Multi-channel space-time slice fusion AMC concentration information extraction method and system

Abstract

The invention is suitable for the technical field of air data monitoring, and particularly relates to a multichannel space-time slice fusion AMC concentration information extraction method and system, and the method comprises the steps: receiving the gas concentration collected by an AMC monitor, and constructing a space-time data cube; selecting a time point in the spatio-temporal data cube, reading a time remainder sub-cube in the spatio-temporal data cube based on the selected time point, and determining spatial features based on the time remainder sub-cube; selecting a spatial position in the spatio-temporal data cube based on the spatial features, reading a position remainder sub-cube based on the selected spatial position, and determining concentration change information of each type of gas by the position remainder sub-cube; marking the gas according to the concentration change information of the gas, querying a preset characteristic convolution kernel of the gas, traversing the spatio-temporal data cube, matching abnormal time and abnormal position, and feeding back to a management end; the system can help a worker to quickly position key data, and is high in data integration level and high in response speed.

Description

Technical Field

[0001] The present invention relates to the technical field of air data monitoring, and in particular, to a method and system for extracting AMC concentration information by multi-channel spatio-temporal slice fusion. Background Art

[0002] AMC (Airborne Molecular Contaminants), that is, airborne molecular contaminants, refers to air pollutants existing in the air in the form of gas or molecules, which have adverse effects on the environment, human health, instrument equipment or industrial processes; in scenarios with extremely high precision requirements (such as clean rooms, chip processing workshops, etc.), AMC concentration monitoring is essential.

[0003] With the development of AMC monitoring equipment, there are many kinds of AMC monitoring equipment in the prior art, which has the monitoring function of multiple channels and multiple positions. Correspondingly, when the number of AMC monitoring equipment is very large, the data obtained is very complicated, and the information extraction process is very difficult. Therefore, how to process the AMC concentration information of multiple channels, multiple positions and multiple moments to improve the information acquisition efficiency and reduce the data processing pressure of the staff is the technical problem that the technical solution of the present invention wants to solve. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for extracting AMC concentration information by multi-channel spatio-temporal slice fusion to solve the problems raised in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions: A method for extracting AMC concentration information by multi-channel spatio-temporal slice fusion, the method comprising: Receiving the gas concentration containing time, position and type collected by an AMC monitor, and constructing a spatio-temporal data cube; the total number of elements of the spatio-temporal data cube is: ; where is the total number of elements, is the total number of gas types to be monitored, is the number of time steps, is the spatial grid size; Selecting time points in the spatio-temporal data cube, reading a time co-subcube in the spatio-temporal data cube based on the selected time points, and determining spatial features based on the time co-subcube; Selecting spatial positions in the spatio-temporal data cube based on the spatial features, reading a position co-subcube based on the selected spatial positions, and determining the concentration change information of each type of gas based on the position co-subcube; Mark the gas according to the gas concentration change information, query the preset characteristic convolution kernel of the gas, traverse the spatio-temporal data cube, match the abnormal time and abnormal position, and feedback to the management terminal.

[0006] Further, the steps of constructing the spatio-temporal data cube by receiving the gas concentration containing time, position and type collected by the AMC monitor include: Obtain the position of the monitor at each moment according to the position collector built in the AMC monitor; Regularly receive the gas concentration containing time and type obtained by the AMC monitor; Statistically calculate the gas concentration of each type at each position according to time, and construct a spatio-temporal data cube; Among them, when obtaining the position, calculate the position change rate in real time, and adjust the position acquisition frequency according to the change rate; when obtaining the gas concentration, calculate the concentration change rate in real time, and adjust the acquisition frequency of the gas concentration of each type based on the concentration change rate and the reference change rate determined by the type.

[0007] Further, the steps of statistically calculating the gas concentration of each type at each position according to time and constructing a spatio-temporal data cube include: Receive the time step set by the management personnel, and determine the time dimension within the preset time range; Receive the position calibration grid set by the management personnel, and determine the coordinate dimension of each coordinate; the grid cell size in the position calibration grid is a preset value; the coordinates do not exceed three parameters; Receive the total number of types set by the management personnel as the type dimension; Statistically calculate the time dimension, coordinate dimension and type dimension, and construct a cube structure; Statistically calculate the gas concentration containing time, position and type based on the cube structure to obtain an intermediate data cube; Perform interpolation in the intermediate data cube based on a preset interpolation algorithm to fill in the empty data positions and obtain a spatio-temporal data cube.

