Gas data anomaly detection method and device

Abstract

The invention provides a gas data anomaly detection method and device. The method comprises the steps that a to-be-detected gas data set is acquired; performing granularity division on the gas data set on the basis of the joint information entropy obtained by calculation of the gas data set to determine a granularity identifier to which each piece of gas data in the gas data set belongs; performing feature extraction on the plurality of gas data belonging to the same type of granularity identification to determine corresponding target features; determining an abnormal probability value sequence corresponding to the gas data set through a pre-trained gas abnormal detection model according to the target features corresponding to the plurality of gas data belonging to the various granularity identifiers; wherein each abnormal probability value in the abnormal probability value sequence is used for indicating the probability that the corresponding gas data in the gas data set is abnormal, so that a large amount of time cost does not need to be consumed, the gas data is marked in advance, and the detection accuracy and reliability of the gas anomaly detection model are improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method and device for detecting gas data anomaly. Background Art

[0002] Gas is a very dangerous gas. In actual scenarios, the dynamic changes of gas are directly related to the safe production of the entire coal mining. Abnormal accumulation of gas may not only lead to gas explosions, causing significant casualties and property losses, but may also cause other serious safety accidents such as gas outbursts. Therefore, accurate and efficient anomaly detection of gas data has always been one of the core tasks of coal mine safety assurance work.

[0003] In related technologies, machine learning models are usually established to detect gas data. However, the gas data collected in actual working scenarios is often affected by multiple environmental factors and does not present the characteristics of an ideal linear distribution. The subsequent use of machine learning models to detect anomalies in gas data will affect the accuracy of the machine learning model in detecting abnormal gas data. Summary of the invention

[0004] The present application aims to solve one of the technical problems in the related art at least to some extent.

[0005] To this end, the first purpose of this application is to propose a gas data anomaly detection method to provide high-quality gas data for the gas anomaly detection model, thereby greatly improving the accuracy and reliability of the gas anomaly detection model.

[0006] The second objective of the present application is to provide a gas data anomaly detection device.

[0007] The third objective of the present application is to provide an electronic device.

[0008] A fourth objective of the present application is to provide a computer-readable storage medium.

[0009] A fifth object of the present application is to provide a computer program product.

[0010] To achieve the above-mentioned purpose, the first aspect of the present application proposes a method for detecting anomalies in gas data, comprising: obtaining a gas data set to be detected; wherein the gas data set includes multiple gas data collected within at least one time period in the mining area to be detected; based on the joint information entropy calculated from the gas data set, granularly dividing the gas data set to determine the granularity identifier to which each of the gas data in the gas data set belongs; for any multiple gas data belonging to the same type of granularity identifier, feature extraction is performed on the multiple gas data belonging to the same type of granularity identifier to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identifier; according to the target features corresponding to the multiple gas data belonging to each type of granularity identifier in the gas data set, a pre-trained gas anomaly detection model is used to determine a sequence of anomaly probability values ​​corresponding to the gas data set; wherein each anomaly probability value in the anomaly probability value sequence is used to indicate the probability of gas data anomaly occurring in the corresponding gas data in the gas data set.

[0011] To achieve the above-mentioned purpose, the second embodiment of the present application proposes a gas data anomaly detection device, comprising: a first acquisition module, used to acquire a gas data set to be detected; wherein the gas data set includes multiple gas data collected within at least one time period of the mining area to be detected; a granularity division module, used to granularize the gas data set based on the joint information entropy calculated from the gas data set to determine the granularity identifier to which each of the gas data in the gas data set belongs; a first determination module, used to extract features from the multiple gas data belonging to the same type of granularity identifier for any multiple gas data belonging to the same type of granularity identifier, so as to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identifier; a second determination module, used to determine the abnormal probability value sequence corresponding to the gas data set through a pre-trained gas anomaly detection model according to the target features corresponding to the multiple gas data belonging to each type of granularity identifier in the gas data set; wherein each abnormal probability value in the abnormal probability value sequence is used to indicate the probability of gas data anomaly occurring in the corresponding gas data in the gas data set.

[0012] To achieve the above-mentioned purpose, the third aspect embodiment of the present application proposes an electronic device, comprising: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the gas data anomaly detection method disclosed in the embodiment of the present application.

[0013] To achieve the above-mentioned purpose, the fourth aspect embodiment of the present application proposes a computer-readable storage medium, in which computer execution instructions are stored. When the computer execution instructions are executed by a processor, they are used to implement the gas data anomaly detection method disclosed in the embodiment of the present application.

[0014] To achieve the above-mentioned purpose, the fifth aspect of the present application proposes a computer program product, including a computer program, which, when executed by a processor, implements the gas data anomaly detection method disclosed in the embodiment of the present application.

[0015] The gas data anomaly detection method and device provided in the present application obtain a gas data set to be detected; wherein the gas data set includes multiple gas data collected within at least one time period of the mining area to be detected; based on the joint information entropy calculated from the gas data set, the gas data set is granularized to determine the granularity identification to which each gas data in the gas data set belongs; for any multiple gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple gas data belonging to the same type of granularity identification to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification; according to the target features corresponding to the multiple gas data belonging to each type of granularity identification in the gas data set, a pre-trained gas anomaly detection model is used to determine the abnormal probability value sequence corresponding to the gas data set; wherein each abnormal probability value in the abnormal probability value sequence is used to indicate the gas data. The probability of gas data anomalies occurring in the corresponding gas data in the gas data set is determined by the gas anomaly detection model. Therefore, before determining the probability of gas data anomalies occurring in the gas data set through the gas anomaly detection model, not only the various factors that may affect the abnormal situation of gas data in the actual working scenario are comprehensively considered, but also the nonlinear characteristics of the distribution of multiple gas data in the gas data set are considered. Therefore, the gas data set is comprehensively divided into granularities not only from the perspective of local fluctuations of gas data but also from the perspective of the macro trend of gas data. Then, based on the gas data belonging to various granularity labels, the probability of gas data anomalies is determined through the gas anomaly detection model. Without consuming a lot of time cost, the gas data can be pre-labeled to provide high-quality gas data for the gas anomaly detection model, which greatly improves the accuracy and reliability of the gas anomaly detection model.

