020] Each adjustable parameter is adjusted according to its corresponding parameter group, and the target rate of change of radiation intensity under no-load conditions is recorded based on the adjustment results.

[0021] Compare the target rate of change with a preset rate of change threshold;

[0022] When the target rate of change is greater than or equal to the preset rate of change threshold, the corresponding adjustable parameter will be used as the influencing parameter.

[0023] Otherwise, the corresponding adjustable parameter will be treated as a non-affecting parameter;

[0024] Retrieve the radiation intensity values ​​corresponding to each parameter in the parameter group that affects the parameters;

[0025] Based on the correspondence between each parameter in the parameter group corresponding to the influencing parameters and the radiation intensity value, determine the target correlation between the influencing parameters and the radiation intensity;

[0026] Adjust the parameters of the influencing parameters according to the target correlation until the output radiation intensity reaches the reference intensity value, thus completing the static calibration of the radiation emitter.

[0027] Preferably, in a calibration method for a cargo density detection module, step 2 involves testing samples of different densities one by one based on a statically calibrated X-ray emitter, and dynamically calibrating the sensor parameters according to the test results, including:

[0028] Based on the statically calibrated X-ray emitter, X-rays are emitted sequentially to samples of different densities, and the sensor is controlled to collect the attenuation signals of the X-rays corresponding to samples of different densities in sequence.

[0029] Obtain the background script file of the sensor, and extract the execution data of the sensor when collecting the attenuation signal of the ray from the background script file;

[0030] Based on the execution data, determine the timing parameters and automatic capture range for the sensor to capture the attenuation signal of the ray, and determine the emission timing and emission range of the ray based on the ray emitter;

[0031] The sensitivity of the sensor is determined by the timing parameters of the attenuation signal of the X-ray captured by the sensor and the automatic capture range, as well as by the X-ray emitter to determine the emission timing and emission range of the X-ray.

[0032] Simultaneously, the output results of the sensor's analysis of the attenuation signal of the X-rays are obtained, and the density detection value of the sensor for each known density sample is determined based on the output results;

[0033] The density detection accuracy of the sensor is determined based on the density detection value and the corresponding known density.

[0034] The system acquires the sensor's performance requirements based on the management terminal and determines the deviation between these requirements and the sensor's sensitivity and density detection accuracy.

[0035] The calibration direction and calibration value in the calibration direction are determined based on the deviation value to improve the sensitivity and density detection accuracy of the sensor.

[0036] The sensor parameters are dynamically calibrated based on the calibration direction and the calibration value in the calibration direction.

[0037] Preferably, a calibration method for a cargo density detection module dynamically calibrates sensor parameters based on the calibration direction and calibration values ​​in the calibration direction, including:

[0038] Obtain the dynamic calibration results of the sensor parameters, and re-examine the performance of the sensor parameters after completing the dynamic calibration;

[0039] After the performance retest is passed, the dynamic calibration information of the sensor parameters is recorded throughout the entire process, and a dynamic calibration report is generated based on the full process record.

[0040] The dynamic calibration report is fed back to the management terminal for recording.

[0041] Preferably, in a calibration method for a cargo density detection module, step 3 involves retrieving historical operating parameters of the cargo density detection module within a target time period after dynamic calibration is completed, including:

[0042] Read the start and end times of the target time period;

[0043] Generate the first index label for the work parameter management library based on the start and end times;

[0044] Obtain the module type of the cargo density detection module, and generate a second index label for the working parameter management library based on the module type;

[0045] Based on the first and second index tags, retrieve the historical operating parameters of the density detection module within the target time period from the operating parameter management library.

[0046] Preferably, in a calibration method for a cargo density detection module, step 4 involves analyzing historical operating parameters to determine if the cargo density detection module has any detection vulnerabilities. If vulnerabilities are found, the cargo density detection module is recalibrated, including:

[0047] Obtain the obtained historical working parameters, read the historical working parameters, and determine the business composition of the historical working parameters;

[0048] Based on the business composition, historical operating parameters are divided into equipment operating parameters and density detection feedback parameters. The equipment operating parameters are then analyzed to determine the fluctuation range of the equipment's own performance.

[0049] When the fluctuation range of the device's own performance meets the baseline operating conditions, it is determined that the device's own performance has no defects; otherwise, it is determined that the device's own performance has a first defect.

[0050] At the same time, the density detection feedback parameters are analyzed to determine the detection deviation value of the sensor for the density parameters, and when the detection deviation value exceeds the preset deviation range, it is determined that there is a second vulnerability in the cargo density detection module.

[0051] The cargo density detection module was recalibrated based on the first and second vulnerabilities.

[0052] Preferably, a calibration method for a cargo density detection module, which recalibrates the cargo density detection module based on a first vulnerability and a second vulnerability, includes:

[0053] The device operating parameters corresponding to the first vulnerability and the density detection feedback parameters corresponding to the second vulnerability are obtained respectively. The device operating parameters and density detection feedback parameters are analyzed respectively to determine the corresponding abnormal parameter fragments.

[0054] The equipment components in the cargo density detection module are traced based on abnormal parameter fragments, and the target components corresponding to recalibration are determined based on the traceability results.

[0055] Based on the vulnerability characteristics of the first and second vulnerabilities, a recalibration strategy for the target component is determined, and the cargo density detection module is recalibrated based on the recalibration strategy.

[0056] Preferably, a calibration method for a cargo density detection module further includes: Step 5, constructing an intelligent calibration model, the specific process of which is as follows:

[0057] Collect historical calibration events and read the calibration data corresponding to each historical calibration event. The calibration data includes: static calibration data, dynamic calibration data, and recalibration data.

[0058] The static calibration data, dynamic calibration data, and recalibration data of each historical calibration event are analyzed to determine the static trigger conditions and the corresponding static calibration dataset for each static calibration, the dynamic trigger conditions and the corresponding dynamic calibration dataset for each dynamic calibration, and the trigger conditions and the corresponding recalibration dataset for each recalibration.

[0059] A static calibration prediction network is constructed by learning the static triggering conditions;

[0060] Learn from dynamic triggering conditions to construct a dynamic calibration prediction network;

[0061] The triggering conditions during recalibration are learned to construct a recalibration prediction network;

[0062] The static calibration prediction network, dynamic calibration prediction network, and recalibration prediction network are integrated to construct the target prediction model;

[0063] The static decision network is determined by learning the static calibration dataset corresponding to the static triggering condition, and the dynamic decision network is determined by learning the dynamic calibration dataset corresponding to the dynamic triggering condition. At the same time, the recalibration decision network is determined by learning the recalibration dataset corresponding to the recalibration triggering condition.

[0064] A target decision-making model is constructed by integrating static decision networks, dynamic decision networks, and recalibration decision networks.

[0065] The target prediction model and the target decision model are associated according to the calibration type to build an intelligent calibration model. When a new calibration event occurs, the target prediction model in the intelligent calibration model is used to determine in real time whether the triggering condition has been met.

