High-altitude radioactive point source dose rate distribution analysis method and system

Abstract

The invention relates to the field of gamma dose rate monitoring, and provides a high-altitude radioactive point source dose rate distribution analysis method and system, and the method comprises the steps: collecting original counting data and energy spectrum data, and obtaining environment parameter data and flight carrier position information; performing statistical calculation on the original counting data to obtain counting statistical parameters; performing nonlinear correction on the original counting data, the energy spectrum data, the environmental parameter data and the counting statistical parameters to obtain corrected detector data and measurement uncertainty; performing fusion calculation on the corrected detector data to obtain a fusion weight, and performing verification processing to obtain fusion dose rate data; performing prediction processing on the fused dose rate data and the environmental parameter data to obtain a range switching signal; and performing reconstruction processing on the fused dose rate data, the range switching signal, the measurement uncertainty and the flight carrier position information to obtain ground dose rate distribution data. According to the invention, the accuracy and reliability of high-altitude radioactive point source dose rate distribution analysis are improved.

Description

Technical Field

[0001] This invention relates to the field of gamma dose rate monitoring, and more particularly to a method and system for analyzing the dose rate distribution of high-altitude radioactive point sources. Background Technology

[0002] Upper-altitude radioactive point source dose rate distribution analysis is a crucial technical means for nuclear emergency response, environmental radiation monitoring, and nuclear safety assessment. It utilizes gamma-ray detectors mounted on flight platforms to rapidly locate ground-based radioactive point sources and reconstruct their dose rate distribution, finding wide application in critical areas such as nuclear accident emergency response, monitoring around nuclear facilities, and radioactive material search. However, in practical applications, upper-altitude gamma-ray detection systems are affected by complex upper-altitude environmental changes, differences in multi-detector responses, dynamic range switching, and atmospheric attenuation, leading to various error sources in the detection data, including Poisson noise, systematic bias, and measurement uncertainty.

[0003] In existing technologies, dose rate distribution analysis of high-altitude radioactive point sources mainly employs traditional interpolation algorithms and basic data fusion methods to achieve basic dose rate distribution reconstruction. However, existing methods do not adequately consider the intrinsic correlation between the multi-source error propagation mechanism of high-altitude detection systems and the physical laws of gamma-ray transmission. This makes it difficult to organically integrate objective radiation physics constraints with actual detector response characteristics, resulting in the inability to achieve high-precision and high-reliability ground dose rate distribution reconstruction. Summary of the Invention

[0004] In view of this, the present invention proposes a method and system for analyzing the dose rate distribution of high-altitude radioactive point sources. This solves the problem that the existing technology does not adequately consider the intrinsic correlation between the multi-source error propagation mechanism and the physical law of gamma-ray transmission in high-altitude detection systems, making it difficult to organically integrate objective radiation physics constraints with actual detector response characteristics, thus resulting in the inability to achieve high-precision and high-reliability ground dose rate distribution reconstruction.

[0005] The technical solution of this invention is implemented as follows: On one hand, this invention provides a method for analyzing the dose rate distribution of high-altitude radioactive point sources, comprising the following steps: Raw counting data and energy spectrum data were collected using multiple gamma-ray detectors, along with environmental parameter data and flight vehicle location information. Statistical calculations were performed on the original count data based on the statistical law of Poisson distribution to obtain the count statistical parameters; A Poisson-constrained physical sensing residual neural network is used to perform nonlinear correction on the original counting data, energy spectrum data, environmental parameter data, and counting statistics parameters to obtain the corrected detector data and measurement uncertainty. The fused detector data is fused using a fusion method based on maximizing mutual information to obtain fusion weights. The fusion weights are then verified using a physical consistency verifier to obtain fused dose rate data. A physically constrained range switching prediction network is used to predict the fused dose rate data and environmental parameter data to obtain the range switching signal. The fused dose rate data, range switching signal, measurement uncertainty, and flight vehicle location information are reconstructed to obtain ground dose rate distribution data.

[0006] Based on the above technical solutions, preferably, the nonlinear correction includes: The original counting data, energy spectrum data, environmental parameter data and counting statistics parameters are input into the Poisson constrained linear layer for Poisson noise constraint processing, and the Poisson constrained features are output. The physical perception residual neural network includes a Poisson constrained linear layer, a physical information residual block and an uncertainty quantification block. The Poisson constraint features are input into the physical information residual block for physical constraint residual learning, and feature mapping is performed based on the physical laws of γ-ray transmission to output physical perception features. The physical sensing features are input into the uncertainty quantification block for Bayesian inference, and the corrected detector data and the measurement uncertainty are output. The Poisson-constrained linear layer constrains the network training process using the Poisson loss function.

