A method and system for monitoring ambient air quality based on multi-source data fusion

Abstract

The present invention discloses an environmental air quality monitoring method and system based on multi-source data fusion, which relates to the field of environmental monitoring technology, including: pulling traffic flow data, geographic information data, air quality data and meteorological data in the monitoring area to the monitoring system; the monitoring system quantifies and fuses the pulled data to form a multi-source data set; uses a spatiotemporal scale analysis algorithm to analyze the multi-source data set, and estimates the influence coefficient of various types of data on air quality; creates an air quality prediction model based on the estimated values ​​of the influence coefficients of various types of data on air quality; uses the air quality prediction model to predict the change trend of air quality in the monitoring area, and generates an air quality monitoring report; uses the real data generated in the monitoring area to regularly tune the air quality prediction model. By fusing information from multiple data sources, the distribution and changes of air quality at different time and space scales are obtained, providing more comprehensive and accurate monitoring information.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental monitoring, and in particular to an environmental air quality monitoring method and system based on multi-source data fusion. Background Art

[0002] Air quality monitoring data is an important basis for environmental management. Through monitoring results, the crux of environmental problems can be discovered in time, and targeted measures can be provided for environmental improvement and governance. At present, the most widely used monitoring method is to manually collect pollutants in the air and analyze the samples. The advantages of this method are simple operation and easy to master, and it is not affected by environmental factors, but it requires more manpower and material resources, and the accuracy of the samples is not high enough. In addition, the existing air quality monitoring system has deficiencies in the feedback mechanism after monitoring, the early warning execution is weak, there is a long time difference in response to feedback, and there is a lack of a unified system for monitoring and early warning linkage. In order to solve the above problems, technical personnel in this field urgently need to develop an environmental air quality monitoring method with high measurement accuracy and integrated monitoring and early warning. Summary of the invention

[0003] The present invention provides an ambient air quality monitoring method based on multi-source data fusion, comprising:

[0004] Step 1: Pull traffic flow data, geographic information data, air quality data and meteorological data in the monitoring area to the monitoring system;

[0005] Step 2: The monitoring system quantifies and integrates the acquired data to form a multi-source data set.

[0006] Step 3: Use the spatiotemporal scale analysis algorithm to analyze the multi-source data set and estimate the impact coefficient of various data on air quality;

[0007] Step 4: Create an air quality prediction model based on the estimated values ​​of the impact coefficients of various data on air quality;

[0008] Step 5. Use the air quality prediction model to predict the changing trend of air quality in the monitoring area and generate an air quality monitoring report;

[0009] Step 6: Use the real data generated in the monitoring area to regularly tune the air quality prediction model.

[0010] As described above, an ambient air quality monitoring method based on multi-source data fusion, in which the monitoring system quantifies and fuses the pulled data to form a multi-source data set, is specifically divided into the following sub-steps:

[0011] Use the system's built-in vector generator to generate the pulled data into vectors;

[0012] Input the generated vectors into the unified quantization function in sequence, and return the quantization result of the pulled data;

[0013] The quantified data are aligned in time and space and stored as a multi-source dataset.

[0014] As described above, an ambient air quality monitoring method based on multi-source data fusion is described, in which a spatiotemporal scale analysis algorithm is used to analyze the multi-source data set and estimate the influence coefficient of each type of data on air quality. The method is specifically divided into the following sub-steps:

[0015] Fixed time and space scales, estimated the impact coefficient of traffic flow data on air quality;

[0016] Fix the spatial scale, adjust the time scale, and estimate the impact coefficient of meteorological data on air quality;

[0017] Fix the time scale, adjust the spatial scale, and estimate the impact coefficient of geographic information data on air quality.

[0018] As described above, an ambient air quality monitoring method based on multi-source data fusion is provided, wherein an air quality prediction model is created based on the estimated values ​​of the influence coefficients of various types of data on air quality, which is specifically divided into the following sub-steps:

[0019] Extract input-output pairs from multi-source data sets and organize them into training data sets;

[0020] Use the training data set to train the air quality prediction model and determine the growth coefficient of each type of data over time;

[0021] The trained air quality prediction model is deployed to the monitoring area.