[0066] When the trigger condition is met, the corresponding calibration type is determined, and the corresponding calibration data is analyzed and matched in the target decision model according to the calibration type. At the same time, the current cargo density detection module is calibrated in real time according to the matching result.

[0067] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0068] By detecting the initial radiation intensity, the radiation emitter can be statically calibrated based on the detection results, laying a solid foundation for cargo density detection and ensuring successful cargo density detection. After static calibration, by testing samples of different densities one by one, the sensor parameters can be dynamically calibrated based on the test results, thereby ensuring the sensor's sensitivity and the accuracy of the detection results. This provides a good detection environment for subsequent cargo density detection, ensuring detection efficiency and accuracy. By collecting and analyzing historical working parameters within the target time period, the cargo density detection module can be recalibrated when detection loopholes are found, effectively ensuring the stable operation, efficiency, and accuracy of the cargo density detection module.

[0069] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0070] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0071] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0072] Figure 1 This is a flowchart of a calibration method for a cargo density detection module according to an embodiment of the present invention;

[0073] Figure 2 This is a flowchart of step 1 in a calibration method for a cargo density detection module according to an embodiment of the present invention;

[0074] Figure 3 This is a flowchart of step 3 in a calibration method for a cargo density detection module according to an embodiment of the present invention. Detailed Implementation

[0075] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0076] Example 1:

[0077] This embodiment provides a calibration method for a cargo density detection module, such as... Figure 1 As shown, it includes:

[0078] Step 1: Detect the initial radiation intensity and perform static calibration on the radiation emitter based on the detection results;

[0079] Step 2: Test samples of different densities one by one based on the statically calibrated X-ray emitter, and dynamically calibrate the sensor parameters according to the test results.

[0080] Step 3: After dynamic calibration is completed, retrieve the historical operating parameters of the cargo density detection module within the target time period;

[0081] Step 4: Analyze historical operating parameters to determine if there are any detection vulnerabilities in the cargo density detection module, and if so, recalibrate the cargo density detection module.

[0082] In this embodiment, the density detection module includes: a radiation emitter, a sensor, etc.

[0083] In this embodiment, dynamic calibration includes optimizing and adjusting sensor sensitivity, response time, and data processing algorithms to ensure detection accuracy and stability; through repeated verification, it ensures that each parameter matches the optimal state, thereby improving overall detection efficiency.

[0084] In this embodiment, the historical working parameters include the working data of the cargo density detection module itself (i.e., the detection results of density detection of cargo) and the feedback data of the user terminal on the detection results after detection.

[0085] In this embodiment, the detection vulnerability may include the component in the density detection module that has a vulnerability and the cause of the vulnerability.

[0086] In this embodiment, the target time period can be pre-set, such as three months, six months, etc.

[0087] The working principle and beneficial effects of the above technical solution are as follows: By detecting the initial radiation intensity, the radiation emitter can be statically calibrated based on the detection results, laying a solid foundation for the density detection of goods and ensuring the smooth detection of goods density. After static calibration, by testing samples of different densities one by one, the sensor parameters can be dynamically calibrated based on the test results, thereby ensuring the sensitivity of the sensor and the accuracy of the detection results. This provides a good detection environment for subsequent goods density detection, ensuring detection efficiency and accuracy. By collecting and analyzing historical working parameters within the target time period, the goods density detection module can be recalibrated when detection loopholes exist, effectively ensuring the stable operation, working efficiency, and working accuracy of the goods density detection module.

[0088] Example 2:

[0089] Based on Example 1, this example provides a calibration method for a cargo density detection module, such as... Figure 2 As shown, in step 1, the initial radiation intensity is detected, and the radiation emitter is statically calibrated based on the detection results, including:

[0090] S101: Record the initial radiation intensity of the radiation emitter multiple times under no-load conditions, analyze the initial radiation intensity, and determine whether static calibration of the radiation emitter is required.

[0091] S102: When it is determined that static calibration of the radiation emitter is required, the adjustment parameters of the radiation emitter are analyzed, and static calibration of the radiation emitter is performed based on the analysis results.

[0092] In this embodiment, an unloaded ray emitter refers to a state in which there are no target objects or loads around the ray emitter.

[0093] In this embodiment, the initial radiation intensity is the intensity value of the radiation emitted by the radiation emitter under no-load conditions; it is a basic performance indicator of the radiation emitter and can be used to measure the radiation emitter's emission capability under conditions without external interference; this value may be affected by factors such as the power supply inside the radiation emitter and the characteristics of the emission device (such as the activity of radioactive materials, the voltage and current of the X-ray generator, etc.).

[0094] In this embodiment, the adjustment parameters are internal parameters related to the calibration of the X-ray emitter; for example, for some X-ray emitters, these include tube voltage adjustment parameters (used to control the energy of the emitted X-rays), tube current adjustment parameters (affecting the intensity of the X-rays, etc.), pulse frequency adjustment parameters (if it is a pulsed X-ray emitter), etc.; these adjustment parameters can be adjusted to change the operating state of the X-ray emitter, thereby achieving static calibration and ensuring that the X-ray emitter can work normally and emit X-rays of accurate intensity when subsequently under load (such as when performing a detection task).

[0095] In this embodiment, static calibration refers to calibration performed on the radiation emitter when it is not in operation.

[0096] The working principle and beneficial effects of the above technical solution are as follows: by recording and analyzing the initial radiation intensity emitted by the radiation emitter multiple times under no-load conditions, the purpose of multiple times is to ensure the accuracy of data analysis and reduce analysis errors. When it is determined that static calibration is required, the adjustment parameters of the radiation emitter are analyzed, thereby effectively ensuring the timeliness and effectiveness of static calibration.

[0097] Example 3:

[0098] Based on Example 2, this example provides a calibration method for a cargo density detection module. Under no-load conditions, the initial radiation intensity emitted by the radiation emitter is recorded multiple times, and the initial radiation intensity is analyzed to determine whether static calibration of the radiation emitter is necessary. This includes:

[0099] Record the initial radiation intensity of the radiation emitter multiple times under no-load conditions, and calculate the average target intensity of the initial radiation intensity of the multiple emissions.

[0100] Based on the attribute information of the radiation emitter, the reference intensity value of the radiation emitter is matched in the preset management library; the average target intensity value is compared with the reference intensity value to determine whether static calibration of the radiation emitter is required;

[0101] If the average target intensity is consistent with the reference intensity value, then it is determined that no static calibration of the ray emitter is required; otherwise, it is determined that static calibration of the ray emitter is required.

[0102] In this embodiment, the initial radiation intensity recorded for each measurement is kept under the same measurement conditions, including ambient temperature and power supply stability.

[0103] In this embodiment, the detector is placed at a fixed distance from the ray emitter during measurement.

[0104] In this embodiment, the reference intensity value is preset by matching a reference intensity value consistent with the current ray emitter in a preset management library.

[0105] The working principle and beneficial effects of the above technical solution are as follows: by calculating the average target intensity of the initial radiation intensity of multiple emissions, the accuracy of the measurement is effectively guaranteed; by comparing the calculated average target intensity with the reference intensity value, it is possible to effectively determine whether static calibration is required, thus ensuring the timeliness of static calibration of the radiation emitter.

