Radiation detection method and system based on big data analysis

Abstract

The invention provides a radiation detection method and system based on big data analysis, and relates to the technical field of data analysis, and the method comprises the steps: obtaining radiation detection data in a target region; the method comprises the following steps: performing preprocessing and feature extraction on radiation detection data to obtain a feature data set, and segmenting the feature data set through a super-plane with a random slope to construct an isolated tree so as to establish a target isolated forest; the target isolated forest comprises a plurality of isolated sub-forests with incompletely same model parameters; and screening out outliers in the radiation detection data by using a target isolated forest, obtaining the total number of times of marking the outliers by the isolated sub-forest, determining the outliers with the total number of times greater than a preset threshold as target outliers, and determining the target sub-regions corresponding to the target outliers as abnormal radiation regions. According to the method, abnormal outliers are identified through the improved isolated forests, anomaly detection is carried out on the same data set through a plurality of isolated forest models, and the radiation detection precision is improved.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a radiation detection method and system based on big data analysis. Background Art

[0002] With the widespread application of nuclear energy and nuclear technology in energy, medicine, industry, and other fields, radiation safety issues are gaining increasing attention. As a crucial means of ensuring radiation safety, the demand for radiation detection is also increasing. Isolation Forest is an anomaly detection algorithm based on the core concept that "anomaly data are small, distinct observations that are more easily isolated." In radiation data scenarios, normal radiation data often follows a certain distribution pattern, while outliers deviate significantly from this pattern. The idea of ​​randomly selecting slopes has been mentioned in the context of Isolation Forest. However, the standard Isolation Forest algorithm produces artifacts when generating anomaly score heatmaps due to the standard binary tree branching operation. These artifacts manifest as unusual distributions of anomaly scores in certain directions, which in turn affects the accuracy of anomaly detection. Summary of the Invention

[0003] In view of this, the present invention proposes a radiation detection method and system based on big data analysis.

[0004] The technical solution of the present invention is implemented as follows: In a first aspect, the present invention provides a radiation detection method based on big data analysis, comprising: Acquire radiation detection data within the target area; Preprocessing and feature extraction are performed on the radiation detection data to obtain a feature data set, and the feature data set is segmented by a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest; the target isolation forest includes a plurality of isolated sub-forests with different model parameters; the model parameters include at least one of the number of isolated trees, the subsampling size, and the maximum depth of the tree; The target isolation forest is used to screen out outliers in the radiation detection data, and the total number of times the outliers are marked by the isolation forest is obtained. The outliers whose total number is greater than a preset threshold are determined as target outliers, and the target sub-area corresponding to the target outlier is determined as the abnormal radiation area.

[0005] On the basis of the above technical solution, preferably, the preprocessing and feature extraction of the radiation detection data to obtain a feature data set includes: performing data cleaning and missing value filling processing on the radiation detection data to obtain first radiation detection data; Converting the first radiation detection data into a unified format and dimension, and associating the data with corresponding geographic location information and time information to obtain second radiation detection data; Visual analysis is performed on the second radiation detection data, and feature extraction is performed to obtain a feature data set including spatial distribution features and temporal distribution features.

[0006] Based on the above technical solution, preferably, the step of segmenting the feature data set by a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest, includes: A binary search tree is constructed using a hyperplane with a random slope and the feature data set; the normal vector of the hyperplane is uniformly selected on the N-dimensional unit sphere, and the intercept is uniformly selected from the data range of the current branch node; For each binary search tree, randomly selecting an intermediate feature and a random value within the range of the intermediate feature from the feature data set as a branch point, dividing the radiation detection data into two parts until each detection data in the radiation detection data becomes a leaf node, and repeatedly constructing multiple binary search trees to form an isolated sub-forest; the intermediate feature includes at least one of an average radiation dose and a maximum radiation intensity; A target isolation forest is determined based on the plurality of isolated sub-forests.