[0008] Further, the steps of selecting a time point in the spatio-temporal data cube, reading a time co-subcube in the spatio-temporal data cube based on the selected time point, and determining the spatial feature based on the time co-subcube include: Extract the time parameter, and query the subcube corresponding to each time point in the time parameter; Input the subcube into a preset numerical normalization model to convert it into a feature value; Arrange the feature values according to time sequence, fit the feature values, and obtain a time change function; Calculate the derivative function of the preset order of the time change function, and query the function value and derivative value of the time point in the time change function and all its derivative functions; Cluster the time points based on the function values and derivative values, and select the category of time points with the largest number of time points as the selected time points; Read the sub-cubes corresponding to the selected time points as the time co-sub-cubes; Perform gradient recognition on each time co-sub-cube to determine the spatial features.

[0009] Further, the step of performing gradient recognition on each time co-sub-cube to determine the spatial features includes: Successively extract various types of data in each time co-sub-cube to obtain a data distribution matrix; Perform gradient recognition on the data distribution matrix to determine the spatial profile; Statistically analyze all types of spatial profiles of all time points, merge the spatial profiles, and obtain a comprehensive profile; the retention degree is determined according to the number of spatial profiles during the merging process; Select the comprehensive profiles with the retention degree reaching the preset retention threshold as the spatial features.

[0010] Further, the step of selecting spatial positions in the spatio-temporal data cube based on the spatial features, reading the position co-sub-cubes based on the selected spatial positions, and determining the concentration change information of various types of gases by the position co-sub-cubes includes: Read the comprehensive profiles in the spatial features and divide the area into sub-areas; Determine the quantity according to the area of each sub-area, and randomly select that quantity of spatial positions in the sub-area; the quantity is proportional to the area; Query the gas concentrations of various types at each time at each spatial position as the position co-sub-cube; the position co-sub-cube is a two-dimensional matrix; Extract the gas concentrations of each type at all times in the position co-sub-cube and input them into a preset fluctuation recognition model to obtain the fluctuation degree of that type; Statistically analyze the fluctuation degrees of all types as the concentration change information of this spatial position.

[0011] Further, the step of marking the gas according to the concentration change information of the gas, querying the preset characteristic convolution kernel of the gas, traversing the spatio-temporal data cube, matching the abnormal time and abnormal position, and feeding back to the management end includes: Query the concentration change information of the selected spatial positions and mark the types with the fluctuation degree reaching the preset fluctuation threshold; Record the total number of marks of each type, and when the total number of marks reaches the preset quantity, query the characteristic convolution kernel corresponding to that type; Traverse the spatio-temporal data cube based on the characteristic convolution kernel, match the abnormal time and abnormal position, and feed back to the management end; Among them, the characteristic convolution kernel is the gas concentration of this type of gas within a period of time under a preset dangerous situation.

[0012] The technical solution of the present invention also provides a multi-channel spatio-temporal slice fusion AMC concentration information extraction system, and the system includes: A cube construction module, configured to receive the gas concentration containing time, position, and type collected by the AMC monitor, and construct a spatio-temporal data cube; the total number of elements of the spatio-temporal data cube is: ; where is the total number of elements, is the total number of gas types to be monitored, is the number of time steps, is the spatial grid size; A spatial feature determination module, configured to select a time point in the spatio-temporal data cube, read a time co-subcube in the spatio-temporal data cube based on the selected time point, and determine spatial features based on the time co-subcube; A gas concentration analysis module, configured to select a spatial position in the spatio-temporal data cube based on the spatial features, read a position co-subcube based on the selected spatial position, and determine the concentration change information of each type of gas from the position co-subcube; An anomaly matching module, configured to mark the gas according to the concentration change information of the gas, query the preset characteristic convolution kernel of the gas, traverse the spatio-temporal data cube, match the abnormal time and abnormal position, and feedback to the management end.

[0013] Furthermore, the cube construction module includes: A position acquisition unit, configured to acquire the position of the monitor at each moment according to the position collector built in the AMC monitor; A gas concentration receiving unit, configured to regularly receive the gas concentration containing time and type acquired by the AMC monitor; A data statistics unit, configured to statistically calculate the gas concentration of each type at each position according to time, and construct a spatio-temporal data cube; Among them, when acquiring the position, the position change rate is calculated in real time, and the position acquisition frequency is adjusted according to the change rate; when acquiring the gas concentration, the concentration change rate is calculated in real time, and the acquisition frequency of the gas concentration of each type is adjusted based on the concentration change rate and the reference change rate determined by the type.

