Urban combined pollution exposure assessment method and early warning device based on multi-source big data

Abstract

The invention discloses an urban combined pollution exposure assessment method and early warning device based on multi-source big data, and the method comprises the steps: constructing a multi-source database which comprises high-precision satellite remote sensing data, ground monitoring station data and mobile phone signaling data; multi-source heterogeneous data fusion analysis including personnel density, traffic capacity, land coverage type, blue-green infrastructure area, building density / height and combined pollution concentration is realized, and on this basis, a multi-factor weighted assignment urban combined pollution exposure risk assessment method is provided. An executable independent operation program is formed, and a portable device which can be used for risk real-time early warning is researched and developed. According to the method, an interaction association mechanism between urban composite air pollution and multi-source influence factors such as personnel, traffic, land utilization, blue-green facilities and space structures is efficiently analyzed, so that accurate assessment and early warning of the exposure risk of the composite air pollution are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban pollution exposure assessment, and specifically relates to an urban composite pollution exposure assessment method and an early warning device based on multi-source big data. Background Art

[0002] Today, air pollution has become the world's largest environmental health risk to humans, causing about 9 million deaths each year, equivalent to one-sixth of the world's deaths. Air pollution is also considered to be closely related to many health risks and premature death. It has been proven that people exposed to PM2.5 are more likely to develop chronic cardiovascular diseases, respiratory diseases and lung cancer, while excessive O3 exposure may increase the probability of chronic obstructive pulmonary disease.

[0003] Since the beginning of the 21st century, China's industrial development and urbanization process have led to a large influx of people into cities. The rapid growth of population density and the rapid development of industrial enterprises have led to a sharp decline in the quality of urban ecological environment. The large-scale emission of automobile exhaust and industrial waste gas has made China's urban air pollution problem increasingly serious. More and more people are exposed to different types of air pollution (PM2.5, O3, NO2, etc.), the so-called urban complex air pollution exposure, which poses a greater threat to human health. Therefore, it is urgent to carry out urban air pollution (especially complex pollution) exposure assessment research to fully understand the pollution exposure level.

[0004] At present, urban air pollution exposure assessment mainly relies on census data and pollutant concentration data from monitoring stations, and obtains the distribution of urban air pollution exposure levels through mathematical modeling and calculation. However, the generation and diffusion process of urban air pollution is extremely complex and is often affected by many different types of factors, including urban spatial structure, land use and development, road traffic, and blue-green infrastructure. For example, high-density development and the use of land with intensive building functions can effectively reduce the ownership rate and travel distance of private cars, improve the efficiency of public transportation, and further reduce vehicle pollution emissions.

[0005] Urban air pollution is the result of the combined effects of urban spatial form and wind environment. High-density development and the presence of large buildings hinder air flow to a certain extent, which can easily lead to the formation of quiet wind areas and aggravate the accumulation of air pollution. Blue-green facilities such as urban green spaces, lakes and wetlands are widely considered to be an important means to reduce air pollution concentrations. Therefore, in order to efficiently realize urban air pollution exposure assessment, it is urgent to fully consider multi-dimensional factors such as personnel, transportation, land use, blue-green facilities, spatial structure, etc., obtain relevant data and deeply explore the complex relationship between various factors and urban air pollution, so as to establish a scientific and reliable exposure assessment method for real-time warning, etc.

[0006] However, urban air pollution exposure assessment still faces the following challenges:

[0007] 1) Data sources for various factors such as personnel, transportation, land use, blue-green facilities, and spatial structure are wide and varied. Existing studies obtain multi-source spatiotemporal data of various factors through satellite remote sensing, ground observation, and other means. The data are complex in format, redundant in information, and inefficient in fusion, making it difficult to achieve efficient data analysis. The complex relationship between various factors and urban pollution is often ignored, and the risk of pollution exposure is seriously underestimated.

[0008] 2) Urban complex pollution is complex in composition and often exhibits non-uniform dynamic propagation and diffusion characteristics, and there are significant interaction mechanisms between various pollutants. Existing research mainly conducts exposure assessment and risk warning for single-category pollutants, and has not yet formed reliable technology or equipment for exposure assessment and warning of complex pollutants, which may lead to a decrease in the efficiency of refined urban pollution control.

[0009] Therefore, in response to the above problems, it is urgent to propose an effective method to fully utilize the advanced technology of multi-source big data to systematically evaluate the level of urban complex air pollution exposure, and on this basis, develop a device that can be used for real-time early warning of pollution risks to provide scientific tools for government management departments to carry out urban complex pollution control. Summary of the invention

[0010] In response to the problems existing in the prior art, the present invention provides an urban complex pollution exposure assessment method and early warning device based on multi-source big data, which efficiently analyzes the interactive correlation mechanism between urban complex air pollution and multi-source influencing factors such as personnel, transportation, land use, blue-green facilities, and spatial structure, so as to achieve accurate assessment and early warning of complex category air pollution exposure risks.