[0016] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 A flow chart of a gas data anomaly detection method provided in an embodiment of the present application Figure 1 ; Figure 2 A flow chart of a gas data anomaly detection method provided in an embodiment of the present application Figure 2 ; Figure 3 An example of a gas data anomaly detection method provided in the present application Figure 3 ; Figure 4 A flow chart of a gas data anomaly detection method provided in an embodiment of the present application Figure 4 ; Figure 5 A schematic diagram of a framework corresponding to the gas data anomaly detection process provided in an embodiment of the present application; Figure 6 An architectural diagram of a gas anomaly detection model provided in an embodiment of the present application; Figure 7 A schematic diagram of the structure of a gas data anomaly detection device provided in an embodiment of the present application; Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0019] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0020] Figure 1 A flow chart of a gas data anomaly detection method provided in an embodiment of the present application Figure 1 .

[0021] Step 101, obtaining a gas data set to be detected; wherein the gas data set includes a plurality of gas data collected within at least one time period in the mining area to be detected.

[0022] It should be noted that the gas data anomaly detection method provided in the embodiment of the present application can be performed by a gas data anomaly detection device (also referred to as a gas data anomaly detector), wherein the gas data anomaly detection device can be implemented by software and / or hardware. The gas data anomaly detection device can be an electronic device, or can be configured in an electronic device to implement the gas data anomaly detection function.

[0023] The embodiment of the present application is described by taking the gas data anomaly detection method configured in an electronic device as an example.

[0024] Among them, the electronic device can be any device with computing capabilities, such as a personal computer, a mobile terminal, a server, etc. The mobile terminal can be, for example, a vehicle-mounted device, a mobile phone, a tablet computer, a personal digital assistant, a wearable device, and other hardware devices with various operating systems, touch screens and / or display screens.

[0025] It should be noted that the specific information of the mining areas to be detected and the specific number of mining areas are determined according to actual conditions, and this embodiment does not make specific limitations. For example, the mining areas to be detected may be coal mining area A and coal mining area B.

[0026] It should be noted that the specific information of the time period is determined according to actual needs, and this embodiment does not make any specific limitations. For example, a time period of the mining area to be detected can be the entire time period of February 2024 in coal mining area A.

[0027] It should be understood that in order to ensure the stability of the production environment during the mining process, the collection of gas data is a long and continuous process. The detection of gas data is not only the detection of gas data at a certain moment, but also the detection of the overall fluctuation trend of gas data at each moment within a period of time. Therefore, gas data at each moment in different time periods can be collected in different time periods, and then the gas data set can be determined.

[0028] Step 102 , based on the joint information entropy calculated for the gas data set, the gas data set is divided into granularities to determine the granularity identifier to which each gas data in the gas data set belongs.

[0029] It is important to understand that since each gas data in the gas data set has various characteristics at different times and in different environments, for example, even in the same mining environment, the gas data collected at different times are not the same. Therefore, in order to distinguish the characteristics of each gas data at different times and in different environments, each gas data can be distinguished, and from the gas data set, the gas data that significantly affects the overall change trend of the gas data and the gas data that significantly affects the local changes of the gas data can be determined.

[0030] Step 103 , for any of the multiple gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple gas data belonging to the same type of granularity identification to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification.

[0031] It should be noted that the target features extracted from the gas data under different types of granularity identification are not the same.

[0032] It is important to understand that, since gas data in actual production environments are affected by many factors, in order to better detect anomalies in gas data and adapt to changes in gas data even under complex working conditions, the granularity identifier to which the gas data belongs can well reflect which aspect of the long-term trend or local fluctuation is representative of the gas data in the subsequent anomaly detection process. Thus, it is possible to determine the granularity identifier suitable for the working environment according to different working conditions, and then focus on the gas data with the granularity identifier suitable for the working environment.

[0033] Step 104 , according to the target features corresponding to the plurality of gas data belonging to various granularity identifiers in the gas data set, a pre-trained gas anomaly detection model is used to determine an abnormal probability value sequence corresponding to the gas data set.

[0034] It should be noted that each abnormal probability value in the abnormal probability value sequence is used to indicate the probability that the corresponding gas data in the gas data set has gas data abnormality.

[0035] For example, the abnormal probability value is in the interval (0,1). Assume that the abnormal probability value sequence includes two abnormal probability values ​​A (0.9) and abnormal probability value B (0.1), where the abnormal probability value A is used to indicate the probability of gas data abnormality in the corresponding gas data A in the gas data set, and the abnormal probability value B is used to indicate the probability of gas data abnormality in the corresponding gas data B in the gas data set. It can be detected that there is a 90% probability of gas data abnormality in gas data A. In actual scenarios, it is more likely to be considered that gas data A is abnormal and needs special attention; there is a 10% probability of gas data abnormality in gas data B. In actual scenarios, it is more likely to be considered that gas data B is normal.

[0036] It should be understood that the determined abnormal probability value sequence is a quantitative representation of the possibility of abnormality in each gas data in the gas data set. However, in actual application scenarios, the abnormal probability value cannot intuitively reflect whether the gas data is abnormal. The abnormal probability value can be further judged to more clearly understand the data status of the gas data.

[0037] In order to intuitively reflect whether the gas data is abnormal, as a possible implementation method, it is judged whether the abnormal probability value corresponding to each gas data exceeds the abnormal threshold value to determine the data status of the gas data.