[0007] Based on the above technical solution, preferably, for each binary search tree, randomly selecting an intermediate feature and a random value within the range of the intermediate feature from the feature data set as a branch point, dividing the radiation detection data into two parts until each detection data in the radiation detection data becomes a leaf node, including: The data in the radiation detection data that is smaller than the hyperplane is divided into the left branch, and the data that is larger than the hyperplane is divided into the right branch. The binary branching operation is repeated on the left and right branches until each detection data in the radiation detection data becomes a leaf node.

[0008] Based on the above technical solution, preferably, the step of segmenting the feature data set by a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest, includes: Parallel computing technology is used to independently train multiple isolated sub-forests simultaneously, and the parameters of each isolated sub-forest are tuned in combination with the Bayesian optimization method to establish a target isolated forest.

[0009] Based on the above technical solution, preferably, the solitary sub-forest includes a first solitary sub-forest and a second solitary sub-forest having different maximum tree depths; and the method of using the target solitary sub-forest to screen out outliers in the radiation detection data, obtaining a total number of times the outliers are marked by the solitary sub-forest, and determining the outliers whose total number is greater than a preset threshold as target outliers includes: Common outliers detected by the first solitary sub-forest and the second solitary sub-forest are obtained, and the common outliers are determined as target outliers.

[0010] Based on the above technical solution, preferably, the step of using the target isolation forest to screen out outliers in the radiation detection data, obtaining the total number of times the outliers are marked by the isolation sub-forest, and determining the outliers whose total number is greater than a preset threshold as target outliers includes: Each of the isolated sub-forests is used to perform anomaly detection on the feature data set, and the number of times each data point in the feature data set is marked as an outlier is recorded, the total number of times each data point is marked as an outlier in all isolated sub-forests is determined, and the outliers whose total number is greater than a preset threshold are determined as target outliers.

[0011] More preferably, the second aspect of the present invention provides a radiation detection system based on big data analysis, comprising: a data acquisition module, a data segmentation module and an anomaly determination module; wherein, The data acquisition module is configured to acquire radiation detection data within the target area; The data segmentation module is configured to preprocess and extract features from the radiation detection data to obtain a feature data set, and segment the feature data set using a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest; the target isolation forest includes a plurality of isolated sub-forests with different model parameters; the model parameters include at least one of the number of isolated trees, the subsampling size, and the maximum depth of the tree; The anomaly determination module is configured to use the target isolation forest to filter out outliers in the radiation detection data, obtain the number of times the outliers are marked by all the isolation sub-forests, determine the outliers with a marking number greater than a preset threshold as target outliers, and determine the target sub-area corresponding to the target outlier as the abnormal radiation area.

[0012] More preferably, the third aspect of the present invention provides an electronic device comprising a processor and a memory; the memory stores a computer program, wherein the computer program, when executed by the processor, implements the radiation detection method based on big data analysis described in the first aspect.

[0013] More preferably, the fourth aspect of the present invention provides a computer storage medium on which a computer program is stored, wherein when the computer program is executed by a processor, it implements the radiation detection method based on big data analysis described in the first aspect.

[0014] The radiation detection method and system based on big data analysis of the present invention have the following beneficial effects compared with the prior art: 1. This method uses a random-slope hyperplane to segment the feature dataset and construct an isolation tree. It then builds multiple isolation subforests with varying model parameters, using these multiple isolation subforests to detect anomalies on the same dataset. Random-slope hyperplanes can segment data at arbitrary angles, allowing for more flexible capture of complex patterns and anomalies. By varying model parameters and constructing multiple subforests, it not only eliminates bias in branching operations but also reduces overfitting of the model to the training data, significantly improving anomaly detection accuracy.

[0015] 2. Utilize geographic location information and time information to analyze the differences in radiation levels in different regions, identify radiation hotspots, and analyze the temporal trend of radiation levels to predict future radiation conditions. On this basis, feature extraction is used to convert the original data into a feature dataset that is more suitable for model training, and a more accurate isolated sub-forest is constructed to provide support for radiation safety management and decision-making.