[0106] Example 4:

[0107] Based on Example 2, this example provides a calibration method for a cargo density detection module. When it is determined that static calibration of the X-ray emitter is required, the adjustment parameters of the X-ray emitter are analyzed, and static calibration of the X-ray emitter is performed based on the analysis results, including:

[0108] When it is determined that static calibration of the X-ray emitter is required, the adjustable parameters in the X-ray emitter are obtained, and a parameter group corresponding to each adjustable parameter is constructed.

[0109] Each adjustable parameter is adjusted according to its corresponding parameter group, and the target rate of change of radiation intensity under no-load conditions is recorded based on the adjustment results.

[0110] Compare the target rate of change with a preset rate of change threshold;

[0111] When the target rate of change is greater than or equal to the preset rate of change threshold, the corresponding adjustable parameter will be used as the influencing parameter.

[0112] Otherwise, the corresponding adjustable parameter will be treated as a non-affecting parameter;

[0113] Retrieve the radiation intensity values ​​corresponding to each parameter in the parameter group that affects the parameters;

[0114] Based on the correspondence between each parameter in the parameter group corresponding to the influencing parameters and the radiation intensity value, determine the target correlation between the influencing parameters and the radiation intensity;

[0115] Adjust the parameters of the influencing parameters according to the target correlation until the output radiation intensity reaches the reference intensity value, thus completing the static calibration of the radiation emitter.

[0116] In this embodiment, adjustable parameters refer to the adjustable parameter information involved in the static calibration of the ray emitter.

[0117] In this embodiment, the parameter group refers to a pre-designed "test sequence" for all adjustable parameters of the radiation emitter. Its core function is to accurately determine the degree of influence of the parameter on the radiation intensity by adjusting the parameter in an orderly manner. The definitions of "adjustment range" and "step size" are only related to the physical characteristics of the parameter itself, equipment limitations, and test requirements, and are unrelated to whether the parameter is ultimately an influencing parameter. Both "adjustment range" and "step size" are pre-set based on experimental conditions, etc. The adjustment range refers to the boundary of the adjustable parameter within a safe and effective range. For example, the distance between the radiation source and the detector (focal length). If the distance is too short (<500mm), it will cause serious radiation scattering interference. If it is too long (>2000mm), it will cause excessive intensity attenuation and have no practical application value. Therefore, the adjustment range is set to 500mm~2000mm. The step size refers to the fixed increment value when adjusting the parameter each time within the adjustment range. It is used to control the resolution and efficiency of the test. The smaller the step size, the more refined the test (capturing minute effects), but the more tests are required. The larger the step size, the more efficient the test, but it may miss minute effects. The step size is mainly based on the potential sensitivity of the parameter to the radiation intensity. For example, if the adjustment range of the distance (focal length) between the radiation source and the detector is 500mm to 2000mm, then according to the attenuation rate of the radiation intensity and actual experimental measurements, the optimal step size is 100mm.

[0118] In this embodiment, the adjustable parameters in the ray emitter include, but are not limited to, tube voltage, tube current, housing temperature, focal length, etc.

[0119] In this embodiment, the influence of multiple parameters on radiation intensity may vary (this can be visually judged by the magnitude of the "target change rate": the larger the target change rate, the more significant the influence of the parameter on radiation intensity). Adjustments are typically made in descending order of influence. The reason is that parameters with a greater influence are more efficient to adjust, quickly reducing the deviation between the actual intensity and the reference value, thus simplifying subsequent adjustments. For example, if the target change rate of tube voltage (U) is 40% (each 10kV adjustment corresponds to a 40% change in radiation intensity), and the target change rate of tube current (I) is 20% (each 1mA adjustment corresponds to a 20% change in radiation intensity), then the tube voltage (with a greater influence) should be adjusted first, followed by the tube current (with a relatively smaller influence).

[0120] In this embodiment, the target correlation between the influencing parameters and the radiation intensity value is determined based on the correspondence between each parameter in the parameter group corresponding to the influencing parameters. The parameters of the influencing parameters are adjusted according to the target correlation until the output radiation intensity reaches the reference intensity value, thus completing the static calibration of the radiation emitter. The adjustment method is determined based on the target correlation, which refers to the specific correspondence between a single influencing parameter and the radiation intensity (such as "radiation intensity is proportional to the square of the tube voltage" or "linearly proportional to the tube current"). The specific process is as follows: First, the parameter with the greatest influence is used to reduce the deviation. For example, if the reference intensity value is 1000 and the current intensity value is 500, then the deviation is 500. The target correlation for the tube voltage (u) is: Intensity = 0.01 * u 2 (That is, when the current tube voltage u = 200kV, the intensity is 400; while when u = 224kV, the intensity is approximately 500). Based on this, the tube voltage is first adjusted. According to the target correlation, the current tube voltage u = 200kV needs to be adjusted to 224kV, thereby increasing the intensity from 400 to 500 (reducing the deviation to 500). Next, other influencing parameters are used to continue approximating the reference value. For example, the target correlation for the tube current (I) is "intensity = 100 × I" (that is, when I = 5mA, the intensity = 500; when I = 10mA, the intensity = 1000). With the tube voltage already adjusted to 224kV (intensity 500), the tube current is adjusted from 5mA to 10mA to achieve an intensity of 1000, completing the calibration.

[0121] In this embodiment, if there is an interaction between multiple influencing parameters, that is, adjusting one parameter will change the correlation between another parameter and intensity (e.g., "intensity = k × U × I", where the effects of U and I are interrelated), then an "iterative adjustment" strategy is adopted: first, fix the other influencing parameters and adjust only the first influencing parameter to make the intensity as close as possible to the reference value (record the parameter value at this time); keep the first parameter unchanged and adjust the second influencing parameter to further reduce the deviation; if the deviation still does not meet the standard, repeat the above steps (fine-tune the first parameter and then adjust the second parameter) until the radiation intensity reaches the reference value (deviation ≤ allowable range, such as ±1%).

[0122] In this embodiment, the preset change rate threshold is pre-set to determine whether an adjustable parameter is an influencing parameter. The method used to set the preset change rate threshold is "experimental verification + scenario adaptation," with the core being finding the "critical value between the rate of change of radiation intensity and the density detection error." The specific steps are as follows: First, determine the maximum allowable density detection error (e.g., ±3%), then test the "relationship between the rate of change of radiation intensity and the density error" experimentally: using a standard cargo with a known density (e.g., density 1.2 g / cm³). 3 Standard wooden block, 7.8g / cm³ 3The test was conducted using a standard iron block. Other parameters were fixed, and only the initial intensity I of the radiation was adjusted. The density detection results were recorded under different rates of change of I. For example, if the experiment found that when the rate of change of I was ≤4%, the density detection error was always ≤±3%; and when the rate of change was >4%, the error exceeded ±3%, then the threshold was set to 4% (ensuring that all intensity fluctuations that could cause density errors to exceed the limit could be covered by parameter adjustment). Secondly, corrections were made based on the characteristics of the goods. For example, different goods (such as cardboard boxes, plastics, and metals) have different attenuation characteristics for radiation, so the threshold must be effective for "intensity fluctuations when radiation interacts with different materials." Finally, the density detection of goods needs to balance "accuracy" and "speed." Since core aspects such as transportation safety and storage of high-value goods are involved, accuracy should be prioritized (e.g., setting the preset rate of change threshold to 2%), even at the cost of speed, to avoid safety accidents or high losses due to density misjudgment. For batch processing of light, low-value goods, the accuracy could be appropriately relaxed (e.g., setting the preset rate of change threshold to 5%), increasing speed to reduce logistics turnover costs and achieving "maximum efficiency without affecting core objectives."