[0007] Based on the above technical solutions, preferably, the physical information residual block adopts a multi-detector collaborative correction mechanism, including: Construct a detector response difference matrix and establish cross-calibration relationships between detectors based on the differences in energy response characteristics between GM tube detectors and scintillator detectors; The high-altitude environment adaptive compensation algorithm is used to dynamically adjust the physical constraint parameters based on the real-time changes in flight altitude and atmospheric density. The physical constraint parameters include the atmospheric attenuation coefficient and the energy conversion factor. A spectral-guided weighting strategy is adopted, which assigns differentiated correction weights to γ-ray counts in different energy ranges based on the energy distribution characteristics of the spectral data. Establish a consistency verification mechanism among detectors, calculate the cross-correlation coefficient of the calibration results of different detectors, and perform a recalibration process when the cross-correlation coefficient is lower than a preset threshold.

[0008] Based on the above technical solutions, preferably, the step of performing fusion calculation on the corrected detector data using a fusion method based on maximizing mutual information to obtain fusion weights, and then verifying the fusion weights using a physical consistency verifier to obtain fused dose rate data, includes: Calculate the mutual information matrix between the corrected detector data, and determine the initial fusion weights of each detector by maximizing the mutual information criterion; The initial fusion weights are verified using a multi-level physical consistency verification strategy, which includes energy conservation verification, spatial distribution consistency verification, and time series continuity verification. The initial fusion weights are adaptively adjusted based on the verification results. When the physical consistency deviation exceeds a preset threshold, the weights are reallocated. The fusion dose rate data is obtained through weighted fusion calculation, and fusion quality evaluation indicators are generated.

[0009] Based on the above technical solutions, preferably, the weight redistribution adopts a detector dynamic reliability assessment algorithm, including: Establish a database of historical detector performance to record the measurement accuracy and stability of each detector under different environmental conditions; A reliability quantification model based on entropy weight theory is adopted to evaluate the real-time reliability level of the detector based on the information entropy change of the detector's current measurement data; The sliding window anomaly detection method is used to monitor the detector output data in real time, identify abnormal fluctuations in the detector output data, and calculate the severity of the anomaly. A reliability-weight mapping function is established to perform nonlinear mapping on the detector reliability assessment results, thereby obtaining the fusion weight correction coefficient. The fusion weight is then dynamically adjusted based on the fusion weight correction coefficient.

[0010] Based on the above technical solutions, preferably, the step of acquiring raw counting data and energy spectrum data through multiple gamma-ray detectors, and obtaining environmental parameter data and flight vehicle position information, includes: The configuration and initialization of multiple gamma-ray detectors mounted on the flight vehicle are performed. The multiple gamma-ray detectors include GM tube detectors and scintillator detectors. The operating voltage and counting threshold of the GM tube detectors are set, and the high voltage and energy calibration parameters of the scintillator detectors are set. Data is collected synchronously by the multiple gamma-ray detectors to obtain the raw count data and energy spectrum data of each detector, as well as environmental parameter data and flight vehicle position information.

[0011] Based on the above technical solutions, preferably, the step of performing statistical calculations on the original counting data based on the Poisson distribution statistical law to obtain counting statistical parameters includes: Based on the statistical law of Poisson distribution, a Poisson distribution model is performed on the original count data to establish a Poisson statistical model of the count data; The Poisson statistical model is used to calculate the count mean, count variance, and statistical uncertainty of each detector, and the count statistics parameters are output.

[0012] Based on the above technical solutions, preferably, the range switching prediction network using physical constraints performs prediction processing on the fused dose rate data and environmental parameter data to obtain the range switching signal, including: The range switching prediction network includes an evolution predictor, an event detector, and a constraint actuator. The evolution predictor and the event detector perform time-series predictive analysis on the fused dose rate data and environmental parameter data to identify dose rate change trends and abnormal events, and generate predictive range switching suggestion signals. The constraint actuator performs physical constraint verification on the predictive range switching suggestion signal, and outputs a constraint-optimized range switching signal based on the physical feasibility of range switching and system stability requirements.

[0013] Based on the above technical solutions, preferably, the reconstructing process of the fused dose rate data, range switching signal, measurement uncertainty, and flight vehicle position information to obtain ground dose rate distribution data includes: The fused dose rate data is processed for range continuity compensation based on the range switching signal, and error propagation calculation is performed based on the measurement uncertainty to obtain compensated and corrected dose rate data and comprehensive uncertainty. Based on the flight vehicle location information, an high-altitude-ground spatial mapping model is established, and a physically constrained spatial reconstruction algorithm is used to reconstruct the compensated and corrected dose rate data to generate the ground dose rate distribution data.

[0014] On the other hand, the present invention also provides a high-altitude radioactive point source dose rate distribution analysis system, the system comprising: The multi-source data acquisition module is used to acquire raw counting data and energy spectrum data through multiple gamma-ray detectors, and to obtain environmental parameter data and flight vehicle position information; The Poisson statistical calculation module is used to perform statistical calculations on the original count data based on the statistical law of Poisson distribution to obtain the count statistical parameters; The physical sensing correction module is used to perform nonlinear correction on the original counting data, energy spectrum data, environmental parameter data and counting statistics parameters using a Poisson-constrained physical sensing residual neural network, so as to obtain the corrected detector data and measurement uncertainty. The fusion weight verification module is used to perform fusion calculation on the corrected detector data using a fusion method based on maximizing mutual information to obtain fusion weights, and to verify the fusion weights according to the physical consistency verifier to obtain fused dose rate data. The range prediction control module is used to perform prediction processing on the fused dose rate data and environmental parameter data using a physically constrained range switching prediction network to obtain a range switching signal. The spatial reconstruction mapping module is used to reconstruct the fused dose rate data, range switching signal, measurement uncertainty and flight vehicle position information to obtain ground dose rate distribution data.