[0016] 3. By varying model parameters to create multiple isolated subforests, the model reduces their bias toward specific data distributions or unusual patterns, thereby improving their generalization ability on unknown data. The randomness introduced into the construction of multiple isolated subforests makes them more robust to anomalies in the data. Even if one isolated subforest is affected by unusual data, the remaining isolated subforests are likely to still provide accurate detection results, ensuring the reliability and accuracy of the final detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A schematic diagram of a process flow of a radiation detection method based on big data analysis provided by an embodiment of the present invention; Figure 2 A schematic diagram of the principle of establishing a target isolation forest according to an embodiment of the present invention; Figure 3A schematic structural diagram of a radiation detection device based on big data analysis provided by an embodiment of the present invention; Figure 4 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] In some embodiments, as Figure 1 As shown, Figure 1 A schematic flow chart of a radiation detection method based on big data analysis provided by an embodiment of the present invention; a radiation detection method based on big data analysis provided by the present invention includes: S110, obtaining radiation detection data within the target area.

[0021] Typically, a radiation sensor network is strategically placed within the target area to ensure coverage of key detection points. These sensors are connected to the data collector via wired or wireless connections. Raw irradiance data from each acquisition interface is acquired at preset intervals (e.g., 200ms) and sent to the analog-to-digital conversion module. The module converts the analog signal into a digital signal, generating intermediate irradiance data and storing it in the module's memory. Calculations are performed on the intermediate irradiance data based on internally stored sensitivity parameters to generate target irradiance data.

[0022] S120, preprocessing and feature extraction of radiation detection data to obtain a feature data set, and segmenting the feature data set through a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest; the target isolation forest includes multiple isolated sub-forests with different model parameters; the model parameters include at least one of the number of isolated trees, the subsampling size, and the maximum depth of the tree.

[0023] In this embodiment, when constructing an isolation tree, a hyperplane with a random slope is used to segment the feature dataset. By randomly selecting segmentation features and split points, the randomness of each segmentation is ensured, enhancing model diversity. The model parameters (such as the number of isolated trees, subsampling size, and maximum tree depth) vary for each isolated subforest. By adjusting the number of trees, model complexity and computational efficiency can be balanced. Controlling the number of samples used in each isolation tree construction introduces randomness and reduces the risk of overfitting. By limiting the maximum tree depth, the model is prevented from becoming overly complex. The detection results of multiple isolated subforests are integrated, combining their strengths and improving the accuracy and robustness of anomaly detection.

[0024] In some embodiments, S120, preprocessing and feature extraction of the radiation detection data to obtain a feature data set includes: Performing data cleaning and missing value filling processing on the radiation detection data to obtain first radiation detection data; Converting the first radiation detection data into a unified format and dimension, and associating the data with corresponding geographic location information and time information to obtain second radiation detection data; Visual analysis is performed on the second radiation detection data, and feature extraction is performed to obtain a feature data set including spatial distribution features and temporal distribution features.

[0025] The data cleaning process can remove noise, erroneous records, and outliers from the data, improving its accuracy and reliability. By using appropriate missing value filling methods (such as mean filling, median filling, interpolation, or model-based prediction filling), more valid data can be retained, avoiding analytical bias caused by missing data. Converting data to a unified format and dimension facilitates subsequent data integration, comparison, and analysis. Correlating radiation detection data with geographic location and time information can reveal the spatial distribution and temporal variation of radiation levels, providing more comprehensive information support for decision-making. Visualization techniques (such as heat maps, line charts, and scatter plots) can intuitively display the spatial distribution and temporal variation characteristics of radiation data. Based on this, feature extraction is performed to convert the raw data into a feature dataset more suitable for model training, thereby improving the performance of the solitary sub-forest model.