[0038] As an example, for each gas data in the gas data set, it is judged whether the abnormal probability value corresponding to the gas data exceeds the abnormal threshold value to determine the data state of the gas data.

[0039] In some examples, if the abnormal probability value corresponding to the gas data exceeds the abnormal threshold, the data state of the gas data is determined to be an abnormal state; if the abnormal probability value corresponding to the gas data is less than the abnormal threshold, the data state of the gas data is determined to be a normal state.

[0040] It should be noted that the specific value of the abnormal threshold is determined according to the actual situation, and this embodiment does not make a specific limitation. For example, 0.5 can be set as the abnormal threshold. If the abnormal probability value corresponding to the gas data exceeds 0.5, the data state of the gas data is abnormal; if the abnormal probability value corresponding to the gas data is less than 0.5, the data state of the gas data is determined to be normal.

[0041] The gas data anomaly detection method provided in the embodiment of the present application obtains a gas data set to be detected; wherein the gas data set includes multiple gas data collected within at least one time period of the mining area to be detected; based on the joint information entropy calculated from the gas data set, the gas data set is granularized to determine the granularity identification to which each gas data in the gas data set belongs; for any multiple gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple gas data belonging to the same type of granularity identification to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification; according to the target features corresponding to the multiple gas data belonging to each type of granularity identification in the gas data set, a pre-trained gas anomaly detection model is used to determine the abnormal probability value sequence corresponding to the gas data set; wherein each abnormal probability value in the abnormal probability value sequence is used to indicate the gas data. The probability of gas data anomalies occurring in the corresponding gas data in the gas data set is determined by the gas anomaly detection model. Therefore, before determining the probability of gas data anomalies occurring in the gas data set through the gas anomaly detection model, not only the various factors that may affect the abnormal situation of gas data in the actual working scenario are comprehensively considered, but also the nonlinear characteristics of the distribution of multiple gas data in the gas data set are considered. Therefore, the gas data set is comprehensively divided into granularities not only from the perspective of local fluctuations of gas data but also from the perspective of the macro trend of gas data. Then, based on the gas data belonging to various granularity labels, the probability of gas data anomalies is determined through the gas anomaly detection model. Without consuming a lot of time cost, the gas data can be pre-labeled to provide high-quality gas data for the gas anomaly detection model, which greatly improves the accuracy and reliability of the gas anomaly detection model.

[0042] In the embodiment of the present application, in order to make it clear how to calculate the joint information entropy based on the gas data set, the gas data set is divided into granularities to determine the granularity identifier to which each gas data in the gas data set belongs, the following is combined with Figure 2 A possible implementation manner of determining a granularity identifier to which each gas data in a gas data set belongs is described by way of example. Figure 2 The process of the gas data anomaly detection method provided in the embodiment of the present application Figure 2 .

[0043] like Figure 2 As shown, the gas data anomaly detection method may include the following steps: Step 201 , obtaining a gas data set to be detected; wherein the gas data set includes a plurality of gas data collected within at least one time period in the mining area to be detected.

[0044] It should be noted that the execution process of step 201 can be implemented in any of the embodiments of the present application. The embodiments of the present application do not limit this and will not be described in detail.

[0045] Step 202, obtaining a preset number of time steps and feature dimensions.

[0046] It should be noted that the granularity identifier includes but is not limited to a coarse-grained identifier, a medium-grained identifier, and a fine-grained identifier.

[0047] It is important to understand that gas data with coarse-grained labels focuses on detecting whether there are abnormalities in the long-term trend of gas data, so as to grasp the macro trend of gas data in the gas data set; gas data with fine-grained labels focuses on detecting whether there are abnormalities in short-term fluctuations and local details of gas data, so as to capture abnormal changes in local gas data in the gas data set; gas data with medium-grained labels focuses on abnormal aspects of gas data, which are between those with coarse-grained labels and those with fine-grained labels.

[0048] It should be noted that the specific value of the time step number is determined according to actual conditions and is not specifically limited in this embodiment.

[0049] It should be noted that the specific information of the feature dimension is determined based on existing technologies and the experience of relevant personnel, and is not specifically limited in this embodiment.

[0050] It should be understood that gas data is affected by many complex factors such as geological structure, ventilation conditions, mining technology, etc. Gas data with different granularity labels have different focuses on detecting gas data anomalies.

[0051] For example, for gas data with coarse-grained identification, the feature of interest may be gas concentration; for gas data with fine-grained identification, the features of interest may be gas concentration, temperature, and humidity.

[0052] Step 203, for any gas data, determine the joint information entropy corresponding to the gas data according to the time step number and the feature dimension.

[0053] For example, the joint information entropy corresponding to the gas data can be expressed as: ; Among them, the gas data set , N represents the number of gas data in the gas data set; T represents the number of time steps, and F represents the feature dimension; for each gas data represents the joint information entropy; is the probability of the occurrence of the fth eigenvalue at the tth time step in the gas data i, obtained by counting all the time steps and eigenvalues ​​in the gas data i, that is, , is the number of times the f-th eigenvalue appears in the t-th time step in the gas data i.

[0054] Step 204, judging whether the joint information entropy corresponding to each gas data is greater than the first information entropy threshold or not greater than the second information entropy threshold, so as to determine the granularity identifier to which each gas data belongs; wherein the first information entropy threshold is greater than the second information entropy threshold.

[0055] It should be noted that the specific value of the first information entropy threshold and the specific value of the second information entropy threshold can be determined according to actual conditions, and this embodiment does not make specific limitations.

[0056] For example, by dividing the historical gas data, it can be found that when the first information entropy threshold is 0.6 and the second information entropy threshold is 0.3, each gas data can be clearly distinguished. Then, the first information entropy threshold can be set to 0.6 and the second information entropy threshold can be set to 0.3.