[0123] In this embodiment, the influencing parameter refers to the parameter that can cause a significant change in radiation intensity. This is a core parameter that needs to be carefully handled and controlled during the calibration process.

[0124] In this embodiment, non-influencing parameters refer to parameters that have a negligible impact on radiation intensity. They can be temporarily ignored during calibration.

[0125] In this embodiment, the reference intensity value refers to a predefined standard output intensity value that the expected ray emitter will reach under no-load conditions.

[0126] The working principle and beneficial effects of the above technical solution are as follows: When it is determined that static calibration of the X-ray emitter is required, by acquiring the adjustable parameters in the X-ray emitter and constructing a parameter group corresponding to each adjustable parameter, the target rate of change of X-ray intensity under no-load conditions can be effectively recorded. By comparing it with a preset rate of change threshold, the influencing parameters of the adjustable parameters can be effectively determined, ensuring the effectiveness of parameter adjustment. By determining the target correlation between the influencing parameters and the X-ray intensity, the parameter values ​​of the influencing parameters can be modulated until the output X-ray intensity reaches the reference intensity, thus completing the static calibration of the X-ray emitter and ensuring the accuracy and effectiveness of static calibration.

[0127] Example 5:

[0128] Based on Example 1, this example provides a calibration method for a cargo density detection module. In step 2, samples of different densities are tested one by one based on the statically calibrated X-ray emitter, and the sensor parameters are dynamically calibrated according to the test results, including:

[0129] Based on the statically calibrated X-ray emitter, X-rays are emitted sequentially to samples of different densities, and the sensor is controlled to collect the attenuation signals of the X-rays corresponding to samples of different densities in sequence.

[0130] Obtain the background script file of the sensor, and extract the execution data of the sensor when collecting the attenuation signal of the ray from the background script file;

[0131] Based on the execution data, determine the timing parameters and automatic capture range for the sensor to capture the attenuation signal of the ray, and determine the emission timing and emission range of the ray based on the ray emitter;

[0132] The sensitivity of the sensor is determined by the timing parameters of the attenuation signal of the X-ray captured by the sensor and the automatic capture range, as well as by the X-ray emitter to determine the emission timing and emission range of the X-ray.

[0133] Simultaneously, the output results of the sensor's analysis of the attenuation signal of the X-rays are obtained, and the density detection value of the sensor for each known density sample is determined based on the output results;

[0134] The density detection accuracy of the sensor is determined based on the density detection value and the corresponding known density.

[0135] The system acquires the sensor's performance requirements based on the management terminal and determines the deviation between these requirements and the sensor's sensitivity and density detection accuracy.

[0136] The calibration direction and calibration value in the calibration direction are determined based on the deviation value to improve the sensitivity and density detection accuracy of the sensor.

[0137] The sensor parameters are dynamically calibrated based on the calibration direction and the calibration value in the calibration direction.

[0138] In this embodiment, dynamic calibration refers to calibration performed while the entire system is in operation (the transmitter emits rays, and the sensor receives signals).

[0139] In this embodiment, the attenuation signal refers to the signal whose intensity decreases after the X-ray penetrates the sample. The degree of attenuation follows a specific physical law (such as the Lambert-Beer law for X-rays) and is directly related to the density and thickness of the sample.

[0140] In this embodiment, the background script file refers to the underlying software, firmware, or configuration file that controls the core functions of the sensor.

[0141] In this embodiment, execution data refers to the key data logs extracted from the background script file that record the actual operation of the sensor. Examples include timestamps, trigger signal records, gain values, exposure times, and internal cache states.

[0142] In this embodiment, the emission timing refers to the time control parameters for the start and stop of the X-ray emitter. The actual emission timing is determined based on the dynamic process of the sample passing through the detection channel (e.g., conveyor belt speed) and the physical characteristics of the X-ray emitter (e.g., emission response time). For example, when the sample departs from the photoelectric sensor at the entrance of the conveyor belt channel, the emitter receives a "sample arrived" signal. The "emission start timing" is set to 0.5s after triggering (which matches the time it takes for the sample to move to the detection center), and the "emission end timing" is set to 0.3s after the sample has completely passed through the detection center (to ensure that the X-ray covers the full thickness of the sample). The emission duration (e.g., 1s) must be greater than the time it takes for the sample to pass through the detection area (e.g., 0.8s).

[0143] Timing parameters refer to the time control parameters for when the sensor starts and ends acquiring "attenuated signals"; for example, trigger delay, integration time, sampling frequency, etc. It ensures that the sensor opens the "window" to receive signals at the correct time. Reverse calibration is performed based on the transmission timing to ensure that the sensor's acquisition period completely covers the transmission period. For example: trigger delay: set to 0.48s (slightly earlier than the transmission start time by 0.5s) to ensure the sensor is ready at the start of transmission; integration time: set to 1.1s (slightly longer than the transmission duration by 1s) to avoid signal loss at sample edges due to mechanical delay; sampling frequency: set to 20Hz (10Hz higher than the transmission pulse frequency) to ensure each transmission pulse is sampled at least twice, improving the signal-to-noise ratio.

[0144] In this embodiment, the emission range refers to the range of radiation intensity output by the emitter. The emission range is determined based on the density and thickness range of the target cargo. The initial radiation intensity range is derived by back-calculating using a preset law (Lambert-Beer Law). For example, for high density (e.g., 10 g / cm³), the emission range is determined by the density and thickness range of the target cargo. 3 For samples with large thicknesses (e.g., 50 cm), the intensity attenuation is extremely low (set to 1 min), so it is necessary to ensure that the minimum initial intensity of the emitter is equal to or greater than the sensor's detection limit; for low densities (e.g., 0.3 g / cm³), the intensity attenuation is extremely low. 3 For samples with small thickness (e.g., 5cm), the intensity after attenuation is relatively high (set as Imax). It is necessary to ensure that the highest initial intensity of the transmitter is less than or equal to the upper limit of sensor detection. Based on the above range, the final emission range is set to (Imin, Imax).