[0026] In some embodiments, S120, segmenting the feature dataset using a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest, includes: A binary search tree is constructed using a hyperplane with a random slope and a feature data set; the normal vector of the hyperplane is uniformly selected on the N-dimensional unit sphere, and the intercept is uniformly selected from the data range of the current branch node; For each binary search tree, an intermediate feature and a random value within the range of the intermediate feature are randomly selected from the feature data set as a branch point, the radiation detection data is divided into two parts until each detection data in the radiation detection data becomes a leaf node, and multiple binary search trees are repeatedly constructed to form an isolated sub-forest; the intermediate feature includes at least one of the average radiation dose and the maximum radiation intensity; Determine the target isolation forest based on multiple isolated subforests.

[0027] For example, see Figure 2 , Figure 2 A schematic diagram of the principle of establishing a target isolation forest provided by an embodiment of the present invention; given an N-dimensional dataset, a random subsample of the data is selected to construct a binary tree. The tree branching process occurs by selecting a random dimension x_i (i∈{1,2,...,N}) of the data. Then, a random value v is selected between the minimum and maximum values ​​of this dimension. If the value of a given data point on this dimension is less than v, the point is sent to the left branch; otherwise, it is sent to the right branch. In this way, the data at the tree node is divided into two parts. This branching process is recursively performed on the dataset until a single point is isolated or a predetermined depth limit is reached. Then, another random tree is constructed using the new random subsample. After a large number of trees (i.e., a forest) are constructed, training is complete. Figure 2 a (left) shows a trained tree. The red line depicts the trajectory of an outlier point as it descends through the tree, while the blue line shows the trajectory of a nominal point. The outlier point is isolated quickly, but the nominal point continues to descend to the maximum depth. Figure 2 In b (right), a complete forest of 60 trees is shown. Each radial line represents a tree, and the outer circle represents the maximum depth limit. The red line shows the trajectory of a single outlier point as it descends through each tree, while the blue line shows the trajectory of a nominal point. On average, the blue line achieves a much larger radius than the red line, allowing for better separation between outliers and nominal points.

[0028] The definition of an isolated sub-forest can be: .

[0029] in, represents the average depth of sample x in multiple itrees, E represents the mathematical expectation, It represents the average path length when the binary tree search is unsuccessful, and n represents the number of samples used to generate each itree. , .

[0030] In some embodiments, for each binary search tree, an intermediate feature and a random value within the range of the intermediate feature are randomly selected from the feature data set as a branch point, and the radiation detection data is divided into two parts until each detection data in the radiation detection data becomes a leaf node, including: The data in the radiation detection data that is smaller than the hyperplane is moved to the left branch, and the data that is larger than the hyperplane is moved to the right branch. The binary branching operation is repeated on the left and right branches until each detection data in the radiation detection data becomes a leaf node.

[0031] In some embodiments, S120, segmenting the feature dataset using a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest, includes: Parallel computing technology is used to train multiple isolated sub-forests independently and simultaneously, and the parameters of each isolated sub-forest are tuned in combination with the Bayesian optimization method to establish the target isolated forest.

[0032] Parallel computing technology allows for the simultaneous and independent training of multiple isolated forests, significantly reducing overall training time and fully utilizing computing resources, avoiding idle and wasted resources. Bayesian optimization methods efficiently search for optimal parameter combinations by constructing a posterior probability distribution of the objective function (such as anomaly detection accuracy). After independently tuning the parameters of each isolated forest, there will be some variation between the models. This diversity helps enhance the robustness and generalization ability of the overall isolated forest, enabling it to better adapt to complex and changing radiation environments. These forests complement each other in the anomaly detection process, collectively improving detection accuracy.

[0033] S130, using the target isolation forest to filter out outliers in the radiation detection data, and obtaining the total number of times the outliers are marked by the isolation forest, determining the outliers whose total number is greater than a preset threshold as target outliers, and determining the target sub-area corresponding to the target outlier as the abnormal radiation area.

[0034] In some embodiments, the solitary sub-forest includes a first solitary sub-forest and a second solitary sub-forest having different maximum tree depths; S130, using the target solitary sub-forest to screen out outliers in the radiation detection data, obtaining a total number of times the outliers are marked by the solitary sub-forests, and determining outliers whose total number of times is greater than a preset threshold as target outliers, including: The common outliers detected by the first solitary subforest and the second solitary subforest are obtained, and the common outliers are determined as target outliers.