[0057] As an example, for the joint information entropy corresponding to each gas data, if the joint information entropy is greater than the first information entropy threshold, the granularity identifier to which the gas data belongs is determined to be a coarse-grained identifier; if the joint information entropy is not greater than the second information entropy threshold, the granularity identifier to which the gas data belongs is determined to be a fine-grained identifier; if the joint information entropy is not greater than the first information entropy threshold and greater than the second information entropy threshold, the granularity identifier to which the gas data belongs is determined to be a medium-grained identifier.

[0058] For example, setting the first information entropy threshold and the second information entropy threshold ( > ), granularity division is performed based on the joint information entropy of gas data: , the gas data is divided into gas data belonging to fine-grained identification; if , the gas data is classified as gas data belonging to the medium granularity mark; if , the gas data is divided into gas data belonging to the coarse-grained identification.

[0059] Step 205 , for any of the multiple gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple gas data belonging to the same type of granularity identification to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification.

[0060] Step 206 , according to the target features corresponding to the multiple gas data belonging to each type of granularity identification in the gas data set, a pre-trained gas anomaly detection model is used to determine the abnormal probability value sequence corresponding to the gas data set.

[0061] It should be noted that the execution process of step 205 and step 206 can be implemented in any way in the embodiments of the present application, and the embodiments of the present application do not limit this and will not be repeated.

[0062] The gas data anomaly detection method provided in the embodiment of the present application obtains a preset number of time steps and a feature dimension; for any gas data, the joint information entropy corresponding to the gas data is determined according to the number of time steps and the feature dimension; it is judged whether the joint information entropy corresponding to each gas data is greater than a first information entropy threshold or is not greater than a second information entropy threshold, so as to determine the granularity identifier to which each gas data belongs; wherein the first information entropy threshold is greater than the second information entropy threshold, and thus, the gas data set is divided according to the joint information entropy corresponding to the gas data, taking into account that the gas data is affected by many complex factors, multi-granularity analysis of the gas data is performed from the perspective of two major categories, macro and local, to lay the foundation for subsequent accurate detection of gas data anomalies, avoid information omissions and misjudgments caused by single granularity analysis, and deepen the understanding of the gas data.

[0063] In the embodiment of the present application, in order to clearly understand that for any multiple gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple gas data belonging to the same type of granularity identification to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification, the following is combined with Figure 3 A possible implementation method for determining at least one target feature corresponding to a plurality of gas data belonging to the same type of granularity identification is described exemplarily. Figure 3 The process of the gas data anomaly detection method provided in the embodiment of the present application Figure 3 .

[0064] like Figure 3 As shown, the gas data anomaly detection method may include the following steps: Step 301 , obtaining a gas data set to be detected; wherein the gas data set includes a plurality of gas data collected within at least one time period in the mining area to be detected.

[0065] Step 302 , based on the joint information entropy calculated for the gas data set, the gas data set is divided into granularities to determine the granularity identifier to which each gas data in the gas data set belongs.

[0066] It should be noted that the execution process of step 301 and step 302 can be implemented in any way in the embodiments of the present application, and the embodiments of the present application do not limit this and will not be repeated.

[0067] Step 303, according to the target particle size identification to which multiple gas data belonging to the same type of particle size identification belong, a target feature extraction method corresponding to the target particle size identification is determined through a preset particle size identification and feature extraction method mapping relationship table.

[0068] It should be noted that the granularity identifier and feature extraction method mapping relationship table includes a plurality of granularity identifiers and corresponding feature extraction methods.

[0069] For example, the mapping relationship table between granularity identifiers and feature extraction methods is shown in Table 1. In Table 1, there are coarse-grained identifiers, medium-grained identifiers and fine-grained identifiers. There is a mapping relationship between the coarse-grained identifier and the principal component analysis method, there is a mapping relationship between the medium-grained identifier and the wavelet transform method, and there is a mapping relationship between the fine-grained identifier and the wavelet transform method. Assuming that the target granularity identifier is a fine-grained identifier, then according to the mapping relationship table between granularity identifiers and feature extraction methods, it is determined that the target feature extraction method corresponding to the target granularity identifier (fine-grained identifier) ​​is the wavelet transform method. Based on the target feature extraction method (wavelet transform method), feature extraction is performed on multiple gas data belonging to the same type of granularity identifier to determine the target features corresponding to the multiple gas data belonging to the same type of granularity identifier.

[0070] Table 1 Mapping relationship between particle size identification and feature extraction method

[0071] Step 304 , based on the target feature extraction method, feature extraction is performed on multiple gas data belonging to the same type of granularity identification to determine target features corresponding to the multiple gas data belonging to the same type of granularity identification.

[0072] It should be understood that the directions of detecting anomalies in gas data belonging to different types of granularity labels are different. Therefore, granularity labels are not easy for computers to understand directly. In order for the gas anomaly detection model to quickly analyze the granularity label to which the gas data belongs, the granularity label can be uniquely encoded.

[0073] As an example, various types of granularity identifiers are obtained; for each type of granularity identifier, the granularity identifier is one-hot encoded to determine the one-hot encoding vector corresponding to the granularity identifier; based on the one-hot encoding vector corresponding to each type of granularity identifier, the one-hot encoding vector associated with the gas data is determined; based on the one-hot encoding vector associated with each gas data, the one-hot encoding vector matrix corresponding to the gas data set is determined.

[0074] For example, the granularity identifier is uniquely encoded, and for each gas data in the gas data set , whose granularity is ( ), then the gas data The corresponding one-hot encoded vector satisfy , then all the gas data unique hot encoding vectors in the gas data set form a matrix .

[0075] Among them, it should be noted that Indicates the maximum granularity identifier; N indicates the number of gas data in the gas data set.