[0015] The high-altitude radioactive point source dose rate distribution analysis method and system of the present invention have the following advantages over the prior art: (1) By constructing a Poisson-constrained physical perception residual neural network, the correction of multi-detector data is realized. High-precision data fusion is achieved by combining the fusion method based on mutual information maximization and the physical consistency verification mechanism. The real-time management of detectors is realized by using a physical constraint range switching prediction network. The accurate mapping of high-altitude-ground dose rate distribution is realized by using a physical constraint spatial reconstruction algorithm, which improves the accuracy and reliability of high-altitude radioactive point source dose rate distribution analysis. (2) By constructing a deep fusion architecture of Poisson constraint linear layer, physical information residual block and uncertainty quantification block, multi-detector collaborative correction was realized. The physical constraint parameters were dynamically adjusted by the high-altitude environment adaptive compensation algorithm, which improved the data correction accuracy in complex high-altitude environments. At the same time, the energy spectrum guided weight allocation strategy and the inter-detector consistency verification mechanism were adopted to ensure the physical rationality of the correction results and the coordination consistency of multi-detector data, thereby enhancing the measurement reliability in dynamic environments. (3) By constructing a fusion method based on maximizing mutual information and a multi-level physical consistency verification strategy, the data of multiple detectors are optimized and fused. Through the detector dynamic reliability assessment algorithm and the activation of backup detectors, the adaptive capability and fault tolerance in complex environments are improved. At the same time, the real-time monitoring and predictive maintenance of detector performance are realized by adopting the entropy weight theory and the sliding window anomaly detection mechanism. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a high-altitude radioactive point source dose rate distribution analysis method according to the present invention. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 This invention provides a method for analyzing the dose rate distribution of high-altitude radioactive point sources, comprising the following steps: Raw counting data and energy spectrum data were collected using multiple gamma-ray detectors, along with environmental parameter data and flight vehicle location information. Statistical calculations were performed on the original count data based on the statistical law of Poisson distribution to obtain the count statistical parameters; A Poisson-constrained physical sensing residual neural network is used to perform nonlinear correction on the original counting data, energy spectrum data, environmental parameter data, and counting statistics parameters to obtain the corrected detector data and measurement uncertainty. The fused detector data is fused using a fusion method based on maximizing mutual information to obtain fusion weights. The fusion weights are then verified using a physical consistency verifier to obtain fused dose rate data. A physically constrained range switching prediction network is used to predict the fused dose rate data and environmental parameter data to obtain the range switching signal. The fused dose rate data, range switching signal, measurement uncertainty, and flight vehicle location information are reconstructed to obtain ground dose rate distribution data.

[0020] Specifically, this embodiment achieves the correction of multi-detector data by constructing a Poisson-constrained physical perception residual neural network, combines a fusion method based on maximizing mutual information and a physical consistency verification mechanism to achieve high-precision data fusion, uses a physically constrained range switching prediction network to achieve real-time management of detectors, and achieves accurate mapping of high-altitude-ground dose rate distribution through a physically constrained spatial reconstruction algorithm, thereby improving the accuracy and reliability of high-altitude radioactive point source dose rate distribution analysis.

[0021] The process of acquiring raw counting data and energy spectrum data through multiple gamma-ray detectors, and obtaining environmental parameter data and flight vehicle location information, includes: The configuration and initialization of multiple gamma-ray detectors mounted on the flight vehicle are performed. The multiple gamma-ray detectors include GM tube detectors and scintillator detectors. The operating voltage and counting threshold of the GM tube detectors are set, and the high voltage and energy calibration parameters of the scintillator detectors are set.

[0022] In one specific embodiment, the operating voltage range of the GM tube detector is set to 400-1200V, and the counting threshold is set to 10%-30% of the signal amplitude. The high voltage range of the scintillator detector is set to 600-1000V, and the energy is calibrated using a standard gamma-ray source to establish the correspondence between the energy channel and the ray energy. The environmental parameter data acquisition parameters are set, including a temperature measurement range of -40℃ to +60℃, a humidity measurement range of 0-100%RH, and a barometric pressure measurement range of 300-1100hPa.

[0023] Data is collected synchronously by the multiple gamma-ray detectors to obtain the raw count data and energy spectrum data of each detector, as well as environmental parameter data and flight vehicle position information.

[0024] In one specific embodiment, the data acquisition frequency is set to 1-10Hz, and a time synchronization mechanism is used to ensure that the data acquisition time of the multiple gamma-ray detectors is consistent, with a time synchronization accuracy of not less than 1ms. The raw counting data is formatted and stored according to the counting rate, counting time, and detector number; The energy spectrum data is formatted and stored according to energy channels, counts, and timestamps; The environmental parameter data and flight vehicle location information are timestamped and stored together with the gamma-ray detection data.