[0035] Here, after independent tuning, the model parameters of the first and second solitary subforests differ somewhat. The first solitary subforest is used to detect anomalies in the feature dataset, recording the detected outliers. The second solitary subforest is used to detect anomalies in the same feature dataset, also recording the detected outliers. The detection results of the first and second solitary subforests are compared to identify outliers that are detected in common by both. These common outliers are designated as target outliers, which have higher credibility and importance.

[0036] In some embodiments, S130, using the target isolation forest to screen out outliers in the radiation detection data, obtaining the total number of times the outliers are marked by the isolation forest, and determining the outliers whose total number of times is greater than a preset threshold as target outliers, includes: Each isolated sub-forest is used to perform anomaly detection on the feature data set, and the number of times each data point in the feature data set is marked as an outlier is recorded. The total number of times each data point is marked as an outlier in all isolated sub-forests is determined, and the outliers whose total number is greater than the preset threshold are identified as target outliers.

[0037] In this embodiment, each solitary sub-forest is used to perform anomaly detection on the feature dataset, recording the number of times each data point is marked as an outlier. For each data point in the feature dataset, the total number of times it is marked as an outlier across all solitary sub-forests is counted. A preset threshold is set, and outliers with a total number of times greater than the threshold are identified as target outliers. While a single model may have false positives or false negatives, statistically analyzing the detection results of multiple models can effectively reduce these occurrences. Data points with a high total number of times are more likely to be true anomalies.

[0038] In some embodiments, see Figure 3 , Figure 3 The radiation detection system 300 based on big data analysis provided by the embodiment of the present invention includes: a data acquisition module 310, a data segmentation module 320 and an abnormality determination module 330; wherein, A data acquisition module 310 is configured to acquire radiation detection data within a target area; The data segmentation module 320 is configured to preprocess and extract features from the radiation detection data to obtain a feature data set, and segment the feature data set using a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest; the target isolation forest includes a plurality of isolated sub-forests with different model parameters; the model parameters include at least one of the number of isolated trees, the subsampling size, and the maximum depth of the tree; The anomaly determination module 330 is configured to use the target isolation forest to filter out outliers in the radiation detection data, and obtain the number of times the outliers are marked by all the isolation sub-forests, determine the outliers with a marking number greater than a preset threshold as target outliers, and determine the target sub-area corresponding to the target outlier as the abnormal radiation area.

[0039] In some embodiments, the data segmentation module 320 is specifically configured to: Performing data cleaning and missing value filling processing on the radiation detection data to obtain first radiation detection data; Converting the first radiation detection data into a unified format and dimension, and associating the data with corresponding geographic location information and time information to obtain second radiation detection data; Visual analysis is performed on the second radiation detection data, and feature extraction is performed to obtain a feature data set including spatial distribution features and temporal distribution features.

[0040] In some embodiments, the data segmentation module 320 is specifically configured to: A binary search tree is constructed using a hyperplane with a random slope and a feature data set; the normal vector of the hyperplane is uniformly selected on the N-dimensional unit sphere, and the intercept is uniformly selected from the data range of the current branch node; For each binary search tree, an intermediate feature and a random value within the range of the intermediate feature are randomly selected from the feature data set as a branch point, the radiation detection data is divided into two parts until each detection data in the radiation detection data becomes a leaf node, and multiple binary search trees are repeatedly constructed to form an isolated sub-forest; the intermediate feature includes at least one of the average radiation dose and the maximum radiation intensity; Determine the target isolation forest based on multiple isolated subforests.

[0041] In some embodiments, the data segmentation module 320 is specifically configured to: The data in the radiation detection data that is smaller than the hyperplane is moved to the left branch, and the data that is larger than the hyperplane is moved to the right branch. The binary branching operation is repeated on the left and right branches until each detection data in the radiation detection data becomes a leaf node.