[0076] Step 305 , according to the target features corresponding to the multiple gas data belonging to each type of granularity identification in the gas data set, a pre-trained gas anomaly detection model is used to determine the abnormal probability value sequence corresponding to the gas data set.

[0077] In order to determine the abnormal probability value sequence corresponding to the gas data set, as a possible implementation method, according to the one-hot encoding vector matrix and the target features corresponding to multiple gas data belonging to various granularity identifiers, the abnormal probability value sequence corresponding to the gas data set is determined through a gas anomaly detection model.

[0078] As an example, according to the one-hot encoding vector matrix corresponding to the gas data set, the granularity identifiers to which multiple gas data in the gas data set belong are determined; according to the target features corresponding to the multiple gas data belonging to each granularity identifier in the gas data set, the target features associated with each granularity identifier are determined; the target features associated with each granularity identifier are input into the gas anomaly detection model to determine the abnormal probability value sequence corresponding to the gas data set.

[0079] The gas data anomaly detection method provided in the embodiment of the present application determines the target feature extraction method corresponding to the target granularity identification according to the target granularity identification to which multiple gas data belonging to the same type of granularity identification belong through a preset granularity identification and feature extraction method mapping relationship table; based on the target feature extraction method, feature extraction is performed on multiple gas data belonging to the same type of granularity identification to determine the target features corresponding to the multiple gas data belonging to the same type of granularity identification, thereby determining the target feature extraction method corresponding to the target granularity identification through the granularity identification and feature extraction method mapping relationship table, thereby focusing on the granularity identification, quickly determining the applicable feature extraction method for the gas data, and through the applicable feature extraction method, completely extracting the target features corresponding to the gas data.

[0080] Figure 4 A flow chart of a gas data anomaly detection method provided in an embodiment of the present application Figure 4 .

[0081] like Figure 4 As shown, the gas data anomaly detection method may include the following steps: Step 401 : Acquire a historical gas data set and an actual abnormal probability value sequence corresponding to the historical gas data set.

[0082] It should be noted that the historical gas data set includes multiple historical gas data collected within at least one historical time period of the detected mining area; each actual abnormality probability value in the actual abnormality probability value sequence is used to indicate the probability of gas data abnormality occurring in the corresponding historical gas data in the historical gas data set.

[0083] Step 402 , based on the joint information entropy calculated for the historical gas data set, the historical gas data set is divided into granularities to determine the granularity identifier to which each historical gas data in the historical gas data set belongs.

[0084] Step 403 , for any of the multiple historical gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple historical gas data belonging to the same type of granularity identification to determine the historical target features corresponding to the multiple historical gas data belonging to the same type of granularity identification.

[0085] It should be noted that for the specific description of steps 401 to 403, reference may be made to other descriptions of this embodiment, which will not be repeated here.

[0086] Step 404 , determining a trained gas anomaly detection model according to historical target features corresponding to a plurality of historical gas data belonging to various granularity identifiers in the historical gas data set and actual anomaly probability value sequences corresponding to the historical gas data set.

[0087] In some embodiments, historical target features corresponding to multiple historical gas data belonging to various granularity identifiers in the historical gas data set are input into the initial gas anomaly detection model to determine the abnormal probability value sequence corresponding to the historical gas data set; according to the abnormal probability value sequence corresponding to the historical gas data set and the actual abnormal probability value sequence, the initial gas anomaly detection model is trained to determine the trained gas anomaly detection model.

[0088] In this embodiment, a historical gas data set and an actual abnormal probability value sequence corresponding to the historical gas data set are obtained; based on the joint information entropy calculated from the historical gas data set, the historical gas data set is granularized to determine the granularity identifier to which each historical gas data in the historical gas data set belongs; for any multiple historical gas data belonging to the same type of granularity identifier, feature extraction is performed on the multiple historical gas data belonging to the same type of granularity identifier to determine the historical target features corresponding to the multiple historical gas data belonging to the same type of granularity identifier; according to the historical target features corresponding to the multiple historical gas data belonging to each type of granularity identifier in the historical gas data set and the actual abnormal probability value sequence corresponding to the historical gas data set, a trained gas anomaly detection model is determined, thereby, through the difference between the abnormal probability value sequence corresponding to the historical gas data set and the actual abnormal probability value sequence, the gas anomaly detection model is continuously trained, thereby greatly improving the accuracy of the gas anomaly detection model in detecting gas data anomalies.

[0089] For example, the framework diagram corresponding to the gas data anomaly detection process is as follows: Figure 5 As shown, X (gas data set) is input, and X is calculated to determine the joint information entropy corresponding to each gas data in the gas data set. Based on the joint information entropy corresponding to each gas data, the gas data set is dynamically divided into multiple granularities to determine the granularity identification to which each gas data belongs. Feature extraction is performed on each gas data to determine the target feature Y1 corresponding to the gas data belonging to the coarse-grained identification, the target feature Y2 corresponding to the gas data belonging to the medium-grained identification, and the target feature Y3 corresponding to the gas data belonging to the fine-grained identification. Y1, Y2 and Y3 are input into the gas anomaly detection model to determine the abnormal probability value sequence corresponding to the gas data set.

[0090] It should be understood that, since the focus for gas data with medium-granularity identification and gas data with fine-granularity identification is on the possibility of subtle abnormalities in the gas data, in order to efficiently detect anomalies in subtle aspects of gas data based on the subsequent gas anomaly detection model, the correlation between multiple gas data with the same type of granularity identification can be strengthened.

[0091] In some embodiments, a first clustering algorithm corresponding to gas data belonging to medium-granularity identification is obtained; a second clustering algorithm corresponding to gas data belonging to fine-granularity identification is obtained; based on the first clustering algorithm, multiple gas data belonging to medium-granularity identification are clustered to determine cluster-optimized gas data belonging to medium-granularity identification; based on the second clustering algorithm, multiple gas data belonging to fine-granularity identification are clustered to determine cluster-optimized gas data belonging to fine-granularity identification.