[0014] Furthermore, the spatial feature determination module includes: A data query unit, configured to extract time parameters and query the subcube corresponding to each time point in the time parameters; A data normalization unit, configured to input the subcube into a preset numerical normalization model and convert it into a feature value; A feature fitting unit, configured to arrange feature values according to time sequence, fit the feature values, and obtain a time variation function; A numerical calculation unit, configured to calculate a derivative function of a preset order of the time variation function, and query function values and derivative values of time points in the time variation function and all its derivative functions; A time point classification unit, configured to cluster time points based on the function values and derivative values, and select the category of time points with the largest number of time points as the selected time points; A data output unit, configured to read sub-cubes corresponding to the selected time points as time co-sub-cubes; A gradient recognition unit, configured to perform gradient recognition on each time co-sub-cube to determine spatial features.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention statistically analyzes the monitoring data of an AMC monitor, constructs a spatio-temporal data cube, extracts data cuts under different parameters from the spatio-temporal data cube, analyzes the data cuts, selects important positions, synchronously determines the gas change conditions at the important positions, and then determines the risk gas types. Furthermore, with the help of a convolutional recognition model, it matches the abnormal positions and abnormal times corresponding to the risk gas types, helping the staff quickly locate key data. The data integration degree is extremely high, the response speed is extremely fast, which helps to improve the data acquisition efficiency and reduce the difficulty for the staff to obtain data. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0017] Figure 1 It is the overall flow block diagram of the AMC concentration information extraction method for multi-channel spatio-temporal slice fusion provided by the embodiment of the present invention.

[0018] Figure 2 It is the structural diagram of the AMC concentration information extraction system for multi-channel spatio-temporal slice fusion provided by the embodiment of the present invention. Detailed Embodiments

[0019] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the following further details the present invention with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0020] Figure 1The overall flowchart of the AMC concentration information extraction method based on multi-channel spatio-temporal slice fusion is shown. In an embodiment of the present invention, a method for extracting AMC concentration information based on multi-channel spatio-temporal slice fusion includes the following steps: Step S100: Receive the gas concentrations containing time, location, and type collected by the AMC monitor, and construct a spatio-temporal data cube. The AMC monitor is a device for detecting gas pollutants. It can be used to collect the gas concentrations of various gas pollutants. Its installation location is not fixed and can be installed on some mobile devices or some fixed facilities. When the AMC monitor obtains the gas concentration, it records the time, location, and type, indicating the gas concentration of what type of gas at what time and where. Count all the obtained gas concentrations within a preset time range (such as one day or one week) and insert them into a multi-dimensional array, called the spatio-temporal data cube. The total number of elements in the spatio-temporal data cube is: ; where is the total number of elements, is the total number of gas types to be monitored, is the number of time steps, is the spatial grid size.

[0021] Step S200: Select time points in the spatio-temporal data cube, read the time co-subcube in the spatio-temporal data cube based on the selected time points, and determine the spatial features based on the time co-subcube. In the spatio-temporal data cube, time is a parameter. After removing the time parameter, the remaining data is still a multi-dimensional data, called the co-subcube. Since it is the co-subcube after removing the time parameter, it is called the time co-subcube. There are a large number of time points in the time parameter. The technical solution of the present invention does not need to analyze each time point. Only need to select the time co-subcubes of some time points, identify the time co-subcubes, and then determine the spatial features.

[0022] Specifically, regarding the cube concept in the above content, an example is given as follows: Suppose there are three types of gas data, and the collection location of the gas data is three-dimensional. Then the data in the spatio-temporal data cube is A[t][x][y][z][a][b][c], where t represents time, x, y, and z respectively represent the three coordinates, and a, b, and c respectively correspond to the three types. This is a seven-dimensional array. After extracting the time parameter, we get A[x][y][z][a][b][c], which is a six-dimensional data and is the co-subcube of the seven-dimensional array. Specifically, the concept of cofactor can be referred to.

[0023] Analyze A[x][y][z][a][b][c] to extract the state of the gas at different coordinates, and spatial features can be obtained. Generally speaking, spatial features are some boundaries. For example, the positions where the gas concentration changes abruptly.

[0024] Step S300: Select spatial positions in the spatio-temporal data cube based on the spatial features, read the position co-subcube based on the selected spatial positions, and the position co-subcube determines the concentration change information of each type of gas; After the spatial features are determined, some important positions are selected in the entire area, and then, taking the positions as a reference, the co-subcube is extracted. Taking the above example as an example, it is to extract A[t][a][b][c]. Analyzing these data can obtain the change situation of each type of gas, that is, the concentration change information of each type of gas.