[0042] In some embodiments, the data segmentation module 320 is specifically configured to: Parallel computing technology is used to train multiple isolated sub-forests independently and simultaneously, and the parameters of each isolated sub-forest are tuned in combination with the Bayesian optimization method to establish the target isolated forest.

[0043] In some embodiments, the solitary sub-forest includes a first solitary sub-forest and a second solitary sub-forest having different maximum tree depths; the abnormality determination module 330 is specifically configured as follows: The common outliers detected by the first solitary subforest and the second solitary subforest are obtained, and the common outliers are determined as target outliers.

[0044] In some embodiments, the abnormality determination module 330 is specifically configured to: Each isolated sub-forest is used to perform anomaly detection on the feature data set, and the number of times each data point in the feature data set is marked as an outlier is recorded. The total number of times each data point is marked as an outlier in all isolated sub-forests is determined, and the outliers whose total number is greater than the preset threshold are identified as target outliers.

[0045] It should be noted that the radiation detection system based on big data analysis provided in the embodiment of the present application and the radiation detection method based on big data analysis provided in the embodiment of the present application are based on the same application concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned radiation detection method based on big data analysis, and the repetitive parts will not be repeated.

[0046] In some embodiments, see Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. An electronic device 400 provided in an embodiment of the present application includes a processor 410 and a memory 420; the memory 420 stores a computer program, wherein the computer program, when executed by the processor, implements the above-mentioned radiation detection method based on big data analysis.

[0047] Specifically, the processor 410 may include, for example, a general-purpose microprocessor, an instruction set processor and / or a related chipset and / or a dedicated microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 410 may also include onboard memory for caching purposes. The processor 410 may be a single processing unit or multiple processing units for executing different actions of the method flow according to the embodiments of the present application.

[0048] Memory 420 can be, for example, any medium capable of containing, storing, conveying, disseminating, or transmitting instructions. For example, memory 420 can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, components, or communication media. Specific examples of memory 420 include: magnetic storage devices, such as magnetic tape or hard disk drives (HDDs); optical storage devices, such as compact discs (CD-ROMs); random access memory (RAM) or flash memory; and / or wired or wireless communication links.

[0049] This application also provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned radiation detection method based on big data analysis. This computer-readable medium may be included in the device / apparatus / system described in the aforementioned embodiments, or it may exist independently and not be incorporated into the device / apparatus / system. The aforementioned computer-readable medium carries one or more programs, and when executed, implements the method according to the embodiments of this application.

[0050] According to an embodiment of the present application, a computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or component. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, radio frequency signals, or any suitable combination thereof.

[0051] Those skilled in the art will understand that the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present application. Therefore, the scope of the present application should not be limited to the above-mentioned embodiments, but should be determined not only by the attached claims, but also by the equivalents of the attached claims. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A radiation detection method based on big data analysis, characterized in that: include: Acquire radiation detection data within the target area; Preprocessing and feature extraction are performed on the radiation detection data to obtain a feature data set, and the feature data set is segmented by a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest; the target isolation forest includes a plurality of isolated sub-forests with different model parameters; the model parameters include at least one of the number of isolated trees, the subsampling size, and the maximum depth of the tree; The target isolation forest is used to screen out outliers in the radiation detection data, and the total number of times the outliers are marked by the isolation forest is obtained. The outliers whose total number is greater than a preset threshold are determined as target outliers, and the target sub-area corresponding to the target outlier is determined as the abnormal radiation area.

2. The radiation detection method based on big data analysis according to claim 1, characterized in that: The preprocessing and feature extraction of the radiation detection data to obtain a feature data set includes: performing data cleaning and missing value filling processing on the radiation detection data to obtain first radiation detection data; Converting the first radiation detection data into a unified format and dimension, and associating the data with corresponding geographic location information and time information to obtain second radiation detection data; Visual analysis is performed on the second radiation detection data, and feature extraction is performed to obtain a feature data set including spatial distribution features and temporal distribution features.