[0022] As described above, an ambient air quality monitoring method based on multi-source data fusion, in which the air quality prediction model is regularly tuned using real data generated in the monitoring area, is specifically divided into the following sub-steps:

[0023] Regularly collect real data generated within the monitoring area;

[0024] Process the collected data into a validation dataset;

[0025] Evaluate the performance changes of air quality prediction models on validation datasets;

[0026] The air quality prediction model is tuned based on the performance evaluation results.

[0027] The present invention also provides an ambient air quality monitoring system based on multi-source data fusion, comprising:

[0028] Multi-source data extraction module, used to extract traffic flow data, geographic information data, air quality data and meteorological data within the monitoring area;

[0029] The multi-source data fusion module is used to quantify and fuse the pulled data to form a multi-source data set;

[0030] Multi-source data analysis module, which is used to analyze multi-source data sets using spatiotemporal scale analysis algorithms and estimate the impact coefficient of various data on air quality;

[0031] An air quality prediction model creation module is used to create an air quality prediction model based on the estimated values ​​of the impact coefficients of various data on air quality;

[0032] An air quality monitoring report generation module is used to use an air quality prediction model to predict the changing trend of air quality in the monitoring area and generate an air quality monitoring report;

[0033] The air quality prediction model tuning module is used to regularly tune the air quality prediction model using real data generated in the monitoring area.

[0034] The beneficial effects achieved by the present invention are as follows: by fusing information from multiple data sources, the distribution and changes of air quality at different time and spatial scales can be obtained, providing more comprehensive and accurate monitoring information; by predicting changes in air quality, air quality warnings can be issued in advance, allowing the public to take measures in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0036] Figure 1 This is a flow chart of an ambient air quality monitoring method for multi-source data fusion provided in Example 1 of the present application;

[0037] Figure 2 This is a schematic diagram of an environmental air quality monitoring system with multi-source data fusion provided in Example 2 of the present application. DETAILED DESCRIPTION

[0038] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0039] Embodiment 1

[0040] like Figure 1 As shown, the first embodiment of the present application provides an ambient air quality monitoring method based on multi-source data fusion, including:

[0041] Step S10: Pull traffic flow data, geographic information data, air quality data and meteorological data in the monitoring area to the monitoring system;

[0042] By connecting to the data export API of different data source systems to the monitoring system, different types of monitoring data can be obtained in real time. The data pulled into the monitoring system will be stored in different data tables according to its type, and each piece of data must record its collection location and collection time. In order to save the storage space of the monitoring system, monitoring data that is too old will be moved to other databases for storage.

[0043] Step S20: The monitoring system quantifies and fuses the pulled data to form a multi-source data set;

[0044] Since the data formats and measurement standards are different, it is impossible to directly analyze the connections between them. Therefore, in this step, the data needs to be processed uniformly and merged into a large data set. The specific steps include the following:

[0045] Step S21: Generate the pulled data into a vector using the system's built-in vector generator;

[0046] The built-in vector generator is a vector generation tool class ported from the NumPy library or Mahout, which can generate input text into vectors.

[0047] Step S22: input the generated vectors into the unified quantization function in sequence, and return the quantization result of the pulled data;

[0048] The unified quantization function is expressed as: , where C is the incoming vector, i.e., the input parameter, , , are the components of the input parameter C on the X, Y, and Z axes respectively, , , The components of the vector formed by the maximum value of the data described by the input parameter C in the historical pull record on the X, Y, and Z axes respectively. , , The components of the vector formed by the minimum value of the data described by the input parameter C in the historical pull record on the X, Y, and Z axes respectively.

[0049] Step S23: aligning the quantified data in time and space and storing them as a multi-source data set;

[0050] Spatiotemporal alignment means storing data with the same collection location and time in the same storage unit. The storage unit can be an array or a sub-dataset. Data pulled from different data sources are labeled with different characters for easy distinction.