[0025] Step S400: Mark the gas according to the concentration change information of the gas, query the preset characteristic convolution kernel of the gas, traverse the spatio-temporal data cube, match the abnormal time and abnormal position, and feedback to the management end; After obtaining the change situation of the gas at the important positions, select the gas with relatively abnormal change situation, query its type, and consider that this type needs to be further detected. Query the preset characteristic convolution kernel of this gas, traverse the entire spatio-temporal data cube. If the match is successful, record the abnormal time and abnormal position. The abnormal time is a time period, and the abnormal position is generally a range. Feedback the abnormal time and abnormal position to the management end.

[0026] Among them, the characteristic convolution kernel is preset. Before the application of the technical solution of the present invention, the staff will pre-simulate some dangerous states (the time range and space range are generally relatively small), record the gas concentration in the same way, and obtain the spatio-temporal data cube in the dangerous state. Since the time range and space range are generally relatively small, the ranges of the parameters of this simulated spatio-temporal data cube are all small, which is called the characteristic convolution kernel. This can be analogized to the image convolution kernel. For example, in an image of 1920*1080, a 16*16 convolution kernel can be used. The image is two-dimensional, and the spatio-temporal data cube in the technical solution of the present invention is multi-dimensional.

[0027] Regarding step S100, the steps of receiving the gas concentration containing time, position and type collected by the AMC monitor and constructing the spatio-temporal data cube include: Obtain the position of the monitor at each moment according to the position collector built in the AMC monitor; Regularly receive the gas concentration containing time and type obtained by the AMC monitor; Count the gas concentration of each type at each position according to the time and construct the spatio-temporal data cube.

[0028] The AMC monitor is built-in with a locator (position collector) for obtaining the position at each moment, and a built-in acquisition module for obtaining the gas concentrations of different types of gases. The AMC monitor also has a clock module (basic module) that can record time. Thus, the position at different moments and the gas concentrations of different types can be obtained. The gas concentrations of each type at each position are statistically analyzed according to time to construct a spatio-temporal data cube.

[0029] It is worth mentioning that when obtaining the position, the position change rate is calculated in real time, and the position acquisition frequency is adjusted according to the change rate; the position acquisition frequency is proportional to the position change rate. When the AMC monitor is installed on a fixed facility, the position change rate is zero, and the position acquisition frequency takes the minimum value. Similarly, when obtaining the gas concentration, the concentration change rate is calculated in real time, and the acquisition frequency of the gas is also proportional to the concentration change rate. When the gas is stable, the acquisition frequency is low. However, the difference is that the importance of different gases is different, which is represented by a reference change rate. The reference change rate is related to the importance of the gas. The more important the gas is, the higher the reference change rate is, and correspondingly, the higher the acquisition frequency is.

[0030] Specifically, the steps of statistically analyzing the gas concentrations of each type at each position according to time to construct a spatio-temporal data cube include: Receiving the time step set by the management personnel to determine the time dimension within the preset time range; Receiving the position calibration grid set by the management personnel to determine the coordinate dimensions of each coordinate; the grid cell size in the position calibration grid is a preset value; the coordinates do not exceed three parameters; Receiving the total number of types set by the management personnel as the type dimension; Statistically analyzing the time dimension, coordinate dimension, and type dimension to construct a cube structure; Statistically analyzing the gas concentrations containing time, position, and type based on the cube structure to obtain an intermediate data cube; Performing interpolation in the intermediate data cube based on a preset interpolation algorithm to fill in the empty data positions to obtain a spatio-temporal data cube.

[0031] In an example of the technical solution of the present invention, the data statistical process is described. The time step set by the management personnel is received, that is, how often to perform statistics, for example, 10s, thereby determining the time dimension within the preset time range, for example, how many time points are obtained in a day; the position calibration grid set by the management personnel is received and inserted into the area to be detected. The position calibration grid indicates how many positions are selected in the area to be detected. The size of the grid unit in the position calibration grid is a preset value. The smaller the size of the grid unit, the more positions are selected. In practical applications, the position calibration grid can be two-dimensional or three-dimensional. If the gas concentration is obtained only in the plane, it is two-dimensional. If the gas concentration is obtained in space, it is three-dimensional.

[0032] The time dimension, coordinate dimension, and type dimension are statistically analyzed to construct a cube structure. Based on the cube structure, the gas concentration containing time, position, and type is statistically analyzed. Since the position and acquisition frequency of the AMC detector are variable, therefore, the acquisition of gas concentration is discontinuous both in time and space. This makes only some data in the cube structure real data, called the transfer data cube; interpolation is performed in the transfer data cube based on a preset interpolation algorithm to fill the empty data positions, and a spatio-temporal data cube is obtained.