[0025] Specifically, this embodiment establishes a standardized multi-detector collaborative working mechanism by precisely configuring, initializing, and optimizing the parameters of the GM tube detector and the scintillator detector, effectively solving the problems of response characteristic differences and environmental adaptability among different types of detectors. By setting a strict time synchronization mechanism (accuracy no less than 1ms) and a standardized data formatting and storage scheme, the spatiotemporal consistency and traceability of multi-detector data are ensured. Simultaneously, through wide-range environmental parameter monitoring (temperature -40℃ to +60℃, humidity 0-100%RH, air pressure 300-1100hPa) and high-precision energy scale calibration, high-quality basic data is provided for Poisson statistical analysis and physical sensing correction.

[0026] The statistical calculation of the original count data based on the Poisson distribution statistical law yields the count statistical parameters, including: Based on the statistical law of Poisson distribution, a Poisson distribution model is constructed for the original count data to establish a Poisson statistical model for the count data.

[0027] In one specific embodiment, the raw counting data is grouped according to a measurement time window, wherein the time window is set to 1 second to 60 seconds; The counting data within each time window are fitted using the probability density function of the Poisson distribution to verify that the counting data conforms to the characteristics of the Poisson distribution. A Poisson statistical model is established that includes a count rate parameter and a time parameter. The property that the mean of the Poisson statistical model equals the variance is used to verify the effectiveness of the modeling.

[0028] The Poisson statistical model is used to calculate the count mean, count variance, and statistical uncertainty of each detector, and the count statistics parameters are output.

[0029] In one specific embodiment, the average count of each detector within each time window is calculated based on the Poisson statistical model, and the average count is equal to the total count within that time window divided by the time length; The count variance is calculated based on the characteristics of the Poisson distribution, and the count variance is numerically equal to the count mean. The statistical uncertainty is calculated based on the count mean, where the statistical uncertainty is the square root of the count mean. The count mean, count variance, and statistical uncertainty are combined and output as the count statistics parameters.

[0030] Specifically, this embodiment constructs a Poisson distribution statistical modeling mechanism to extract statistical characteristics from the original counting data, solving the problem that traditional methods ignore the random fluctuations in counting data. By setting a flexible time window (1-60 seconds) and data fitting verification based on the probability density function of the Poisson distribution, the physical correctness and mathematical rigor of the counting statistical model are ensured. At the same time, the inherent property that the mean of the Poisson distribution equals the variance is used to verify the effectiveness of the modeling, and the statistical uncertainty is accurately calculated based on the square root property of the count mean, thereby improving the accuracy of the entire system in handling random counting fluctuations and the quantification accuracy of statistical uncertainty.

[0031] The nonlinear correction includes: The original counting data, energy spectrum data, environmental parameter data and counting statistics parameters are input into the Poisson constrained linear layer for Poisson noise constraint processing, and the Poisson constrained features are output. The physical perception residual neural network includes a Poisson constrained linear layer, a physical information residual block and an uncertainty quantification block. The Poisson constraint features are input into the physical information residual block for physical constraint residual learning, and feature mapping is performed based on the physical laws of γ-ray transmission to output physical perception features. The physical sensing features are input into the uncertainty quantification block for Bayesian inference, and the corrected detector data and the measurement uncertainty are output. The Poisson-constrained linear layer constrains the network training process through the Poisson loss function, and the physical information residual block ensures that the output conforms to the gamma-ray decay law through the physical consistency loss function.

[0032] The physical information residual block employs a multi-detector collaborative correction mechanism, including: Construct a detector response difference matrix and establish cross-calibration relationships between detectors based on the differences in energy response characteristics between GM tube detectors and scintillator detectors; The high-altitude environment adaptive compensation algorithm is used to dynamically adjust the physical constraint parameters based on the real-time changes in flight altitude and atmospheric density. The physical constraint parameters include the atmospheric attenuation coefficient and the energy conversion factor. A spectral-guided weighting strategy is adopted, which assigns differentiated correction weights to γ-ray counts in different energy ranges based on the energy distribution characteristics of the spectral data. Establish a consistency verification mechanism among detectors, calculate the cross-correlation coefficient of the calibration results of different detectors, and perform a recalibration process when the cross-correlation coefficient is lower than a preset threshold to ensure the physical consistency of data from multiple detectors.

[0033] In one specific embodiment, the Poisson loss function is calculated as follows: ; in, The loss function is a Poisson constraint. The input dataset for the physical perception residual neural network is... The expected output label set of the physical perception residual neural network; The number of detectors (including GM tubes and scintillator detectors). This represents the number of time sampling points; For the first The detector at the The actual count value at each time point. For the first The detector at the The mean of the predicted counts at each time point; For the first The Poisson variance weighting coefficients of each detector For the first The log-likelihood weighting coefficients of each detector For collaborative correction weights among detectors; For gamma function, This is the detector response difference correction function. This represents the parameter indicating the difference in response between detectors.