[0051] Step S30: Analyze the multi-source data set using a spatiotemporal scale analysis algorithm to estimate the impact coefficient of each type of data on air quality;

[0052] The spatiotemporal scale analysis algorithm is used to analyze the correlation between data from different data sources and air quality at spatiotemporal scales, so as to estimate the impact coefficient of various data on air quality. It specifically includes the following sub-steps:

[0053] Step S31: Fixing the time and space scale, estimating the impact coefficient of traffic flow data on air quality;

[0054] The scale on the time scale is the collection time, and each scale interval is one hour. The scale on the spatial scale is the collection location, and there is no interval. Fixed time and space scale means freezing the time in a certain hour and fixing the location at a collection location. In order to exclude other influencing factors as much as possible, when choosing a fixed collection location, the most stable weather location should be selected, and when choosing the freeze time, the hour with the smallest change in weather data should be selected. Extract the traffic flow data and air quality data at the collection location corresponding to the fixed scale of the spatial scale and the collection time corresponding to the fixed scale of the time scale from the multi-source data set, and bring them into the formula: The influence coefficient of traffic flow on air quality is obtained ,in , They are the traffic flow data and air quality data collected at the i+1th collection time contained in the fixed time scale, , is the traffic flow data and air quality data collected at the i-th collection time contained in the fixed time scale. i ranges from 1 to n-1, and n is the number of collection times contained in the fixed time scale. A formal parameter is set to avoid the denominator being zero, and the value is one hundred thousandth.

[0055] Step S32: fix the spatial scale, adjust the time scale, and estimate the impact coefficient of meteorological data on air quality;

[0056] First, fix the spatial scale to ignore the influence of geographic information, then adjust the time scale, read the meteorological data, traffic flow data, and air quality data at each time scale, and bring them into the formula: The estimated value of the impact coefficient of meteorological data on air quality is obtained ,in , , They represent the meteorological data, air quality data, and traffic flow data at the j+1th time scale, respectively. , , They represent the meteorological data, air quality data, and traffic flow data at the jth time scale, respectively. j ranges from 1 to m-1, and m is the total length of the time scale. is the impact coefficient of traffic flow data on air quality.

[0057] Step S33: fix the time scale, adjust the spatial scale, and estimate the influence coefficient of geographic information data on air quality;

[0058] The time scale is fixed to reduce the impact of factors other than geographic information on air quality, and then the geographic information data, meteorological data, and traffic flow data at different spatial scales (i.e. different collection locations) are read and brought into the formula:

[0059] The estimated value of the impact coefficient of geographic information on air quality is obtained ,in , , , They represent the geographic information data, air quality data, traffic flow data and meteorological data at the k+1th spatial scale respectively. , , , They represent geographic information data, air quality data, traffic flow data, and meteorological data on the kth spatial scale, respectively. k ranges from 1 to w-1, and w is the total length of the spatial scale. , They are the impact coefficients of traffic flow data and meteorological data on air quality.

[0060] Step S40: creating an air quality prediction model based on the estimated values ​​of the influence coefficients of various types of data on air quality;

[0061] According to the real-time data of various data and their influence coefficients on air quality, real-time air quality data can be obtained. To predict air quality, the growth coefficient of various data over time needs to be considered, which includes the following sub-steps:

[0062] Step S41: extracting input-output pairs from the multi-source data sets and arranging them into training data sets;

[0063] First, the multi-source data sets are grouped according to the collection location, and the data in the group are sorted according to the collection time; then the geographic information data, traffic flow data, and meteorological data within the 1st to 10th hour in the group are extracted as input, and the air quality data from the 11th to 20th hour is used as output to form the first input-output pair. Then, the geographic information data, traffic flow data, and meteorological data within the 2nd to 11th hour are used as input, and the air quality data from the 12th to 21st hour is used as output to form the second input-output pair, and so on, until all the input-output pairs contained in all groups are extracted, and the extracted input-output pairs are stored in the training data set.