[0033] Regarding the method of filling the empty data positions, a conventional interpolation process can be used. For a certain type of gas, spatial interpolation can be performed on the gas concentration at each moment, and each empty data is set to the nearest known data. Of course, a diffusion model may also be introduced to predict it; temporal interpolation can also be performed on the gas concentration at each position, query the gas concentration at the known moment, fit the gas concentration function, and then determine the gas concentration at the unknown moment.

[0034] Regarding step S200, the steps of selecting a time point in the spatio-temporal data cube, reading a time co-subcube in the spatio-temporal data cube based on the selected time point, and determining the spatial feature based on the time co-subcube include: Extract time parameters and query the subcube corresponding to each time point in the time parameters; Input the subcube into a preset numerical normalization model to convert it into a feature value; Arrange the feature values according to the time sequence, fit the feature values, and obtain a time variation function; Calculate the derivative function of a preset order of the time variation function, and query the function value and derivative value of the time point in the time variation function and all its derivative functions; Cluster the time points based on the function value and derivative value, select the category of time points with the largest number of time points as the selected time points; Read the subcube corresponding to the selected time points as the time co-subcube; Perform gradient recognition on each time co-subcube to determine spatial features.

[0035] In an example of the technical solution of the present invention, a specific data screening solution is provided. The key point is to select important time points among all time points within a cycle. The specific operations are as follows: Extract time parameters, query the subcube corresponding to each time point in the time parameters, input the subcube into a preset numerical normalization model to convert it into an eigenvalue. The numerical normalization model is used to convert a multi-dimensional array into a single value. First, perform non-dimensionalization processing on all data, and then perform summation. It has no practical significance, and the specific normalization method is not limited here. Since each time point corresponds to a numerical value, arrange the eigenvalues according to the time sequence, fit the eigenvalues to obtain a time variation function, calculate the derivative function of a preset order of the time variation function, query the function value and derivative value of the time point in the time variation function and all its derivative functions. The order of the derivative function is at least one and at most three. Any more would not have much meaning and would make the subsequent clustering results more diverse; cluster the time points based on the function value and derivative value, and time points with sufficiently similar states can be grouped into one category. Select the category with the largest number of time points as the selected time points, and read the subcubes corresponding to the selected time points as time co-subcubes.

[0036] The time influence of the subcube selected through the above content is extremely low. When used to identify spatial features, the influence of time is very small; perform gradient recognition on each time co-subcube to obtain spatial features.

[0037] As an example of the technical solution of the present invention, the steps of performing gradient recognition on each time co-subcube to determine spatial features include: Successively extract various types of data in each time co-subcube to obtain a data distribution matrix; Perform gradient recognition on the data distribution matrix to determine the spatial profile; Statistically analyze all types of spatial profiles of all time points, and merge the spatial profiles to obtain a comprehensive profile; the retention degree is determined according to the number of spatial profiles during the merging process; Select the comprehensive profile with a retention degree reaching a preset retention threshold as the spatial feature.

[0038] Taking the above example as a reference, successively extract various types of data in each time co-subcube. The obtained data is A[t][x][y][z][a], where t and a are known and are fixed values. It is actually a[x][y][z], that is, the gas concentration of gas type a at different positions, which is called the data distribution matrix. Each type at each time point can obtain a data distribution matrix.

[0039] Perform gradient recognition on the data distribution matrix to determine the spatial profile; count all types of spatial profiles at all time points, merge the spatial profiles, and obtain a comprehensive profile. Among them, the merging process is to merge profiles that are close enough into one profile, record how many data distribution matrix results each comprehensive profile is merged from, determine the retention degree, the larger the number, the higher the retention degree, eliminate the comprehensive profiles with a lower retention degree (such as spatial profiles that only appear in a few data distribution matrices), and select the comprehensive profiles whose retention degree reaches the preset retention threshold as spatial features.

[0040] Regarding step S300, the steps of selecting spatial positions in the spatio-temporal data cube based on the spatial features and reading the position co-subcube, where the position co-subcube determines the concentration change information of each type of gas, include: Read the comprehensive profile in the spatial features and divide the area into sub-areas; Determine the quantity according to the area of each sub-area, and randomly select that quantity of spatial positions in the sub-area; the quantity is proportional to the area; Query the gas concentrations of each type at each time at each spatial position as the position co-subcube; the position co-subcube is a two-dimensional matrix; Extract the gas concentrations of each type of gas at all times in the position co-subcube and input them into a preset fluctuation recognition model to obtain the fluctuation degree of that type; Count the fluctuation degrees of all types as the concentration change information of this spatial position.