[0145] In this embodiment, the sensor's sensitivity refers to first accurately comparing the sensor's trigger delay, integration time, and other "timing parameters" with the transmitter's emission timing to analyze the synchronization accuracy of signal acquisition; the smaller the delay, the higher the synchronization. Secondly, the sensor's "automatic acquisition range" (such as gain and range) is matched with the transmitter's "emission range" (ray intensity) to assess whether the signal is acquired without distortion within the sensor's optimal range. Finally, the sensor's response efficiency is calculated based on the synchronization accuracy and range matching degree, thereby quantifying its sensitivity: Sensitivity = (Synchronization Accuracy Score / 100) × (Range Matching Degree) × 100%. For example, with a synchronization delay of 0.05ms (100 points) and a matching degree of 95%, the sensitivity = 1 × 0.95 × 100% = 95%.

[0146] In this embodiment, the automatic capture range refers to the range of signal strength that the sensor can effectively receive. This is achieved by matching the attenuation signal range corresponding to the transmission range and adjusting the sensor hardware parameters (such as gain and range). The "gain calibration parameters" in the background script file (such as amplifying weak signals by 10 times and strong signals by 2 times) ensure that the attenuation signal falls within the linear response range of the sensor (such as 20% to 80% of the range to avoid nonlinear errors). At this time, the automatic capture range can be quantified as "automatic capture lower limit to automatic capture upper limit (corresponding to gain of 10x to 2x)".

[0147] In this embodiment, the performance requirements are set by the user based on experimental test results and needs, and uploaded to the management terminal; that is, the performance that the sensor needs to achieve.

[0148] The beneficial effects of the above technical solution are as follows: First, by emitting X-rays to samples of different densities through a statically calibrated X-ray emitter, and collecting the attenuation signals of the corresponding X-rays through a sensor, data support is provided for determining the current performance parameters of the sensor. Second, by retrieving and analyzing the execution data during data acquisition by the sensor, the sensitivity of the sensor can be effectively determined. At the same time, by analyzing the density detection value of the sensor, the density detection accuracy of the sensor can be effectively determined. Finally, based on the indicator requirements of the management terminal, it is determined whether the sensitivity and density detection accuracy of the sensor need to be calibrated. When calibration is required, the sensor parameters are dynamically calibrated by the deviation between the indicator requirements and the sensor's sensitivity and density detection accuracy, ensuring the reliability and accuracy of the sensor parameter calibration.

[0149] Example 6:

[0150] Based on Example 5, this example provides a calibration method for a cargo density detection module, which dynamically calibrates sensor parameters based on the calibration direction and the calibration value in the calibration direction, including:

[0151] Obtain the dynamic calibration results of the sensor parameters, and re-examine the performance of the sensor parameters after completing the dynamic calibration;

[0152] After the performance retest is passed, the dynamic calibration information of the sensor parameters is recorded throughout the entire process, and a dynamic calibration report is generated based on the full process record.

[0153] The dynamic calibration report is fed back to the management terminal for recording.

[0154] In this embodiment, performance retesting refers to re-checking the sensor parameters to determine whether the sensor has passed dynamic calibration.

[0155] The beneficial effects of the above technical solution are: by monitoring the dynamic calibration results of sensor parameters, and by re-checking the performance of sensor parameters after completing dynamic calibration, and by recording the dynamic calibration information of sensor parameters, the entire calibration process of the sensor can be effectively recorded.

[0156] Example 7:

[0157] Based on Example 1, this example provides a calibration method for a cargo density detection module, such as... Figure 3 As shown, in step 3, after dynamic calibration is completed, the historical operating parameters of the cargo density detection module within the target time period are retrieved, including:

[0158] S301: Read the start and end times of the target time period;

[0159] S302: Generate the first index label of the working parameter management library based on the start time and end time;

[0160] S303: Obtain the module type of the cargo density detection module and generate the second index label of the working parameter management library based on the module type;

[0161] S304: Retrieve the historical working parameters of the density detection module within the target time period from the working parameter management library based on the first index tag and the second index tag.

[0162] In this embodiment, the first index label is used to mark the time between the start time point and the end time point, which is beneficial for retrieving parameter data for that time period from the working parameter management library.

[0163] In this embodiment, the second index tag is used to mark the module type of the cargo density module, which is beneficial for retrieving parameter data consistent with the module type from the working parameter management library.

[0164] In this embodiment, the working parameter management library is pre-set.

[0165] The beneficial effect of the above technical solution is that by determining the first index label and the second index label respectively, the accuracy of retrieving historical working parameters for the target time period from the working parameter management library is effectively guaranteed.

[0166] Example 8:

[0167] Based on Example 1, this example provides a calibration method for a cargo density detection module. In step 4, historical operating parameters are analyzed to determine whether the cargo density detection module has any detection vulnerabilities. If a detection vulnerability exists, the cargo density detection module is recalibrated, including:

[0168] Obtain the obtained historical working parameters, read the historical working parameters, and determine the business composition of the historical working parameters;

[0169] Based on the business composition, historical operating parameters are divided into equipment operating parameters and density detection feedback parameters. The equipment operating parameters are then analyzed to determine the fluctuation range of the equipment's own performance.

[0170] When the fluctuation range of the device's own performance meets the baseline operating conditions, it is determined that the device's own performance has no defects; otherwise, it is determined that the device's own performance has a first defect.

[0171] At the same time, the density detection feedback parameters are analyzed to determine the detection deviation value of the sensor for the density parameters, and when the detection deviation value exceeds the preset deviation range, it is determined that there is a second vulnerability in the cargo density detection module.

[0172] The cargo density detection module was recalibrated based on the first and second vulnerabilities.

[0173] In this embodiment, the service composition refers to the equipment's own operation services and the services for detecting density included in the historical operating parameters.

[0174] In this embodiment, the fluctuation range of the device's own performance refers to the range of changes in the extreme values ​​of data amplitude during device operation.

[0175] In this embodiment, the baseline operating conditions refer to the parameter operating status of the device during normal operation.

[0176] In this embodiment, the preset deviation range is a pre-set standard used to determine whether a second vulnerability exists. The preset deviation range is pre-set based on the density characteristics of the goods, the purpose of the inspection, and other actual conditions, according to the experience of the personnel.

[0177] The beneficial effects of the above technical solution are: by analyzing historical working parameters, it is possible to accurately and effectively determine whether there are loopholes in the equipment itself in the cargo density detection module, as well as whether there are loopholes in the accuracy and sensitivity of density detection. Then, when loopholes exist, the cargo density detection module is recalibrated according to the specific circumstances of the loopholes, thus ensuring the accuracy and reliability of the cargo density detection module.

[0178] Example 9:

[0179] Based on Example 8, this example provides a calibration method for a cargo density detection module, which recalibrates the cargo density detection module based on a first vulnerability and a second vulnerability, including:

[0180] The device operating parameters corresponding to the first vulnerability and the density detection feedback parameters corresponding to the second vulnerability are obtained respectively. The device operating parameters and density detection feedback parameters are analyzed respectively to determine the corresponding abnormal parameter fragments.

[0181] The equipment components in the cargo density detection module are traced based on abnormal parameter fragments, and the target components corresponding to recalibration are determined based on the traceability results.

[0182] Based on the vulnerability characteristics of the first and second vulnerabilities, a recalibration strategy for the target component is determined, and the cargo density detection module is recalibrated based on the recalibration strategy.