[0064] Step S42: using the training data set to train the air quality prediction model and determine the growth coefficient of each type of data over time;

[0065] During the training process, the growth coefficients of various types of data over time are iterated in accordance with the minimum loss principle to obtain the optimal solution for each growth coefficient.

[0066] The air quality prediction model is expressed as:

[0067] , where Y is the prediction result, , , The input data are the traffic flow data, meteorological data, and geographic information data collected at the time t. , , are the growth coefficients of traffic flow data, meteorological data, and geographic information data over time, , , They are the influence coefficients of traffic flow data, meteorological data, and geographic information data on air quality, respectively. t ranges from 1 to T, where T is the length of the prediction period.

[0068] Step S43: deploying the trained air quality prediction model to the monitoring area;

[0069] It should be noted that the training data set used in each monitoring area is different, so the trained air quality prediction model is only applicable to the respective monitoring area.

[0070] Step S50: using the air quality prediction model to predict the change trend of the air quality in the monitoring area and generate an air quality monitoring report;

[0071] Each report record in the air quality monitoring report includes three parts, namely, air quality monitoring points, real-time air quality monitoring data and air quality prediction data, among which the air quality monitoring points are the collection sites of multi-source data; first, the monitoring system will group the multi-source data sets according to the collection sites; then extract the traffic flow data, meteorological data and geographic information data within the most recent prediction period T in each group and input them into the air quality prediction model respectively, and output the prediction data of air quality at different collection sites; finally, the names of different collection sites, the real-time monitoring data and prediction data of air quality at the collection sites are combined into report records to form a complete air quality monitoring report.

[0072] Step S60: Regularly tune the air quality prediction model using real data generated in the monitoring area;

[0073] The real data generated in the monitoring area is an important basis for model tuning. Regular use of real data to tune the air quality prediction model to ensure the timeliness and accuracy of the model includes the following sub-steps:

[0074] Step S61: Regularly collect real data generated in the monitoring area;

[0075] Includes recent traffic flow data, meteorological data, geographic information data, and air quality data in the detection area.

[0076] Step S62: Processing the collected data into a verification data set;

[0077] The processing method is consistent with the model's training dataset.

[0078] Step S63: evaluating the performance change of the air quality prediction model on the validation data set;

[0079] The inverse of the MES loss function is taken to evaluate the performance of the model on the training dataset.

[0080] Step S64: optimizing the air quality prediction model according to the performance evaluation result;

[0081] If the performance of the model is found to be declining, the validation data set is used to perform incremental training on the model. During the training process, the various parameters of the model are adjusted to minimize the MES loss value to achieve the effect of model optimization. The tuned air quality prediction model is used to replace the original model to complete the model version upgrade. If the performance of the model on the validation data set is stable, the original model is still used for subsequent predictions.

[0082] Embodiment 2

[0083] like Figure 2 As shown, the second embodiment of the present application provides an ambient air quality monitoring system based on multi-source data fusion, including: a multi-source data pulling module 21, a multi-source data fusion module 22, a multi-source data parsing module 23, an air quality prediction model creation module 24, an air quality monitoring report generation module 25, and an air quality prediction model tuning module 26;

[0084] The multi-source data extraction module 21 is used to extract traffic flow data, geographic information data, air quality data and meteorological data in the monitoring area;

[0085] By connecting to the data export API of different data source systems to the monitoring system, different types of monitoring data can be obtained in real time. The data pulled into the monitoring system will be stored in different data tables according to its type, and each piece of data must record its collection location and collection time. In order to save the storage space of the monitoring system, monitoring data that is too old will be moved to other databases for storage.

[0086] The multi-source data fusion module 22 is used to quantize and fuse the pulled data to form a multi-source data set; it includes a vector generation submodule, a unified quantization submodule, and a spatiotemporal synchronization submodule;

[0087] The vector generation submodule is used to generate vectors from the pulled data using the built-in vector generator;

[0088] The built-in vector generator is a vector generation tool class ported from the NumPy library or Mahout, which can generate input text into vectors.