[0092] It should be noted that the gas anomaly detection model is composed of a convolutional neural network (CNN), a maximum pooling layer and a long short-term memory network (LSTM).

[0093] For example, the architecture diagram of the gas anomaly detection model is as follows: Figure 6 As shown, the process of performing anomaly detection on gas data may include: inputting the target features associated with the coarse-grained identifier, the target features associated with the medium-grained identifier, and the target features associated with the fine-grained identifier into the gas anomaly detection model; the convolution layer included in the gas anomaly detection model: performing a convolution operation by using a convolution function (Conv3-32), wherein the parameters in Conv3-32 are set to filters=32, kernel_size=3, and the activation function is the ReLU function, so as to determine the convolved data, and input the convolved data into the maximum pooling layer; the maximum pooling layer (Max-Pool) included in the gas anomaly detection model: wherein the parameter in the maximum pooling function is set to pool_size=2, performing a maximum pooling operation on the convolved data to determine the pooled data, and inputting the pooled data into the LSTM layer; the long short-term memory network (LSTM) layer included in the gas anomaly detection model: the number of units is 32, and the data output by the LSTM layer is determined; the fusion and output layer included in the gas anomaly detection model: connecting the data output by the LSTM layer under various granularity identifiers to output anomaly probability values.

[0094] In order to implement the above embodiments, the present application also proposes a gas data anomaly detection device.

[0095] Figure 7 A schematic diagram showing the structure of a gas data anomaly detection device provided in an embodiment of the present application.

[0096] like Figure 7 As shown, the gas data anomaly detection device 700 includes: a first acquisition module 701, a granularity division module 702, a first determination module 703 and a second determination module 704.

[0097] The first acquisition module 701 is used to acquire a gas data set to be detected; wherein the gas data set includes a plurality of gas data collected in at least one time period of the mining area to be detected; A granularity division module 702 is used to divide the gas data set into granularities based on the joint information entropy calculated for the gas data set, so as to determine the granularity identifier to which each gas data in the gas data set belongs; The first determination module 703 is used to extract features from the multiple gas data belonging to the same type of granularity identification, so as to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification; The second determination module 704 is used to determine the abnormal probability value sequence corresponding to the gas data set through a pre-trained gas anomaly detection model based on the target features corresponding to multiple gas data belonging to various granularity identifiers in the gas data set; wherein each abnormal probability value in the abnormal probability value sequence is used to indicate the probability of gas data anomaly occurring in the corresponding gas data in the gas data set.

[0098] In one embodiment of the present application, the granularity identifier includes but is not limited to a coarse granularity identifier, a medium granularity identifier, and a fine granularity identifier; the granularity division module 702 is specifically used to: Get the preset number of time steps and feature dimensions; For any gas data, determine the joint information entropy corresponding to the gas data according to the time step number and feature dimension; It is judged whether the joint information entropy corresponding to each gas data is greater than the first information entropy threshold or not greater than the second information entropy threshold to determine the granularity identifier to which each gas data belongs; wherein the first information entropy threshold is greater than the second information entropy threshold.

[0099] In one embodiment of the present application, the granularity division module 702 is specifically used to: For each gas data's corresponding joint information entropy, if the joint information entropy is greater than a first information entropy threshold, determining that the granularity identifier to which the gas data belongs is a coarse granularity identifier; If the joint information entropy is not greater than the second information entropy threshold, determining that the granularity identifier to which the gas data belongs is a fine-granularity identifier; If the combined information entropy is not greater than the first information entropy threshold and greater than the second information entropy threshold, it is determined that the granularity identifier to which the gas data belongs is the medium granularity identifier.

[0100] In one embodiment of the present application, the first determining module 703 is specifically configured to: According to the target particle size identification to which multiple gas data belonging to the same type of particle size identification belong, a target feature extraction method corresponding to the target particle size identification is determined through a preset particle size identification and feature extraction method mapping relationship table; wherein the particle size identification and feature extraction method mapping relationship table includes multiple particle size identifications and corresponding feature extraction methods; Based on the target feature extraction method, feature extraction is performed on multiple gas data belonging to the same type of granularity identification to determine the target features corresponding to the multiple gas data belonging to the same type of granularity identification.

[0101] In one embodiment of the present application, the device further includes a one-hot encoding module, specifically configured to: Get various granularity identifiers; For each type of granularity identifier, perform one-hot encoding on the granularity identifier to determine a one-hot encoding vector corresponding to the granularity identifier; Determine the one-hot encoding vector associated with the gas data based on the one-hot encoding vectors corresponding to the various granularity identifiers; According to the one-hot encoding vectors associated with each gas data, the one-hot encoding vector matrix corresponding to the gas data set is determined.

[0102] In one embodiment of the present application, the second determining module 704 is specifically configured to: Determine the granularity identifiers to which the plurality of gas data in the gas data set respectively belong according to the one-hot encoding vector matrix corresponding to the gas data set; According to target features corresponding to a plurality of gas data belonging to various granularity identifiers in the gas data set, target features associated with various granularity identifiers are determined; The target features associated with various granularity identifiers are input into the gas anomaly detection model to determine the anomaly probability value sequence corresponding to the gas data set.

[0103] In one embodiment of the present application, the device further includes a gas anomaly detection model module, which is specifically used to: Acquire a historical gas data set and a sequence of actual abnormal probability values ​​corresponding to the historical gas data set; wherein the historical gas data set includes a plurality of historical gas data collected in at least one historical time period of the detected mining area, and each actual abnormal probability value in the sequence of actual abnormal probability values ​​is used to indicate the probability of gas data abnormality occurring in the corresponding historical gas data in the historical gas data set; Based on the joint information entropy calculated from the historical gas data set, the historical gas data set is divided into granularities to determine the granularity identifier to which each historical gas data in the historical gas data set belongs; For any multiple historical gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple historical gas data belonging to the same type of granularity identification to determine the historical target features corresponding to the multiple historical gas data belonging to the same type of granularity identification; The trained gas anomaly detection model is determined according to historical target features corresponding to multiple historical gas data belonging to various granularity identifiers in the historical gas data set and actual anomaly probability value sequences corresponding to the historical gas data set.