[0034] The formula for dynamically adjusting the physical constraint parameters of the high-altitude environment adaptive compensation algorithm is as follows: ; in, Adaptive physical constraint parameters for high-altitude environments; These are the baseline physical parameters under standard conditions; For height The relevant atmospheric attenuation coefficient; The height difference relative to the reference height; This is the temperature correction factor. This is the pressure correction factor; The current ambient temperature. Standard temperature (20℃); Due to current environmental pressures, Standard pressure (101.325 kPa); This is an environmental response function related to the detector type. Atmospheric density, The energy of gamma rays; It is an exponential function.

[0035] Specifically, this embodiment achieves multi-detector collaborative correction by constructing a deep fusion architecture of a Poisson-constrained linear layer, a physical information residual block, and an uncertainty quantification block. It improves the data correction accuracy in complex high-altitude environments by dynamically adjusting physical constraint parameters through an adaptive compensation algorithm for high-altitude environments. At the same time, it adopts an energy spectrum-guided weight allocation strategy and a consistency verification mechanism between detectors to ensure the physical rationality of the correction results and the coordination consistency of multi-detector data, thereby enhancing the measurement reliability in dynamic environments.

[0036] The process involves fusing the corrected detector data using a mutual information maximization-based fusion method to obtain fusion weights, and then verifying these fusion weights using a physical consistency verifier to obtain fused dose rate data, including: Calculate the mutual information matrix between the corrected detector data, and determine the initial fusion weights of each detector by maximizing the mutual information criterion; The initial fusion weights are verified using a multi-level physical consistency verification strategy, which includes energy conservation verification, spatial distribution consistency verification, and time series continuity verification. The initial fusion weights are adaptively adjusted based on the verification results. When the physical consistency deviation exceeds a preset threshold, the weights are reallocated. The fusion dose rate data is obtained through weighted fusion calculation, and a fusion quality evaluation index is generated. The fusion quality evaluation index is used to guide parameter optimization in subsequent processing steps.

[0037] The weight redistribution employs a detector dynamic reliability assessment algorithm, including: Establish a database of historical detector performance to record the measurement accuracy and stability of each detector under different environmental conditions; A reliability quantification model based on entropy weight theory is adopted to evaluate the real-time reliability level of the detector based on the information entropy change of the detector's current measurement data; The sliding window anomaly detection method is used to monitor the detector output data in real time, identify abnormal fluctuations in the detector output data, and calculate the severity of the anomaly. A reliability-weight mapping function is established to perform nonlinear mapping on the detector reliability assessment results, thereby obtaining the fusion weight correction coefficient. The fusion weight is then dynamically adjusted based on the fusion weight correction coefficient. When a significant degradation in detector performance is detected, the backup detector activation mechanism is automatically triggered to ensure the continuity and accuracy of the fusion dose rate data.

[0038] In one specific embodiment, the formula for calculating the fusion weight of the fusion method based on maximizing mutual information is: ; in, For the fusion weight vector; The parameters that maximize the objective function The value; To correct the number of data channels of the detector, For the first and the Individual detector calibration data in weight Mutual information under certain conditions For the first calibrated detector data for each detector; The weight regularization coefficient; For the regularization term of the weight vector, ; This is the physical consistency weighting coefficient; This is a multi-level physical consistency verification function.

[0039] The calculation formula for the reliability quantification model based on entropy weight theory is as follows: ; ; in, For the first Each detector at time Dynamic reliability score; This is the kurtosis parameter of the sigmoid function; For the first Each detector at time Information entropy value; The entropy threshold for reliability determination; The number of historical performance evaluation indicators; For the first The detector Time decay weights for historical performance metrics; For the first Each detector at time Anomaly detection correction factor based on sliding window; To adjust the sliding window size; For the first Each detector at time Measurement data, For the first in the sliding window The mean of data from each detector For the first The normal fluctuation standard deviation of each detector The hyperbolic tangent activation function is used. It is an exponential function.

[0040] Specifically, this embodiment optimizes and fuses multi-detector data by constructing a fusion method based on maximizing mutual information and a multi-level physical consistency verification strategy. Through the detector dynamic reliability assessment algorithm and the activation of backup detectors, it improves the adaptability and fault tolerance in complex environments. At the same time, it uses entropy weight theory and sliding window anomaly detection mechanism to realize real-time monitoring and predictive maintenance of detector performance.

[0041] The physically constrained range switching prediction network performs prediction processing on the fused dose rate data and environmental parameter data to obtain the range switching signal, including: The range switching prediction network includes an evolution predictor, an event detector, and a constraint actuator. The evolution predictor and the event detector perform time-series prediction analysis on the fused dose rate data and environmental parameter data to identify dose rate change trends and abnormal events, and generate predictive range switching suggestion signals.