[0041] In an example of the technical solution of the present invention, read the comprehensive profile in the spatial features, divide the area into sub-areas, determine the quantity according to the area of each sub-area, the larger the area, the more the quantity, randomly select that quantity of spatial positions in the sub-area, query the gas concentrations of each type at each time at each spatial position as the position co-subcube; taking the above example as a reference, the gas concentrations of each type at each time at each spatial position are A[t][a], A[t][b], and A[t][c], which actually obtain a[t], b[t], and c[t], which are some time series, input them into a preset fluctuation recognition model to obtain the fluctuation degree of that type, and the fluctuation recognition model can use an existing fluctuation recognition model, which is generally related to variance and periodicity, count the fluctuation degrees of all types as the concentration change information of this spatial position; in practical applications, the types of gases to be monitored are far more than three, so there will be many obtained fluctuation degrees.

[0042] Regarding step S400, the steps of marking the gas according to the gas concentration change information, querying the preset characteristic convolution kernel of the gas, traversing the spatio-temporal data cube, matching the abnormal time and abnormal position, and feeding back to the management end include: Query the concentration change information of the selected spatial position, and mark the types with the fluctuation degree reaching the preset fluctuation threshold; Record the total number of marks of each type. When the total number of marks reaches the preset quantity, query the characteristic convolution kernel corresponding to this type; Based on the characteristic convolution kernel, traverse the spatio-temporal data cube, match the abnormal time and abnormal position, and feed back to the management end.

[0043] In an example of the technical solution of the present invention, each spatial position is analyzed separately. Query the fluctuation degrees of all types of each spatial position, mark the types with the fluctuation degree reaching the preset fluctuation threshold, and analyze each spatial position in turn. Record the total number of marks of each type. When the total number of marks is large enough, it indicates that the gas of this type appears abnormally at multiple positions in the entire area. At this time, query the characteristic convolution kernel corresponding to this type; Based on the characteristic convolution kernel, traverse the spatio-temporal data cube, match the abnormal time and abnormal position, and feed back to the management end; Combining the description of the characteristic convolution kernel in the above content, the characteristic convolution kernel is the gas concentration of this type of gas within a period of time under the preset dangerous conditions.

[0044] Combined with Figure 2 , the technical solution of the present invention also provides a multi-channel spatio-temporal slice fusion AMC concentration information extraction system. The system 10 includes: A cube construction module 11, configured to receive the gas concentration containing time, position, and type collected by the AMC monitor, and construct a spatio-temporal data cube; The total number of elements of the spatio-temporal data cube is: ; In the formula, is the total number of elements, is the total number of gas types to be monitored, is the number of time steps, is the spatial grid size; A spatial feature determination module 12, configured to select a time point in the spatio-temporal data cube, read a time co-subcube in the spatio-temporal data cube based on the selected time point, and determine a spatial feature based on the time co-subcube; A gas concentration analysis module 13, configured to select a spatial position in the spatio-temporal data cube based on the spatial feature, read a position co-subcube based on the selected spatial position, and determine the concentration change information of each type of gas from the position co-subcube; Anomaly matching module 14 is used to mark gases according to the gas concentration change information, query the preset characteristic convolution kernels of the gases, traverse the spatio-temporal data cube, match the abnormal time and abnormal position, and feedback to the management terminal.

[0045] Specifically, the cube construction module 11 includes: A position acquisition unit for acquiring the position of the monitor at each moment according to the position collector built in the AMC monitor; A gas concentration receiving unit for regularly receiving the gas concentration containing time and type obtained by the AMC monitor; A data statistics unit for statistically calculating the gas concentrations of various types at each position according to time and constructing a spatio-temporal data cube; Among them, when acquiring the position, the position change rate is calculated in real time, and the position acquisition frequency is adjusted according to the change rate; when acquiring the gas concentration, the concentration change rate is calculated in real time, and the acquisition frequency of the gas concentrations of various types is adjusted based on the concentration change rate and the reference change rate determined by the type.