[0183] In this embodiment, abnormal parameter segments refer to parameter ranges that do not conform to the baseline operating conditions and the preset deviation range.

[0184] In this embodiment, the recalibration strategy can be, for example, if the vulnerability is characterized by parameter drift (such as emission intensity attenuation), then a physical calibration strategy is adopted for hardware components such as the X-ray emitter. This physical calibration strategy addresses vulnerabilities caused by hardware aging, mechanical misalignment, performance degradation, etc. (such as X-ray emitter intensity drift, sensor physical position deviation, etc.). Calibration is achieved by directly adjusting the hardware state or structure. Specific operations include: First, physical calibration of the X-ray emitter (corresponding to the "emission intensity attenuation" vulnerability). If the vulnerability is characterized by "X-ray emission intensity attenuating over time (parameter drift)," leading to abnormal density detection feedback parameters (such as low-density goods being misjudged as high-density), then the physical calibration strategy includes: tube voltage / tube current calibration: Using a dedicated high-voltage meter and ammeter connected to the X-ray emitter, the tube voltage (e.g., drifting from a set 150kV to 140kV) is adjusted back to the reference value (150kV) by adjusting the hardware knob or dedicated calibration interface. Simultaneously, the tube current (e.g., drifting from 20mA to 18mA) is calibrated to the rated value to ensure initial... The radiation intensity is restored to the design standard; radiation source position calibration: if the relative position of the radiation source and the detection channel is offset due to mechanical vibration (e.g., deviating from the central axis by 5mm), the spatial coordinates of the radiation source are calibrated using a laser positioning instrument. By adjusting the screws of the fixing bracket, the radiation source is reset to the reference position (within ±0.5mm of the central axis) to ensure that the radiation beam uniformly covers the detection area; if the characteristic is a logical error (e.g., analysis algorithm deviation), a software parameter correction strategy is adopted for the sensor or processing module. The software parameter correction strategy is used to solve vulnerabilities caused by algorithm deviation, firmware configuration errors, signal processing logic failures, etc. (e.g., sensor data parsing errors, outdated density calculation model parameters, etc.). Calibration is achieved by modifying the software configuration and updating the algorithm parameters. Specific operations include: parameter correction of the sensor signal processing software (corresponding to the "analysis algorithm deviation" vulnerability). If the vulnerability characteristic is "sensor's parsing logic error for attenuation signals (e.g., outdated gain coefficient)", causing the density detection deviation to exceed the preset range (e.g., actual density 2.0g / cm³), 3 The cargo was tested at 2.5g / cm³. 3The software parameter correction strategy includes: gain coefficient update: log in to the sensor's backend system through the management terminal, retrieve the configuration file of the signal processing module, and correct outdated gain coefficients (such as the default gain of 1.2 that is no longer applicable due to component aging) to measured values ​​(such as the gain of 1.05 determined after calibration with standard samples) to ensure that the amplification factor of the attenuated signal matches the current hardware status; filter parameter adjustment: if the noise in the attenuated signal collected by the sensor is too large (such as fluctuations of ±5% due to unreasonable software filter parameters), modify the cutoff frequency of the filter algorithm (such as adjusting it from 10Hz to 5Hz) or the sliding window size (such as increasing it from 5 sampling points to 10) in the background script file to reduce the interference of noise on signal analysis; the core of the strategy is to restore the performance of the target component in a targeted manner based on the principle of "hardware problems are adjusted in hardware, software problems are repaired in software".

[0185] The beneficial effects of the above technical solution are as follows: by analyzing the device operating parameters corresponding to the first vulnerability and the density detection feedback parameters corresponding to the second vulnerability respectively, the abnormal parameter fragments can be determined based on the analysis results, and then the corresponding device components can be determined based on the abnormal parameter fragments. At the same time, the first vulnerability and the second vulnerability are analyzed to determine the vulnerability characteristics. Finally, the device components can be recalibrated based on the vulnerability characteristics, which ensures the reliability of the cargo density detection module and improves the accuracy of cargo density detection.

[0186] Example 10:

[0187] Based on Example 1, this example provides a calibration method for a cargo density detection module, which further includes: Step 5, constructing an intelligent calibration model, the specific process of which is as follows:

[0188] Collect historical calibration events and read the calibration data corresponding to each historical calibration event. The calibration data includes: static calibration data, dynamic calibration data, and recalibration data.

[0189] The static calibration data, dynamic calibration data, and recalibration data of each historical calibration event are analyzed to determine the static trigger conditions and the corresponding static calibration dataset for each static calibration, the dynamic trigger conditions and the corresponding dynamic calibration dataset for each dynamic calibration, and the trigger conditions and the corresponding recalibration dataset for each recalibration.

[0190] A static calibration prediction network is constructed by learning the static triggering conditions;

[0191] Learn from dynamic triggering conditions to construct a dynamic calibration prediction network;

[0192] The triggering conditions during recalibration are learned to construct a recalibration prediction network;

[0193] The static calibration prediction network, dynamic calibration prediction network, and recalibration prediction network are integrated to construct the target prediction model;

[0194] The static decision network is determined by learning the static calibration dataset corresponding to the static triggering condition, and the dynamic decision network is determined by learning the dynamic calibration dataset corresponding to the dynamic triggering condition. At the same time, the recalibration decision network is determined by learning the recalibration dataset corresponding to the recalibration triggering condition.

[0195] A target decision-making model is constructed by integrating static decision networks, dynamic decision networks, and recalibration decision networks.

[0196] The target prediction model and the target decision model are associated according to the calibration type to build an intelligent calibration model. When a new calibration event occurs, the target prediction model in the intelligent calibration model is used to determine in real time whether the triggering condition has been met.

[0197] When the trigger condition is met, the corresponding calibration type is determined, and the corresponding calibration data is analyzed and matched in the target decision model according to the calibration type. At the same time, the current cargo density detection module is calibrated in real time according to the matching result.

[0198] In this embodiment, a historical calibration event refers to a complete record of every calibration (whether static, dynamic, or recalibrated) performed in the past, and is considered an independent event.

[0199] In this embodiment, calibration data—specific data contained in an event package—can be categorized as follows: Static calibration data: such as influencing parameters, rate of change, and final adjusted parameter values; Dynamic calibration data: such as sensor timing parameters, accuracy, deviation values, and final calibrated sensitivity; Recalibration data: such as abnormal parameter fragments, traced target components, and the recalibration strategies adopted.

[0200] In this embodiment, the triggering condition refers to the root cause or state characteristic that triggers a calibration event.

[0201] In this embodiment, the target prediction model refers to the final prediction model after integrating or synthesizing the static / dynamic / recalibration prediction networks. That is, it is used to comprehensively analyze real-time data to determine whether calibration is needed and what type of calibration is needed.

[0202] In this embodiment, the target decision model refers to the final decision model after integration or synthesis through a static / dynamic / recalibration decision network (i.e., the optimal calibration parameter adjustment strategy to be adopted under certain triggering conditions). That is, once the prediction model determines that a certain calibration is needed, the decision model outputs a set of optimal specific calibration parameters and schemes that have been verified by historical data.