[0089] The unified quantization submodule is used to input the generated vectors into the unified quantization function in sequence and return the quantization results of the pulled data;

[0090] The unified quantization function is expressed as: , where C is the incoming vector, i.e., the input parameter, , , are the components of the input parameter C on the X, Y, and Z axes respectively, , , The components of the vector formed by the maximum value of the data described by the input parameter C in the historical pull record on the X, Y, and Z axes respectively. , , The components of the vector formed by the minimum value of the data described by the input parameter C in the historical pull record on the X, Y, and Z axes respectively.

[0091] The spatiotemporal synchronization submodule is used to align the quantified data in time and space and store them as a multi-source data set;

[0092] Spatiotemporal alignment means storing data with the same collection location and time in the same storage unit. The storage unit can be an array or a sub-dataset. Data pulled from different data sources are labeled with different characters for easy distinction.

[0093] The multi-source data analysis module 23 is used to analyze the multi-source data set using a spatiotemporal scale analysis algorithm to estimate the influence coefficient of various data on air quality; including a traffic flow data influence coefficient estimation submodule, a meteorological data influence coefficient estimation submodule, and a geographic information data influence coefficient estimation submodule;

[0094] Traffic flow data impact coefficient estimation submodule is used to estimate the impact coefficient of traffic flow data on air quality at a fixed time and space scale;

[0095] The scale on the time scale is the collection time, and each scale interval is one hour. The scale on the spatial scale is the collection location, and there is no interval. Fixed time and space scale means freezing the time in a certain hour and fixing the location at a collection location. In order to exclude other influencing factors as much as possible, when choosing a fixed collection location, the most stable weather location should be selected, and when choosing the freeze time, the hour with the smallest change in weather data should be selected. Extract the traffic flow data and air quality data at the collection location corresponding to the fixed scale of the spatial scale and the collection time corresponding to the fixed scale of the time scale from the multi-source data set, and bring them into the formula: The influence coefficient of traffic flow on air quality is obtained ,in , They are the traffic flow data and air quality data collected at the i+1th collection time contained in the fixed time scale, , is the traffic flow data and air quality data collected at the i-th collection time contained in the fixed time scale. i ranges from 1 to n-1, and n is the number of collection times contained in the fixed time scale. A formal parameter is set to avoid the denominator being zero, and the value is one hundred thousandth.

[0096] The meteorological data impact coefficient estimation submodule is used to fix the spatial scale, adjust the time scale, and estimate the impact coefficient of meteorological data on air quality;

[0097] First, fix the spatial scale to ignore the influence of geographic information, then adjust the time scale, read the meteorological data, traffic flow data, and air quality data at each time scale, and bring them into the formula: The estimated value of the impact coefficient of meteorological data on air quality is obtained ,in , , They represent the meteorological data, air quality data, and traffic flow data at the j+1th time scale, respectively. , , They represent the meteorological data, air quality data, and traffic flow data at the jth time scale, respectively. j ranges from 1 to m-1, and m is the total length of the time scale. is the impact coefficient of traffic flow data on air quality.

[0098] The geographic information data impact coefficient estimation submodule is used to fix the time scale, adjust the spatial scale, and estimate the impact coefficient of geographic information data on air quality;

[0099] The time scale is fixed to reduce the impact of factors other than geographic information on air quality, and then the geographic information data, meteorological data, and traffic flow data at different spatial scales (i.e. different collection locations) are read and brought into the formula: The estimated value of the impact coefficient of geographic information on air quality is obtained ,in , , , They represent the geographic information data, air quality data, traffic flow data and meteorological data at the k+1th spatial scale respectively. , , , They represent geographic information data, air quality data, traffic flow data, and meteorological data on the kth spatial scale, respectively. k ranges from 1 to w-1, and w is the total length of the spatial scale. , They are the impact coefficients of traffic flow data and meteorological data on air quality.