[0104] In one embodiment of the present application, the gas anomaly detection model module is specifically used to: Inputting historical target features corresponding to multiple historical gas data belonging to various granularity identifiers in the historical gas data set into the initial gas anomaly detection model to determine an abnormal probability value sequence corresponding to the historical gas data set; According to the abnormal probability value sequence corresponding to the historical gas data set and the actual abnormal probability value sequence, the initial gas anomaly detection model is trained to determine the trained gas anomaly detection model.

[0105] The gas data anomaly detection device provided in the embodiment of the present application obtains a gas data set to be detected; wherein the gas data set includes multiple gas data collected within at least one time period of the mining area to be detected; based on the joint information entropy calculated from the gas data set, the gas data set is granularized to determine the granularity identification to which each gas data in the gas data set belongs; for any multiple gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple gas data belonging to the same type of granularity identification to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification; according to the target features corresponding to the multiple gas data belonging to each type of granularity identification in the gas data set, a pre-trained gas anomaly detection model is used to determine the abnormal probability value sequence corresponding to the gas data set; wherein each abnormal probability value in the abnormal probability value sequence is used to indicate the gas data. The probability of gas data anomalies occurring in the corresponding gas data in the gas data set is determined by the gas anomaly detection model. Therefore, before determining the probability of gas data anomalies occurring in the gas data set through the gas anomaly detection model, not only the various factors that may affect the abnormal situation of gas data in the actual working scenario are comprehensively considered, but also the nonlinear characteristics of the distribution of multiple gas data in the gas data set are considered. Therefore, the gas data set is comprehensively divided into granularities not only from the perspective of local fluctuations of gas data but also from the perspective of the macro trend of gas data. Then, based on the gas data belonging to various granularity labels, the probability of gas data anomalies is determined through the gas anomaly detection model. Without consuming a lot of time cost, the gas data can be pre-labeled to provide high-quality gas data for the gas anomaly detection model, which greatly improves the accuracy and reliability of the gas anomaly detection model.

[0106] It should be noted that the above explanation of the embodiment of the gas data anomaly detection method is also applicable to the gas data anomaly detection device of this embodiment, and will not be repeated here.

[0107] In order to implement the above embodiments, the present application also proposes an electronic device.

[0108] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0109] like Figure 8As shown, the electronic device 800 may include: a processor 802, and a memory 803 communicatively connected to the processor 802; the memory 803 stores computer-executable instructions; the processor 802 executes the computer-executable instructions stored in the memory 803 to implement the method provided in the aforementioned embodiment. Furthermore, the electronic device 800 further includes: The transceiver 801 is used for communication between the memory 803 and the processor 802 .

[0110] The memory 803 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0111] If the memory 803, the processor 802 and the transceiver 801 are implemented independently, the transceiver 801, the memory 803 and the processor 802 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0112] Optionally, in a specific implementation, if the memory 803, the processor 802 and the transceiver 801 are integrated on a chip, the memory 803, the processor 802 and the transceiver 801 can communicate with each other through an internal interface.

[0113] The processor 802 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0114] In order to implement the above embodiments, the present application also proposes a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the methods provided by the above embodiments.

[0115] In order to implement the above embodiments, the present application also proposes a computer program product, including a computer program, which implements the methods provided by the above embodiments when executed by a processor.

[0116] The collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in this application are in compliance with relevant laws and regulations and do not violate public order and good morals.

[0117] It should be noted that personal information from users should be collected for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. In addition, such collection / sharing should be carried out after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign the agreement / authorization including authorization of relevant user information before the user uses the function. In addition, any necessary steps should be taken to protect and safeguard access to such personal information data and ensure that others who have access to personal information data comply with its privacy policy and procedures.

[0118] The present application is expected to provide an implementation scheme for users to selectively block the use or access of personal information data. That is, the present disclosure is expected to provide hardware and / or software to prevent or block access to such personal information data. Once the personal information data is no longer needed, the risk can be minimized by limiting data collection and deleting the data. In addition, when applicable, such personal information is de-identified to protect the privacy of the user.

[0119] In the description of the aforementioned embodiments, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they contradict each other.

[0120] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0121] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0122] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0123] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0124] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0125] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0126] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for detecting gas data anomaly, characterized in that: The following steps are involved: Acquire a gas data set to be detected; wherein the gas data set includes a plurality of gas data collected in at least one time period of the mining area to be detected; Based on the joint information entropy calculated for the gas data set, the gas data set is divided into granularities to determine a granularity identifier to which each gas data in the gas data set belongs; For any of the multiple gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple gas data belonging to the same type of granularity identification to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification; According to the target features corresponding to multiple gas data belonging to each type of granularity identification in the gas data set, a pre-trained gas anomaly detection model is used to determine the abnormal probability value sequence corresponding to the gas data set; wherein each abnormal probability value in the abnormal probability value sequence is used to indicate the probability of gas data abnormality occurring in the corresponding gas data in the gas data set.

2. The method according to claim 1, characterized in that in, The granularity identifier includes but is not limited to a coarse granularity identifier, a medium granularity identifier and a fine granularity identifier; The step of dividing the gas data set into granularities based on the joint information entropy calculated based on the gas data set to determine a granularity identifier to which each gas data in the gas data set belongs includes: Get the preset number of time steps and feature dimensions; For any of the gas data, determining the joint information entropy corresponding to the gas data according to the time step number and the feature dimension; It is judged whether the joint information entropy corresponding to each of the gas data is greater than a first information entropy threshold or is not greater than a second information entropy threshold to determine the granularity identifier to which each of the gas data belongs; wherein the first information entropy threshold is greater than the second information entropy threshold.