[0046] Specifically, the spatial feature determination module 12 includes: A data query unit for extracting time parameters and querying the sub-cubes corresponding to each time point in the time parameters; A data normalization unit for inputting the sub-cube into a preset numerical normalization model and converting it into a feature value; A feature fitting unit for arranging the feature values according to the time sequence, fitting the feature values, and obtaining a time change function; A numerical calculation unit for calculating the derivative function of a preset order of the time change function and querying the function values and derivative values of the time points in the time change function and all its derivative functions; A time point classification unit for clustering the time points based on the function values and derivative values, and selecting the category of time points with the largest number of time points as the selected time points; A data output unit for reading the sub-cube corresponding to the selected time points as the time co-sub-cube; A gradient recognition unit for performing gradient recognition on each time co-sub-cube to determine the spatial features.

[0047] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied to other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A method for extracting AMC concentration information by multi-channel spatio-temporal slice fusion, characterized in that The method includes: Receive the gas concentration containing time, location, and type collected by the AMC monitor, and construct a spatio-temporal data cube; the total number of elements in the spatio-temporal data cube is: ; where is the total number of elements, is the total number of gas types to be monitored, is the number of time steps, is the spatial grid size; Selecting time points in the spatio-temporal data cube, reading time co-sub-cubes in the spatio-temporal data cube based on the selected time points, and determining spatial features based on the time co-sub-cubes; Selecting spatial positions in the spatio-temporal data cube based on the spatial features, reading position co-sub-cubes based on the selected spatial positions, and determining the concentration change information of each type of gas from the position co-sub-cubes; Marking the gas according to the concentration change information of the gas, querying the preset feature convolution kernels of the gas, traversing the spatio-temporal data cube, matching abnormal times and abnormal positions, and feeding back to the management terminal.

2. The method for extracting AMC concentration information by multi-channel spatio-temporal slice fusion according to claim 1, wherein The step of receiving the gas concentrations containing time, position, and type collected by the AMC monitor and constructing the spatio-temporal data cube includes: Obtaining the position of the monitor at each moment according to the position collector built in the AMC monitor; Regularly receiving the gas concentrations containing time and type obtained by the AMC monitor; Counting the gas concentrations of each type at each position according to time, and constructing the spatio-temporal data cube; Among them, when obtaining the position, calculating the position change rate in real time, and adjusting the position acquisition frequency according to the change rate; when obtaining the gas concentration, calculating the concentration change rate in real time, and adjusting the acquisition frequency of the gas concentration of each type based on the concentration change rate and the reference change rate determined by the type.

3. The method for extracting AMC concentration information by multi-channel spatio-temporal slice fusion according to claim 2, wherein The step of counting the gas concentrations of each type at each position according to time and constructing the spatio-temporal data cube includes: Receiving the time step set by the management personnel and determining the time dimension within the preset time range; Receiving the position calibration grid set by the management personnel and determining the coordinate dimensions of each coordinate; the grid cell size in the position calibration grid is a preset value; the coordinates do not exceed three parameters; Receiving the total number of types set by the management personnel as the type dimension; Counting the time dimension, coordinate dimension, and type dimension, and constructing a cube structure; Counting the gas concentrations containing time, position, and type based on the cube structure to obtain an intermediate data cube; Performing interpolation in the intermediate data cube based on a preset interpolation algorithm to fill in the empty data positions and obtain the spatio-temporal data cube.

4. The method for extracting AMC concentration information by multi-channel spatio-temporal slice fusion according to claim 1, characterized in that, The step of selecting time points in the spatio-temporal data cube, reading time co-sub-cubes in the spatio-temporal data cube based on the selected time points, and determining spatial features based on the time co-sub-cubes includes: Extracting time parameters and querying the sub-cubes corresponding to each time point in the time parameters; Inputting the sub-cubes into a preset numerical normalization model to convert them into eigenvalue; Arranging the eigenvalues according to time sequence, fitting the eigenvalues, and obtaining a time variation function; Calculating the derivative function of a preset order of the time variation function, and querying the function values and derivative values of the time points in the time variation function and all its derivative functions; Clustering the time points based on the function values and derivative values, and selecting the category of time points with the largest number of time points as the selected time points; Reading the sub-cubes corresponding to the selected time points as time co-sub-cubes; Performing gradient recognition on each time co-sub-cube to determine spatial features.

5. The method for extracting AMC concentration information by multi-channel spatio-temporal slice fusion according to claim 4, wherein The step of performing gradient recognition on each time co-sub-cube to determine spatial features includes: Extract various types of data in each time co - sub - cube in sequence to obtain a data distribution matrix; Perform gradient recognition on the data distribution matrix to determine the spatial profile; Statistically analyze all spatial profiles of all types at all time points, merge the spatial profiles, and obtain a comprehensive profile; the retention degree is determined according to the number of spatial profiles during the merging process; Select the comprehensive profiles whose retention degree reaches the preset retention threshold as spatial features.