[0042] In one specific embodiment, the evolution predictor is constructed using a long short-term memory network to perform multi-step long-time-series prediction on the fusion dose rate data, with the prediction time window set to the future 1-30 seconds; A dose rate mutation detection model is established through the event detector, and a dose rate change rate threshold is set. When the detected dose rate change rate exceeds the threshold, it is identified as an abnormal event. An environmental factor correlation analysis mechanism is established to correct the dose rate prediction results based on the changes in temperature, humidity, and air pressure in the environmental parameter data. Based on the matching degree analysis between the dose rate prediction and the current measurement range, a predictive range switching suggestion signal is generated, which includes the switching direction, switching timing, and switching urgency.

[0043] The constraint actuator performs physical constraint verification on the predictive range switching suggestion signal, and outputs a constraint-optimized range switching signal based on the physical feasibility of range switching and system stability requirements.

[0044] In one specific embodiment, a physical constraint model for range switching is established, which includes detector response time constraints, range overlap interval constraints, and switching frequency limitation constraints. The feasibility of the predictive range switching suggestion signal is constrained to verify whether the switching operation meets the physical characteristics of the detector and the stability requirements of the system. A switching cost evaluation algorithm is used to calculate the impact of each range switch on measurement continuity and accuracy. When the switching cost exceeds a preset threshold, the switching operation is delayed or canceled. A multi-objective optimization strategy is introduced to minimize the switching frequency while ensuring measurement accuracy, and the optimal switching strategy is selected through the Pareto optimal solution. The range switching signal, after physical constraints and optimization, is output, and a switching execution timing table is generated to guide the actual switching operation of the detector hardware.

[0045] Specifically, this embodiment constructs a physical constraint range switching prediction network, comprising a three-layer collaborative architecture of an evolution predictor, an event detector, and a constraint actuator. This network proactively predicts dose rate change trends and identifies abnormal events, resolving the data loss and measurement interruption issues caused by traditional passive range switching. Through multi-step long-sequence prediction (1-30 second prediction window) using a long short-term memory network and an environmental factor correlation analysis mechanism, the timeliness and accuracy of range switching are improved. This embodiment employs physical constraint verification, switching cost assessment, and multi-objective optimization strategies to ensure the physical feasibility and stability of range switching operations, providing reliable switching timing guidance and enhancing adaptability and measurement continuity assurance in complex dynamic environments.

[0046] The reconstruction processing of the fused dose rate data, range switching signal, measurement uncertainty, and flight vehicle position information yields ground dose rate distribution data, including: The fused dose rate data is processed for range continuity compensation based on the range switching signal, and error propagation calculation is performed based on the measurement uncertainty to obtain compensated and corrected dose rate data and overall uncertainty.

[0047] In one specific embodiment, a range switching compensation model is established, the range switching time is identified based on the range switching signal, and the dose rate data before and after the switching is continuously corrected to eliminate data jumps caused by range switching. A weighted moving average algorithm is used to smooth the dose rate data in the range switching interval, and the weighting coefficient is dynamically adjusted according to the reliability of the data before and after the switching. An uncertainty propagation chain model is established, and the measurement uncertainty, correction uncertainty, and fusion uncertainty are combined and calculated to obtain the comprehensive uncertainty; By introducing time correlation analysis, the uncertainty calculation results are corrected based on the time series characteristics of dose rate data, thereby improving the accuracy of uncertainty assessment.

[0048] Based on the flight vehicle location information, an high-altitude-ground spatial mapping model is established, and a physically constrained spatial reconstruction algorithm is used to reconstruct the compensated and corrected dose rate data to generate the ground dose rate distribution data.

[0049] In one specific embodiment, a three-dimensional radiative transfer model is constructed, and the geometric relationship between the high-altitude measurement point and the ground grid point is established based on the position information of the flight vehicle, and the attenuation coefficient of γ rays in the atmosphere is calculated. A Bayesian spatial interpolation algorithm is used, combined with the compensated and corrected dose rate data and the overall uncertainty, to probabilistically reconstruct the ground dose rate; A terrain-constrained optimization mechanism is introduced to correct the radiation transmission path based on ground elevation data, thereby improving the reconstruction accuracy of mountainous and complex terrain areas; A multi-scale adaptive mesh refinement strategy is established, which increases mesh density in high dose rate gradient regions and decreases mesh density in low gradient regions to optimize computational efficiency. The output includes the ground dose rate distribution data, including mean dose rate, uncertainty range, and confidence level information, and generates distribution quality evaluation indicators.

[0050] Specifically, this embodiment achieves data continuity assurance and accurate error propagation during range switching by constructing a range switching compensation model and an uncertainty propagation chain model. Through the weighted moving average algorithm and time correlation analysis mechanism, the smoothness of the data in the switching interval and the accuracy of uncertainty assessment are improved. At the same time, by adopting a physically constrained three-dimensional radiative transfer model and a Bayesian spatial interpolation algorithm, combined with terrain constraint optimization and multi-scale grid adaptive refinement strategy, a high-precision spatial mapping from high-altitude measurement data to ground dose rate distribution is achieved, providing high-confidence ground dose rate distribution information for nuclear emergency response and environmental monitoring.