[0100] The air quality prediction model creation module 24 is used to create an air quality prediction model based on the estimated values ​​of the influence coefficients of various types of data on air quality; it includes a training data set creation submodule, a model training submodule, and a model delivery submodule;

[0101] The training data set creation submodule is used to extract input-output pairs from multi-source data sets and organize them into training data sets;

[0102] First, the multi-source data sets are grouped according to the collection location, and the data in the group are sorted according to the collection time; then the geographic information data, traffic flow data, and meteorological data within the 1st to 10th hour in the group are extracted as input, and the air quality data from the 11th to 20th hour is used as output to form the first input-output pair. Then, the geographic information data, traffic flow data, and meteorological data within the 2nd to 11th hour are used as input, and the air quality data from the 12th to 21st hour is used as output to form the second input-output pair, and so on, until all the input-output pairs contained in all groups are extracted, and the extracted input-output pairs are stored in the training data set.

[0103] The model training submodule is used to train the air quality prediction model using the training data set and determine the growth coefficient of various types of data over time;

[0104] During the training process, the growth coefficients of various types of data over time are iterated in accordance with the minimum loss principle to obtain the optimal solution for each growth coefficient.

[0105] The air quality prediction model is expressed as:

[0106] , where Y is the prediction result, , , The input data are the traffic flow data, meteorological data, and geographic information data collected at the time t. , , are the growth coefficients of traffic flow data, meteorological data, and geographic information data over time, , , They are the influence coefficients of traffic flow data, meteorological data, and geographic information data on air quality, respectively. t ranges from 1 to T, where T is the length of the prediction period.

[0107] The model delivery submodule is used to deliver the trained air quality prediction model to the monitoring area;

[0108] It should be noted that the training data set used in each monitoring area is different, so the trained air quality prediction model is only applicable to the respective monitoring area.

[0109] An air quality monitoring report generating module 25 is used to use an air quality prediction model to predict the changing trend of air quality in the monitoring area and generate an air quality monitoring report;

[0110] Each report record in the air quality monitoring report includes three parts, namely, air quality monitoring points, real-time air quality monitoring data and air quality prediction data, among which the air quality monitoring points are the collection sites of multi-source data; first, the monitoring system will group the multi-source data sets according to the collection sites; then extract the traffic flow data, meteorological data and geographic information data within the most recent prediction period T in each group and input them into the air quality prediction model respectively, and output the prediction data of air quality at different collection sites; finally, the names of different collection sites, the real-time monitoring data and prediction data of air quality at the collection sites are combined into report records to form a complete air quality monitoring report.

[0111] An air quality prediction model tuning module 26, for regularly tuning the air quality prediction model using real data generated in the monitoring area;

[0112] The real data generated in the monitoring area is an important basis for model tuning. Regular use of real data to tune the air quality prediction model to ensure the timeliness and accuracy of the model includes the following sub-steps:

[0113] 1. Regularly collect real data generated in the monitoring area;

[0114] Includes recent traffic flow data, meteorological data, geographic information data, and air quality data in the detection area.

[0115] 2. Process the collected data into a validation dataset;

[0116] The processing method is consistent with the model's training dataset.

[0117] 3. Evaluate the performance changes of the air quality prediction model on the validation dataset;

[0118] The inverse of the MES loss function is taken to evaluate the performance of the model on the training dataset.

[0119] 4. Optimize the air quality prediction model based on the performance evaluation results;

[0120] If the performance of the model is found to be declining, the validation data set is used to perform incremental training on the model. During the training process, the various parameters of the model are adjusted to minimize the MES loss value to achieve the effect of model optimization. The tuned air quality prediction model is used to replace the original model to complete the model version upgrade. If the performance of the model on the validation data set is stable, the original model is still used for subsequent predictions.

[0121] Corresponding to the above embodiment, an embodiment of the present invention provides a computer storage medium, including: at least one memory and at least one processor;

[0122] The memory is used to store one or more program instructions;

[0123] The processor is used to run one or more program instructions to execute an environmental air quality monitoring method based on multi-source data fusion.

[0124] Corresponding to the above-mentioned embodiments, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer storage medium contains one or more program instructions, and the one or more program instructions are used by a processor to execute an environmental air quality monitoring method based on multi-source data fusion.