3. The radiation detection method based on big data analysis according to claim 1, characterized in that: The step of segmenting the feature data set by a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest, includes: A binary search tree is constructed using a hyperplane with a random slope and the feature data set; the normal vector of the hyperplane is uniformly selected on the N-dimensional unit sphere, and the intercept is uniformly selected from the data range of the current branch node; For each binary search tree, randomly selecting an intermediate feature and a random value within the range of the intermediate feature from the feature data set as a branch point, dividing the radiation detection data into two parts until each detection data in the radiation detection data becomes a leaf node, and repeatedly constructing multiple binary search trees to form an isolated sub-forest; the intermediate feature includes at least one of an average radiation dose and a maximum radiation intensity; A target isolation forest is determined based on the plurality of isolated sub-forests.

4. The radiation detection method based on big data analysis according to claim 3, characterized in that: For each binary search tree, randomly selecting an intermediate feature and a random value within the range of the intermediate feature from the feature data set as a branch point, dividing the radiation detection data into two parts until each detection data in the radiation detection data becomes a leaf node, including: The data in the radiation detection data that is smaller than the hyperplane is divided into the left branch, and the data that is larger than the hyperplane is divided into the right branch. The binary branching operation is repeated on the left and right branches until each detection data in the radiation detection data becomes a leaf node.

5. The radiation detection method based on big data analysis according to claim 1, characterized in that: The step of segmenting the feature data set by a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest, includes: Parallel computing technology is used to independently train multiple isolated sub-forests simultaneously, and the parameters of each isolated sub-forest are tuned in combination with the Bayesian optimization method to establish a target isolated forest.

6. The radiation detection method based on big data analysis according to claim 5, characterized in that: The solitary sub-forest includes a first solitary sub-forest and a second solitary sub-forest with different maximum tree depths; and using the target solitary sub-forest to screen out outliers in the radiation detection data, obtaining a total number of times the outliers are marked by the solitary sub-forest, and determining outliers whose total number is greater than a preset threshold as target outliers, includes: Common outliers detected by the first solitary sub-forest and the second solitary sub-forest are obtained, and the common outliers are determined as target outliers.

7. The radiation detection method based on big data analysis according to claim 1, characterized in that: The method of using the target isolation forest to screen out outliers in the radiation detection data, obtaining a total number of times the outliers are marked by the isolation forest, and determining an outlier with a total number greater than a preset threshold as a target outlier includes: Each of the isolated sub-forests is used to perform anomaly detection on the feature data set, and the number of times each data point in the feature data set is marked as an outlier is recorded, the total number of times each data point is marked as an outlier in all isolated sub-forests is determined, and the outliers whose total number is greater than a preset threshold are determined as target outliers.

8. A radiation detection system based on big data analysis, characterized in that: include: Data acquisition module, data segmentation module and abnormality determination module; wherein, The data acquisition module is configured to acquire radiation detection data within the target area; The data segmentation module is configured to preprocess and extract features from the radiation detection data to obtain a feature data set, and segment the feature data set using a hyperplane with a random slope to construct an isolation tree, thereby establishing a target isolation forest; the target isolation forest includes a plurality of isolated sub-forests with different model parameters; the model parameters include at least one of the number of isolated trees, the subsampling size, and the maximum depth of the tree; The anomaly determination module is configured to use the target isolation forest to filter out outliers in the radiation detection data, obtain the number of times the outliers are marked by all the isolation sub-forests, determine the outliers with a marking number greater than a preset threshold as target outliers, and determine the target sub-area corresponding to the target outlier as the abnormal radiation area.

9. An electronic device comprising a processor and a memory; the memory stores a computer program, wherein: When the computer program is executed by the processor, the computer program implements the radiation detection method based on big data analysis according to any one of claims 1 to 7.

10. A computer storage medium, characterized in that A computer program is stored thereon, wherein when the computer program is executed by a processor, the radiation detection method based on big data analysis as described in any one of claims 1 to 7 is implemented.

Images