[0051] The present invention also provides a high-altitude radioactive point source dose rate distribution analysis system, the system comprising: The multi-source data acquisition module is used to acquire raw counting data and energy spectrum data through multiple gamma-ray detectors, and to obtain environmental parameter data and flight vehicle position information; The Poisson statistical calculation module is used to perform statistical calculations on the original count data based on the statistical law of Poisson distribution to obtain the count statistical parameters; The physical sensing correction module is used to perform nonlinear correction on the original counting data, energy spectrum data, environmental parameter data and counting statistics parameters using a Poisson-constrained physical sensing residual neural network, so as to obtain the corrected detector data and measurement uncertainty. The fusion weight verification module is used to perform fusion calculation on the corrected detector data using a fusion method based on maximizing mutual information to obtain fusion weights, and to verify the fusion weights according to the physical consistency verifier to obtain fused dose rate data. The range prediction control module is used to perform prediction processing on the fused dose rate data and environmental parameter data using a physically constrained range switching prediction network to obtain a range switching signal. The spatial reconstruction mapping module is used to reconstruct the fused dose rate data, range switching signal, measurement uncertainty and flight vehicle position information to obtain ground dose rate distribution data.

[0052] Specifically, this embodiment of an upper-altitude radioactive point source dose rate distribution analysis system, through the construction of six functional modules, realizes the entire process from multi-source data acquisition to ground distribution reconstruction, solving the problems of inconsistent data transmission between modules, disjointed processing flow, and low system integration in traditional systems. The standardized interface of the multi-source data acquisition module and the statistical modeling of the Poisson statistical calculation module provide a high-quality data foundation for subsequent processing. The deep coupling of the physical sensing correction module and the fusion weight verification module achieves high-precision fusion of multi-detector data. The intelligent prediction capability of the range prediction control module and the three-dimensional modeling of the spatial reconstruction mapping module ensure the system's adaptability and reconstruction accuracy. Thus, a highly integrated, modular, and scalable upper-altitude radiation monitoring system integrating data acquisition, intelligent processing, predictive control, and spatial reconstruction is constructed, improving the accuracy and reliability of nuclear emergency response.

[0053] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for analyzing the dose rate distribution of a high-altitude radioactive point source, characterized in that, Includes the following steps: Raw counting data and energy spectrum data were collected using multiple gamma-ray detectors, along with environmental parameter data and flight vehicle location information. Statistical calculations were performed on the original count data based on the statistical law of Poisson distribution to obtain the count statistical parameters; A Poisson-constrained physical sensing residual neural network is used to perform nonlinear correction on the original counting data, energy spectrum data, environmental parameter data, and counting statistics parameters to obtain the corrected detector data and measurement uncertainty. The fused detector data is fused using a fusion method based on maximizing mutual information to obtain fusion weights. The fusion weights are then verified using a physical consistency verifier to obtain fused dose rate data. A physically constrained range switching prediction network is used to predict the fused dose rate data and environmental parameter data to obtain the range switching signal. The fused dose rate data, range switching signal, measurement uncertainty, and flight vehicle location information are reconstructed to obtain ground dose rate distribution data.

2. The method for analyzing dose rate distribution of a high-altitude radioactive point source as described in claim 1, characterized in that, The nonlinear correction includes: The original counting data, energy spectrum data, environmental parameter data and counting statistics parameters are input into the Poisson constrained linear layer for Poisson noise constraint processing, and the Poisson constrained features are output. The physical perception residual neural network includes a Poisson constrained linear layer, a physical information residual block and an uncertainty quantification block. The Poisson constraint features are input into the physical information residual block for physical constraint residual learning, and feature mapping is performed based on the physical laws of γ-ray transmission to output physical perception features. The physical sensing features are input into the uncertainty quantification block for Bayesian inference, and the corrected detector data and the measurement uncertainty are output. The Poisson-constrained linear layer constrains the network training process using the Poisson loss function.

3. The method for analyzing dose rate distribution of a high-altitude radioactive point source as described in claim 2, characterized in that, The physical information residual block employs a multi-detector collaborative correction mechanism, including: Construct a detector response difference matrix and establish cross-calibration relationships between detectors based on the differences in energy response characteristics between GM tube detectors and scintillator detectors; The high-altitude environment adaptive compensation algorithm is used to dynamically adjust the physical constraint parameters based on the real-time changes in flight altitude and atmospheric density. The physical constraint parameters include the atmospheric attenuation coefficient and the energy conversion factor. A spectral-guided weighting strategy is adopted, which assigns differentiated correction weights to γ-ray counts in different energy ranges based on the energy distribution characteristics of the spectral data. Establish a consistency verification mechanism among detectors, calculate the cross-correlation coefficient of the calibration results of different detectors, and perform a recalibration process when the cross-correlation coefficient is lower than a preset threshold.