3. The method according to claim 2, characterized in that The step of judging whether the joint information entropy corresponding to each of the gas data exceeds a first information entropy threshold or is not greater than a second information entropy threshold, so as to determine the granularity identifier to which each of the gas data belongs; wherein the first information entropy threshold is greater than the second information entropy threshold, includes: For the joint information entropy corresponding to each of the gas data, if the joint information entropy is greater than the first information entropy threshold, determining that the granularity identifier to which the gas data belongs is the coarse granularity identifier; If the joint information entropy is not greater than the second information entropy threshold, determining that the granularity identifier to which the gas data belongs is the fine-granularity identifier; If the joint information entropy is not greater than the first information entropy threshold and greater than the second information entropy threshold, it is determined that the granularity identifier to which the gas data belongs is the medium granularity identifier.

4. The method according to claim 1, characterized in that The step of extracting features from the plurality of gas data belonging to the same type of granularity identifier to determine at least one target feature corresponding to the plurality of gas data belonging to the same type of granularity identifier includes: According to the target particle size identification to which the plurality of gas data belonging to the same type of particle size identification belong, a target feature extraction method corresponding to the target particle size identification is determined through a preset particle size identification and feature extraction method mapping relationship table; wherein the particle size identification and feature extraction method mapping relationship table includes a plurality of particle size identifications and corresponding feature extraction methods; Based on the target feature extraction method, feature extraction is performed on the multiple gas data belonging to the same type of granularity identification to determine the target feature corresponding to the multiple gas data belonging to the same type of granularity identification.

5. The method according to claim 4, characterized in that in, After extracting features from the plurality of gas data belonging to the same type of granularity identifier to determine at least one target feature corresponding to the plurality of gas data belonging to the same type of granularity identifier, the method further comprises: Obtaining various types of granularity identifiers; For each type of granularity identifier, one-hot encoding is performed on the granularity identifier to determine a one-hot encoding vector corresponding to the granularity identifier; Determining the one-hot encoding vector associated with the gas data based on the one-hot encoding vector corresponding to each type of the granularity identifier; According to the one-hot encoding vectors associated with each of the gas data, a one-hot encoding vector matrix corresponding to the gas data set is determined.

6. The method according to claim 5, characterized in that The method of determining the abnormal probability value sequence corresponding to the gas data set by a pre-trained gas anomaly detection model according to the target features corresponding to the plurality of gas data under each type of granularity identification in the gas data set includes: Determining the granularity identifiers to which the plurality of gas data in the gas data set respectively belong according to the one-hot encoding vector matrix corresponding to the gas data set; Determine the target feature associated with each type of granularity identifier according to the target features corresponding to the plurality of gas data belonging to each type of granularity identifier in the gas data set; The target features associated with each type of granularity identifier are input into a gas anomaly detection model to determine the anomaly probability value sequence corresponding to the gas data set.

7. The method according to claim 1, characterized in that The method for obtaining the gas anomaly detection model includes: Acquire a historical gas data set and a sequence of actual abnormal probability values ​​corresponding to the historical gas data set; wherein the historical gas data set includes a plurality of historical gas data collected in at least one historical time period of the detected mining area, and each actual abnormal probability value in the sequence of actual abnormal probability values ​​is used to indicate the probability of gas data abnormality occurring in the corresponding historical gas data in the historical gas data set; Based on the joint information entropy calculated from the historical gas data set, the historical gas data set is divided into granularities to determine the granularity identifier to which each piece of the historical gas data in the historical gas data set belongs; For any of the multiple historical gas data belonging to the same type of granularity identification, feature extraction is performed on the multiple historical gas data belonging to the same type of granularity identification to determine the historical target features corresponding to the multiple historical gas data belonging to the same type of granularity identification; The trained gas anomaly detection model is determined according to the historical target features corresponding to the multiple historical gas data belonging to each type of granularity identification in the historical gas data set and the actual abnormal probability value sequence corresponding to the historical gas data set.

8. The method according to claim 7, characterized in that The determining of the gas anomaly detection model according to the historical target features corresponding to the plurality of historical gas data belonging to each type of the granularity identifier in the historical gas data set and the actual anomaly probability value sequence corresponding to the historical gas data set includes: Inputting the historical target features corresponding to the plurality of historical gas data belonging to each type of the granularity identification in the historical gas data set into the initial gas anomaly detection model to determine an abnormal probability value sequence corresponding to the historical gas data set; The initial gas anomaly detection model is trained according to the anomaly probability value sequence corresponding to the historical gas data set and the actual anomaly probability value sequence to determine the trained gas anomaly detection model.

9. A gas data anomaly detection device, characterized in that: The following steps are involved: A first acquisition module is used to acquire a gas data set to be detected; wherein the gas data set includes a plurality of gas data collected in at least one time period of the mining area to be detected; A granularity division module, configured to divide the gas data set into granularities based on the joint information entropy calculated for the gas data set, so as to determine a granularity identifier to which each gas data in the gas data set belongs; A first determination module is used to extract features from the multiple gas data belonging to the same type of granularity identification, so as to determine at least one target feature corresponding to the multiple gas data belonging to the same type of granularity identification; The second determination module is used to determine the abnormal probability value sequence corresponding to the gas data set through a pre-trained gas anomaly detection model according to the target features corresponding to multiple gas data belonging to each type of granularity identification in the gas data set; wherein each abnormal probability value in the abnormal probability value sequence is used to indicate the probability of the corresponding gas data in the gas data set having a gas data anomaly.

10. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 8.

Images