[0125] The embodiments disclosed in the present invention provide a computer-readable storage medium, in which computer program instructions are stored. When the computer program instructions are executed on a computer, the computer executes the above-mentioned environmental air quality monitoring method based on multi-source data fusion.

[0126] In the embodiment of the present invention, the processor may be an integrated circuit chip having the ability to process signals. The processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0127] The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present invention can be directly embodied as a hardware decoding processor for execution, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The processor reads the information in the storage medium and completes the steps of the above method in combination with its hardware.

[0128] The storage medium may be a memory, which may be, for example, a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memory.

[0129] Among them, the non-volatile memory can be a read-only memory (ROM), a programmable ROM (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory.

[0130] The volatile memory may be a random access memory (RAM) which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDRSDRAM), enhanced synchronous DRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus DRAM (DRRAM).

[0131] The storage media described in the embodiments of the present invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0132] Those skilled in the art will appreciate that in one or more of the above examples, the functions described in the present invention can be implemented using a combination of hardware and software. When software is used, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any media that facilitates the transmission of computer programs from one place to another. Storage media can be any available media that can be accessed by general or special-purpose computers.

[0133] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for monitoring ambient air quality based on multi-source data fusion, characterized in that: include: Step S10: Pull traffic flow data, geographic information data, air quality data and meteorological data within the monitoring area to the monitoring system; Step S20: The monitoring system quantifies and fuses the acquired data to form a multi-source data set; Step S30: parse the multi-source data set using a spatiotemporal scale analysis algorithm to estimate the impact coefficient of various data on air quality, which is specifically divided into the following sub-steps: The influence coefficient of traffic flow data on air quality is estimated by fixing the time and space scale. The estimation formula is: ,in is the estimated value of the impact coefficient of traffic flow on air quality, , They are the traffic flow data and air quality data collected at the i+1th collection time contained in the fixed time scale, , is the traffic flow data and air quality data collected at the i-th collection time contained in the fixed time scale. i ranges from 1 to n-1, and n is the number of collection times contained in the fixed time scale. A formal parameter is set to avoid the denominator being zero, and the value is one hundred thousandth; Fix the spatial scale, adjust the time scale, and estimate the impact coefficient of meteorological data on air quality. The estimation formula is: ,in is the estimated value of the impact coefficient of meteorological data on air quality, , , They represent the meteorological data, air quality data, and traffic flow data at the j+1th time scale, respectively. , , They represent the meteorological data, air quality data, and traffic flow data at the jth time scale, respectively. j ranges from 1 to m-1, and m is the total length of the time scale. Fix the time scale, adjust the spatial scale, and estimate the impact coefficient of geographic information data on air quality. The estimation formula is: ,in is the estimated value of the impact coefficient of geographic information on air quality, , , , They represent the geographic information data, air quality data, traffic flow data and meteorological data at the k+1th spatial scale respectively. , , , They represent the geographic information data, air quality data, traffic flow data and meteorological data on the kth spatial scale respectively, k ranges from 1 to w-1, and w is the total length of the spatial scale; Step S40: Create an air quality prediction model based on the estimated values ​​of the influence coefficients of various data on air quality, which is specifically divided into the following sub-steps: Extract input-output pairs from multi-source data sets and organize them into training data sets; First, the multi-source data sets are grouped by the collection location, and the data in the group are sorted according to the collection time; then the geographic information data, traffic flow data, and meteorological data within the 1st to 10th hour in the group are extracted as input, and the air quality data from the 11th to 20th hour is used as output to form the first input-output pair, and then the geographic information data, traffic flow data, and meteorological data within the 2nd to 11th hour are used as input, and the air quality data from the 12th to 21st hour is used as output to form the second input-output pair, and so on, until all the input-output pairs contained in the group are extracted, and the extracted input-output pairs are stored in the training data set; Use the training data set to train the air quality prediction model and determine the growth coefficient of each type of data over time; The air quality prediction model is expressed as: , where Y is the prediction result, , , The input data are the traffic flow data, meteorological data, and geographic information data collected at the time t. , , are the growth coefficients of traffic flow data, meteorological data, and geographic information data over time, , , are the influence coefficients of traffic flow data, meteorological data, and geographic information data on air quality, respectively. t ranges from 1 to T, where T is the length of the prediction period; Deploy the trained air quality prediction model to the monitoring area; Step S50: Use the air quality prediction model to predict the change trend of air quality in the monitoring area and generate an air quality monitoring report; Step S60: Regularly tune the air quality prediction model using real data generated in the monitoring area.