4. The method for analyzing dose rate distribution of a high-altitude radioactive point source as described in claim 1, characterized in that, The process involves fusing the corrected detector data using a mutual information maximization-based fusion method to obtain fusion weights, and then verifying these fusion weights using a physical consistency verifier to obtain fused dose rate data, including: Calculate the mutual information matrix between the corrected detector data, and determine the initial fusion weights of each detector by maximizing the mutual information criterion; The initial fusion weights are verified using a multi-level physical consistency verification strategy, which includes energy conservation verification, spatial distribution consistency verification, and time series continuity verification. The initial fusion weights are adaptively adjusted based on the verification results. When the physical consistency deviation exceeds a preset threshold, the weights are reallocated. The fusion dose rate data is obtained through weighted fusion calculation, and fusion quality evaluation indicators are generated.

5. The method for analyzing dose rate distribution of a high-altitude radioactive point source as described in claim 4, characterized in that, The weight redistribution employs a detector dynamic reliability assessment algorithm, including: Establish a database of historical detector performance to record the measurement accuracy and stability of each detector under different environmental conditions; A reliability quantification model based on entropy weight theory is adopted to evaluate the real-time reliability level of the detector based on the information entropy change of the detector's current measurement data; The sliding window anomaly detection method is used to monitor the detector output data in real time, identify abnormal fluctuations in the detector output data, and calculate the severity of the anomaly. A reliability-weight mapping function is established to perform nonlinear mapping on the detector reliability assessment results, thereby obtaining the fusion weight correction coefficient. The fusion weight is then dynamically adjusted based on the fusion weight correction coefficient.

6. The method for analyzing dose rate distribution of a high-altitude radioactive point source as described in claim 1, characterized in that, The process of acquiring raw counting data and energy spectrum data through multiple gamma-ray detectors, and obtaining environmental parameter data and flight vehicle location information, includes: The configuration and initialization of multiple gamma-ray detectors mounted on the flight vehicle are performed. The multiple gamma-ray detectors include GM tube detectors and scintillator detectors. The operating voltage and counting threshold of the GM tube detectors are set, and the high voltage and energy calibration parameters of the scintillator detectors are set. Data is collected synchronously by the multiple gamma-ray detectors to obtain the raw count data and energy spectrum data of each detector, as well as environmental parameter data and flight vehicle position information.

7. The method for analyzing dose rate distribution of a high-altitude radioactive point source as described in claim 1, characterized in that, The statistical calculation of the original count data based on the Poisson distribution statistical law yields the count statistical parameters, including: Based on the statistical law of Poisson distribution, a Poisson distribution model is performed on the original count data to establish a Poisson statistical model of the count data; The Poisson statistical model is used to calculate the count mean, count variance, and statistical uncertainty of each detector, and the count statistics parameters are output.

8. The method for analyzing dose rate distribution of a high-altitude radioactive point source as described in claim 1, characterized in that, The physically constrained range switching prediction network performs prediction processing on the fused dose rate data and environmental parameter data to obtain the range switching signal, including: The range switching prediction network includes an evolution predictor, an event detector, and a constraint actuator. The evolution predictor and the event detector perform time-series predictive analysis on the fused dose rate data and environmental parameter data to identify dose rate change trends and abnormal events, and generate predictive range switching suggestion signals. The constraint actuator performs physical constraint verification on the predictive range switching suggestion signal, and outputs a constraint-optimized range switching signal based on the physical feasibility of range switching and system stability requirements.

9. The method for analyzing dose rate distribution of a high-altitude radioactive point source as described in claim 1, characterized in that, The reconstruction processing of the fused dose rate data, range switching signal, measurement uncertainty, and flight vehicle position information yields ground dose rate distribution data, including: The fused dose rate data is processed for range continuity compensation based on the range switching signal, and error propagation calculation is performed based on the measurement uncertainty to obtain compensated and corrected dose rate data and comprehensive uncertainty. Based on the flight vehicle location information, an high-altitude-ground spatial mapping model is established, and a physically constrained spatial reconstruction algorithm is used to reconstruct the compensated and corrected dose rate data to generate the ground dose rate distribution data.

10. A high-altitude radioactive point source dose rate distribution analysis system, used to perform a high-altitude radioactive point source dose rate distribution analysis method as described in any one of claims 1-9, characterized in that, The system includes: The multi-source data acquisition module is used to acquire raw counting data and energy spectrum data through multiple gamma-ray detectors, and to obtain environmental parameter data and flight vehicle position information; The Poisson statistical calculation module is used to perform statistical calculations on the original count data based on the statistical law of Poisson distribution to obtain the count statistical parameters; The physical sensing correction module is used to perform nonlinear correction on the original counting data, energy spectrum data, environmental parameter data and counting statistics parameters using a Poisson-constrained physical sensing residual neural network, so as to obtain the corrected detector data and measurement uncertainty. The fusion weight verification module is used to perform fusion calculation on the corrected detector data using a fusion method based on maximizing mutual information to obtain fusion weights, and to verify the fusion weights according to the physical consistency verifier to obtain fused dose rate data. The range prediction control module is used to perform prediction processing on the fused dose rate data and environmental parameter data using a physically constrained range switching prediction network to obtain a range switching signal. The spatial reconstruction mapping module is used to reconstruct the fused dose rate data, range switching signal, measurement uncertainty and flight vehicle position information to obtain ground dose rate distribution data.

Images