[0203] In this embodiment, the calibration type refers to static calibration, dynamic calibration, and recalibration.

[0204] In this embodiment, the static decision network is based on static calibration data, primarily consisting of "influencing parameters (such as tube voltage and tube current), rate of change, and final adjustment value." These are mostly structured discrete / continuous features (e.g., voltage adjustment value is continuous, parameter type is discrete) and lack strong temporal correlation. The algorithm for the static decision network is generated through training using gradient boosting trees (such as XGBoost or LightGBM) or random forests. Its network construction includes: Input: Static triggering conditions (e.g., feature vectors such as "ray intensity deviation > 5%" and "tube voltage fluctuation exceeding ±3kV"); Output: Specific parameter adjustment values ​​for static calibration (e.g., continuous / discrete actions such as "tube voltage adjusted from 150kV to 155kV" and "tube current adjusted from 20mA to 22mA"); Structure: Taking XGBoost as an example, multiple decision trees are integrated, with each tree learning the residual of the previous tree, ultimately outputting the optimal adjustment value (e.g., predicting the adjustment amount through a regression task).

[0205] In this embodiment, the dynamic decision network is based on the fact that the dynamic calibration data involves "sensor timing parameters (trigger delay, sampling frequency), sensitivity, and deviation value," which have weak temporal sequence characteristics (e.g., sensitivity changes at different times may be related), and needs to capture the dynamic correlation between "synchronization accuracy and calibration strategy." The algorithm of the dynamic decision network is generated by training a Long Short-Term Memory (LSTM) network. Its network construction includes: input: temporal features of dynamic triggering conditions (e.g., "trigger delay values ​​of the past 5 moments" and "sensitivity change sequence over time"); output: parameter adjustments for dynamic calibration (e.g., "integration time shortened from 1.1s to 1.0s" and "gain value adjusted from 10× to 8×"); structure: LSTM layer (capturing temporal features) + fully connected layer (outputting specific adjustment values), and dynamic data is processed through a temporal sliding window (e.g., taking data from the past 10 moments as the input window).

[0206] In this embodiment, the recalibration decision network is based on recalibration data containing "abnormal parameter fragments (such as hardware fluctuation range, software logic error codes), target components (ray emitters / sensors), and recalibration strategies." Essentially, it's a mapping from "vulnerability characteristics to remediation strategies," mixing categorical features (component type, vulnerability type) and continuous features (deviation value, fluctuation amplitude). The recalibration network algorithm is implemented using a decision tree + rule engine. Network construction includes: Input: recalibration trigger conditions (such as "hardware fluctuation range exceeds baseline ±10%" or "density detection deviation > 0.3 g / cm³"). 3 Features); Output: Specific parameters of the recalibration strategy (e.g., "Physical calibration: tube voltage is adjusted back to 150kV±1kV" "Software correction: filter cutoff frequency is adjusted from 10Hz to 5Hz"); Structure: Decision tree (classifies vulnerability types) → Rule engine (filters unreasonable strategies) → Fully connected layer (outputs specific adjustment values).

[0207] In this embodiment, the construction of the target decision model is based on the collaborative hub of a static decision network, a dynamic decision network, and a recalibration decision network. The core is: according to the calibration type, the corresponding decision network is called and the results are fused. Through an integrated learning framework, the "calibration type" (static / dynamic / recalibration) output by the target prediction model and the real-time trigger condition features are used as input. The fusion logic is as follows: if the calibration type is "static calibration", the output of the static decision network is directly called, and at the same time, the dynamic decision network is used to verify "whether static adjustment affects the dynamic performance of the sensor" (such as whether the tube voltage adjustment causes sensitivity fluctuation). If there is a conflict, it is fine-tuned through a meta-model (such as logistic regression); if it is "recalibration", the strategy of the recalibration decision network is called first, and the static / dynamic decision networks provide auxiliary parameters (such as the timing parameters of the sensor that need to be updated synchronously after hardware calibration).

[0208] In this embodiment, the intelligent calibration model is built based on an end-to-end integration of a "target prediction model + target decision model," requiring the implementation of an automated process of "real-time data input → trigger judgment → strategy output." It is an artificial intelligence system with complete "perception-decision-execution" capabilities. Its architecture is as follows: Data access layer: real-time acquisition of data such as ray emitter parameters, sensor signals, and density detection results; Prediction layer: The target prediction model (as described in the integrated network above) determines whether the triggering conditions and calibration type are met; Decision layer: The target decision model calls the corresponding decision network according to the type, outputting specific calibration parameters (such as adjustment values ​​and strategy steps); Execution layer: The decision results are converted into executable instructions for the device (such as controlling the voltage knob of the ray emitter or modifying the background script parameters of the sensor).

[0209] The working principle and beneficial effects of the above technical solution are as follows: by learning to construct a target prediction model and a target decision model, and by associating the target prediction model and the target decision model, the accuracy and comprehensiveness of the intelligent calibration model can be effectively guaranteed, and the intelligence of the calibration of the cargo density detection module can be effectively guaranteed.

[0210] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A calibration method for a cargo density detection module, characterized in that, include: Step 1: Detect the initial radiation intensity and perform static calibration on the radiation emitter based on the detection results; Step 2: Test samples of different densities one by one based on the statically calibrated X-ray emitter, and dynamically calibrate the sensor parameters according to the test results. Step 3: After dynamic calibration is completed, retrieve the historical operating parameters of the cargo density detection module within the target time period; Step 4: Analyze historical operating parameters to determine if there are any detection vulnerabilities in the cargo density detection module, and if so, recalibrate the cargo density detection module.

2. The calibration method for a cargo density detection module according to claim 1, characterized in that, In step 1, the initial radiation intensity is detected, and the radiation emitter is statically calibrated based on the detection results, including: Record the initial radiation intensity of the radiation emitter multiple times under no-load conditions, analyze the initial radiation intensity, and determine whether static calibration of the radiation emitter is required. When it is determined that static calibration of the X-ray emitter is required, the adjustment parameters of the X-ray emitter are analyzed, and static calibration of the X-ray emitter is performed based on the analysis results.

3. The calibration method for a cargo density detection module according to claim 2, characterized in that, Record the initial radiation intensity of the radiation emitter multiple times under no-load conditions, and analyze the initial radiation intensity to determine whether static calibration of the radiation emitter is required, including: Record the initial radiation intensity of the radiation emitter multiple times under no-load conditions, and calculate the average target intensity of the initial radiation intensity of the multiple emissions. Based on the attribute information of the radiation emitter, the reference intensity value of the radiation emitter is matched in the preset management library; the average target intensity value is compared with the reference intensity value to determine whether static calibration of the radiation emitter is required; If the average target intensity is consistent with the reference intensity value, then it is determined that no static calibration of the ray emitter is required; otherwise, it is determined that static calibration of the ray emitter is required.