2. The method for monitoring ambient air quality based on multi-source data fusion according to claim 1, characterized in that: The monitoring system quantifies and integrates the acquired data to form a multi-source data set, which is divided into the following sub-steps: Use the system's built-in vector generator to generate the pulled data into vectors; The generated vectors are input into the unified quantization function in sequence, and the quantization result of the pulled data is returned; The quantified data are aligned in time and space and stored as a multi-source dataset.

3. The method for monitoring ambient air quality based on multi-source data fusion according to claim 1, characterized in that: The air quality prediction model is regularly tuned using real data generated within the monitoring area, which is divided into the following sub-steps: Regularly collect real data generated within the monitoring area; Process the collected data into a validation dataset; Evaluate the performance changes of air quality prediction models on validation datasets; The air quality prediction model is tuned based on the performance evaluation results.

4. An ambient air quality monitoring system based on multi-source data fusion, used to execute an ambient air quality monitoring method based on multi-source data fusion as claimed in any one of claims 1 to 3, characterized in that: include: Multi-source data extraction module, used to extract traffic flow data, geographic information data, air quality data and meteorological data within the monitoring area; The multi-source data fusion module is used to quantify and fuse the pulled data to form a multi-source data set; Multi-source data analysis module, which is used to analyze multi-source data sets using spatiotemporal scale analysis algorithms and estimate the impact coefficient of various data on air quality; An air quality prediction model creation module is used to create an air quality prediction model based on the estimated values ​​of the impact coefficients of various data on air quality; An air quality monitoring report generation module is used to use an air quality prediction model to predict the changing trend of air quality in the monitoring area and generate an air quality monitoring report; The air quality prediction model tuning module is used to regularly tune the air quality prediction model using real data generated in the monitoring area.

5. The ambient air quality monitoring system based on multi-source data fusion according to claim 4 is characterized in that: The multi-source data fusion module includes a vector generation submodule, a unified quantization submodule, and a time-space synchronization submodule; The vector generation submodule is used to generate vectors from the pulled data using the built-in vector generator; The unified quantization submodule is used to input the generated vectors into the unified quantization function in sequence and return the quantization results of the pulled data; The spatiotemporal synchronization submodule is used to align the quantified data in time and space and store them as a multi-source data set.

6. The ambient air quality monitoring system based on multi-source data fusion according to claim 4 is characterized in that: The multi-source data analysis module includes a traffic flow data influence coefficient estimation submodule, a meteorological data influence coefficient estimation submodule, and a geographic information data influence coefficient estimation submodule; Traffic flow data impact coefficient estimation submodule is used to estimate the impact coefficient of traffic flow data on air quality at a fixed time and space scale; The meteorological data impact coefficient estimation submodule is used to fix the spatial scale, adjust the time scale, and estimate the impact coefficient of meteorological data on air quality; The geographic information data impact coefficient estimation submodule is used to fix the time scale, adjust the spatial scale, and estimate the impact coefficient of geographic information data on air quality.

7. The ambient air quality monitoring system based on multi-source data fusion according to claim 4 is characterized in that: The air quality prediction model creation module includes a training data set creation submodule, a model training submodule, and a model delivery submodule; The training data set creation submodule is used to extract input-output pairs from multi-source data sets and organize them into training data sets; The model training submodule is used to train the air quality prediction model using the training data set and determine the growth coefficient of various types of data over time; The model delivery submodule is used to deliver the trained air quality prediction model to the monitoring area.

Images