6. The method for extracting AMC concentration information by multi-channel spatio-temporal slice fusion according to claim 1, wherein The steps of selecting spatial positions in the spatio - temporal data cube based on the spatial features, reading the position co - sub - cube based on the selected spatial positions, and the position co - sub - cube determining the concentration change information of various types of gases include: Read the comprehensive profile in the spatial feature and divide the area into sub - areas; Determine the number according to the area of each sub - area, and randomly select this number of spatial positions in the sub - area; the number is proportional to the area; Query the gas concentrations of various types at each time at each spatial position as the position co - sub - cube; the position co - sub - cube is a two - dimensional matrix; Extract the gas concentrations of each type of gas at all times in the position co - sub - cube and input them into a preset fluctuation recognition model to obtain the fluctuation degree of this type; Statistically analyze the fluctuation degrees of all types as the concentration change information of this spatial position.

7. The method for extracting AMC concentration information by multi-channel spatio-temporal slice fusion according to claim 6, wherein The steps of marking the gas according to the concentration change information of the gas, querying the preset characteristic convolution kernel of the gas, traversing the spatio - temporal data cube, matching the abnormal time and abnormal position, and feeding back to the management terminal include: Query the concentration change information of the selected spatial positions, and mark the types whose fluctuation degree reaches the preset fluctuation threshold; Record the total number of marks of each type. When the total number of marks reaches the preset quantity, query the characteristic convolution kernel corresponding to this type; Traverse the spatio - temporal data cube based on the characteristic convolution kernel, match the abnormal time and abnormal position, and feed back to the management terminal; Among them, the characteristic convolution kernel is the gas concentration of this type of gas within a period of time under preset dangerous conditions.

8. An AMC concentration information extraction system with multi-channel spatio-temporal slice fusion, characterized in that, The system includes: A cube construction module is used to receive gas concentrations containing time, location, and type collected by an AMC monitor and construct a spatio-temporal data cube; the total number of elements of the spatio-temporal data cube is: ; where is the total number of elements, is the total number of gas types to be monitored, is the number of time steps, is the spatial grid size; A spatial feature determination module, which is used to select time points in the spatio - temporal data cube, read the time co - sub - cube in the spatio - temporal data cube based on the selected time points, and determine spatial features based on the time co - sub - cube; A gas concentration analysis module, which is used to select spatial positions in the spatio - temporal data cube based on the spatial features, read the position co - sub - cube based on the selected spatial positions, and the position co - sub - cube determines the concentration change information of various types of gases; An abnormal matching module, which is used to mark the gas according to the concentration change information of the gas, query the preset characteristic convolution kernel of the gas, traverse the spatio - temporal data cube, match the abnormal time and abnormal position, and feed back to the management terminal.

9. The AMC concentration information extraction system for multi-channel spatio-temporal slice fusion according to claim 8, characterized in that, The cube construction module includes: A position acquisition unit, which is used to acquire the position of the monitor at each moment according to the position collector built in the AMC monitor; A gas concentration receiving unit, which is used to regularly receive the gas concentrations containing time and type obtained by the AMC monitor; A data statistics unit, which is used to statistically analyze the gas concentrations of various types at each position according to time and construct a spatio - temporal data cube; When obtaining the position, the position change rate is calculated in real time, and the position acquisition frequency is adjusted according to the change rate; when obtaining the gas concentration, the concentration change rate is calculated in real time, and the acquisition frequency of the gas concentration of each type is adjusted based on the concentration change rate and the reference change rate determined by the type.

10. The AMC concentration information extraction system for multi-channel spatio-temporal slice fusion according to claim 8, wherein The spatial feature determination module includes: A data query unit, configured to extract time parameters and query the sub-cubes corresponding to each time point in the time parameters; A data normalization unit, configured to input the sub-cubes into a preset numerical normalization model and convert them into feature values; A feature fitting unit, configured to arrange the feature values in chronological order, fit the feature values, and obtain a time variation function; A numerical calculation unit, configured to calculate the derivative function of a preset order of the time variation function, and query the function values and derivative values of the time points in the time variation function and all its derivative functions; A time point classification unit, configured to cluster the time points based on the function values and derivative values, and select the category of time points with the largest number of time points as the selected time points; A data output unit, configured to read the sub-cubes corresponding to the selected time points as time co-sub-cubes; A gradient recognition unit, configured to perform gradient recognition on each time co-sub-cube to determine the spatial features.

Images