4. The calibration method for a cargo density detection module according to claim 2, characterized in that, When it is determined that static calibration of the radiation emitter is required, the adjustment parameters of the radiation emitter are analyzed, and static calibration is performed on the radiation emitter based on the analysis results, including: When it is determined that static calibration of the X-ray emitter is required, the adjustable parameters in the X-ray emitter are obtained, and a parameter group corresponding to each adjustable parameter is constructed. Each adjustable parameter is adjusted according to its corresponding parameter group, and the target rate of change of radiation intensity under no-load conditions is recorded based on the adjustment results. Compare the target rate of change with a preset rate of change threshold; When the target rate of change is greater than or equal to the preset rate of change threshold, the corresponding adjustable parameter will be used as the influencing parameter. Otherwise, the corresponding adjustable parameter will be treated as a non-affecting parameter; Retrieve the radiation intensity values ​​corresponding to each parameter in the parameter group that affects the parameters; Based on the correspondence between each parameter in the parameter group corresponding to the influencing parameters and the radiation intensity value, determine the target correlation between the influencing parameters and the radiation intensity; Adjust the parameters of the influencing parameters according to the target correlation until the output radiation intensity reaches the reference intensity value, thus completing the static calibration of the radiation emitter.

5. The calibration method for a cargo density detection module according to claim 1, characterized in that, In step 2, samples of different densities are tested one by one based on the statically calibrated X-ray emitter, and the sensor parameters are dynamically calibrated based on the test results, including: Based on the statically calibrated X-ray emitter, X-rays are emitted sequentially to samples of different densities, and the sensor is controlled to collect the attenuation signals of the X-rays corresponding to samples of different densities in sequence. Obtain the background script file of the sensor, and extract the execution data of the sensor when collecting the attenuation signal of the ray from the background script file; Based on the execution data, determine the timing parameters and automatic capture range for the sensor to capture the attenuation signal of the ray, and determine the emission timing and emission range of the ray based on the ray emitter; The sensitivity of the sensor is determined by the timing parameters of the attenuation signal of the X-ray captured by the sensor and the automatic capture range, as well as by the X-ray emitter to determine the emission timing and emission range of the X-ray. Simultaneously, the output results of the sensor's analysis of the attenuation signal of the X-rays are obtained, and the density detection value of the sensor for each known density sample is determined based on the output results; The density detection accuracy of the sensor is determined based on the density detection value and the corresponding known density. The system acquires the sensor's performance requirements based on the management terminal and determines the deviation between these requirements and the sensor's sensitivity and density detection accuracy. The calibration direction and calibration value in the calibration direction are determined based on the deviation value to improve the sensitivity and density detection accuracy of the sensor. The sensor parameters are dynamically calibrated based on the calibration direction and the calibration value in the calibration direction.

6. The calibration method for a cargo density detection module according to claim 5, characterized in that, Dynamic calibration of sensor parameters is performed based on the calibration direction and the calibration value in the calibration direction, including: Obtain the dynamic calibration results of the sensor parameters, and re-examine the performance of the sensor parameters after completing the dynamic calibration; After the performance retest is passed, the dynamic calibration information of the sensor parameters is recorded throughout the entire process, and a dynamic calibration report is generated based on the full process record. The dynamic calibration report is fed back to the management terminal for recording.

7. The calibration method for a cargo density detection module according to claim 1, characterized in that, In step 3, after dynamic calibration is completed, the historical operating parameters of the cargo density detection module within the target time period are retrieved, including: Read the start and end times of the target time period; Generate the first index label for the work parameter management library based on the start and end times; Obtain the module type of the cargo density detection module, and generate a second index label for the working parameter management library based on the module type; Based on the first and second index tags, retrieve the historical operating parameters of the density detection module within the target time period from the operating parameter management library.

8. The calibration method for a cargo density detection module according to claim 1, characterized in that, In step 4, historical operating parameters are analyzed to determine if there are any detection vulnerabilities in the cargo density detection module. If vulnerabilities are found, the cargo density detection module is recalibrated, including: Obtain the obtained historical working parameters, read the historical working parameters, and determine the business composition of the historical working parameters; Based on the business composition, historical operating parameters are divided into equipment operating parameters and density detection feedback parameters. The equipment operating parameters are then analyzed to determine the fluctuation range of the equipment's own performance. When the fluctuation range of the device's own performance meets the baseline operating conditions, it is determined that the device's own performance has no defects; otherwise, it is determined that the device's own performance has a first defect. At the same time, the density detection feedback parameters are analyzed to determine the detection deviation value of the sensor for the density parameters, and when the detection deviation value exceeds the preset deviation range, it is determined that there is a second vulnerability in the cargo density detection module. The cargo density detection module was recalibrated based on the first and second vulnerabilities.

9. A calibration method for a cargo density detection module according to claim 8, characterized in that, The cargo density detection module was recalibrated based on the first and second vulnerabilities, including: The device operating parameters corresponding to the first vulnerability and the density detection feedback parameters corresponding to the second vulnerability are obtained respectively. The device operating parameters and density detection feedback parameters are analyzed respectively to determine the corresponding abnormal parameter fragments. The equipment components in the cargo density detection module are traced based on abnormal parameter fragments, and the target components corresponding to recalibration are determined based on the traceability results. Based on the vulnerability characteristics of the first and second vulnerabilities, a recalibration strategy for the target component is determined, and the cargo density detection module is recalibrated based on the recalibration strategy.

10. The calibration method for a cargo density detection module according to claim 1, characterized in that, It also includes: Step 5, constructing the intelligent calibration model, the specific process of which is as follows: Collect historical calibration events and read the calibration data corresponding to each historical calibration event. The calibration data includes: static calibration data, dynamic calibration data, and recalibration data. The static calibration data, dynamic calibration data, and recalibration data of each historical calibration event are analyzed to determine the static trigger conditions and the corresponding static calibration dataset for each static calibration, the dynamic trigger conditions and the corresponding dynamic calibration dataset for each dynamic calibration, and the trigger conditions and the corresponding recalibration dataset for each recalibration. A static calibration prediction network is constructed by learning the static triggering conditions; Learn from dynamic triggering conditions to construct a dynamic calibration prediction network; The triggering conditions during recalibration are learned to construct a recalibration prediction network; The static calibration prediction network, dynamic calibration prediction network, and recalibration prediction network are integrated to construct the target prediction model; The static decision network is determined by learning the static calibration dataset corresponding to the static triggering condition, and the dynamic decision network is determined by learning the dynamic calibration dataset corresponding to the dynamic triggering condition. At the same time, the recalibration decision network is determined by learning the recalibration dataset corresponding to the recalibration triggering condition. A target decision-making model is constructed by integrating static decision networks, dynamic decision networks, and recalibration decision networks. The target prediction model and the target decision model are associated according to the calibration type to build an intelligent calibration model. When a new calibration event occurs, the target prediction model in the intelligent calibration model is used to determine in real time whether the triggering condition has been met. When the trigger condition is met, the corresponding calibration type is determined, and the corresponding calibration data is analyzed and matched in the target decision model according to the calibration type. At the same time, the current cargo density detection module is calibrated in real time according to the matching result.

Images