[0011] In order to solve the above technical problems, the present invention is implemented through the following technical solutions:

[0012] A method for urban composite pollution exposure assessment based on multi-source big data, including:

[0013] Step 1) Select the target area and divide its city map into grids, and use the smallest grid unit obtained after grid division as the grid unit map for personnel density distribution calculation and pollution exposure assessment;

[0014] Step 2) obtaining personnel location (mobile phone signaling) data and population census data in the target area, and using them to calculate population density distribution data for each grid unit map in the target area;

[0015] Step 3) obtaining high-resolution land satellite remote sensing data of the target area, and using it to calculate the traffic capacity distribution data, land cover type distribution data, blue-green infrastructure area data and building density / height data of each unit grid map in the target area;

[0016] Step 4) obtaining ground monitoring station data of the target area, and combining it with high-resolution land satellite remote sensing data to calculate high-resolution distribution data of composite pollution concentration for each unit grid map in the target area;

[0017] Step 5) Based on the calculated population density distribution data, traffic capacity data, land cover type data, blue-green infrastructure area data, building density / height data, and high-resolution distribution data of composite pollution concentration, a multi-source database is constructed using a multi-source relational data fusion (MSF) framework, and multi-source heterogeneous data is deeply analyzed to obtain a dimensionless multi-source database;

[0018] Step 6) According to the dimensionless multi-source database, extract the urban characteristic factors of traffic capacity, land cover type, blue-green infrastructure area, building density / height, and then calculate the population density distribution data of various characteristic factors weighted by this factor, and then calculate the cumulative concentration of composite pollutants in combination with the high-resolution distribution data of dimensionless composite pollution concentration, and finally calculate the composite pollution exposure risk of the target area.

[0019] Furthermore, in step 1, the ArcGIS-based fishing net division technology is used to achieve grid division of the urban map of the target area, and the minimum grid unit size of the urban map of the target area after grid division is 1km×1km.

[0020] Furthermore, in step 2, the population density distribution data of each unit grid map in the target area is calculated as follows:

[0021] First, the personnel location (mobile phone signaling) data and population census data of the target area are used as influencing factors, and the distance from each grid unit map to the influencing factors is calculated. The distance data is classified according to the 1 / 4 standard deviation (determines the number of classifications) and quantile (determines the range of each classification);

[0022] Then, according to the importance of different influencing factors (importance ranking: personnel location data > census data), a corresponding weight is assigned to each classified distance data;

[0023] Finally, according to each distance data and its corresponding weight coefficient, the population density distribution data of each unit grid map in the target area is calculated.

[0024] Furthermore, in step 3, the calculation methods of the traffic capacity distribution data, land cover type distribution data, blue-green infrastructure area data and building density / height data of each unit grid map in the target area are as follows:

[0025] 1) The calculation method of traffic capacity data is:

[0026] According to the characteristics of vehicles in the image, texture features are extracted from the high-resolution land satellite remote sensing data (satellite images) of the target area, and the local binary pattern (LBP) is calculated to obtain the texture feature image. On this basis, the support vector machine is used to classify the vehicles, and the traffic capacity data of each unit grid map in the target area is calculated according to different vehicle types.

[0027] 2) The calculation method of land cover type distribution data is:

[0028] First, GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data are obtained through high-resolution land satellite remote sensing data (satellite images) in the target area; then the obtained GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data are preprocessed including radiation calibration, geometric correction, and projection conversion; then based on the multi-scale segmentation principle, the preprocessed GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data are image segmented to extract features including spectrum, texture, and shape; finally, the random forest algorithm is used for feature selection, and the CART algorithm is used for training and classification to obtain the land cover type distribution data of each unit grid map in the target area;

[0029] 3) The calculation method for blue and green infrastructure area data is:

[0030] Using the NDVI and MNDWI calculation formulas, the appropriate judgment threshold is determined according to the spectral curve characteristics and experiments, the blue-green infrastructure is extracted based on the high-resolution land satellite remote sensing data (satellite images) of the target area, and the blue-green infrastructure area data of each unit grid map in the target area is calculated;

[0031] 4) The calculation method for building density / height data is:

[0032] Through the imaging geometry model of buildings and shadows, a single high-resolution satellite remote sensing image in the high-resolution land satellite remote sensing data of the target area is used to calculate the building density / height data of each unit grid map in the target area.

[0033] Furthermore, in step 4, the calculation method of the high-resolution distribution data of the composite pollution concentration of each unit grid map in the target area is:

[0034] Through satellite remote sensing technology, the PM2.5, O3, and NO2 concentration fields in the target area are extracted. Based on the divided grid unit map, the diffusion distance of adjacent grid unit maps in the concentration field is calculated to obtain the diffusion distance surface. Then, the shortest path distance between each grid unit map and the ground monitoring station in the study area is calculated. The pollution monitoring data is then combined for inverse distance weighted interpolation to achieve high-resolution interpolation of different types of pollutant concentrations. Finally, based on the distribution results of different types of pollution concentrations, high-resolution distribution data of composite pollution concentrations are obtained.

[0035] Furthermore, in step 5, the method for constructing the dimensionless multi-source database is:

[0036] First, feature extraction is performed on attribute data including population density distribution data, traffic capacity data, land cover type data, blue-green infrastructure area data, building density / height data and high-resolution distribution data of composite pollution concentration to measure the similarity between attribute data, solve the similarity distance and form a distance matrix, and then find the best matching result based on the obtained distance matrix, so that the sum of the distances between all attribute data matching attributes is minimized, that is, a preliminarily screened multi-source database is obtained; then, based on the preliminarily screened multi-source database, a data cleaning method based on dynamically configurable rules and an extreme value normalization method are used to achieve in-depth analysis of multi-source heterogeneous data including multi-source data cleaning and dimensionless processing, thereby obtaining a dimensionless multi-source database; wherein,

[0037] 1) The specific steps of implementing multi-source data cleaning using a data cleaning method based on dynamically configurable rules are as follows:

[0038] Fill missing values, select the median of similar data as the filling value, and perform standard deviation test on noise data, further check data consistency, that is, check whether the data conforms to the predefined format, and finally implement redundant data detection based on similarity calculation;

[0039] 2) The range normalization method is used to realize dimensionless processing of multi-source data, that is, the original data is mapped to the interval [0,1] through linear transformation, so that data of different magnitudes can be compared and weighted;

[0040] Based on the above data analysis process, a dimensionless multi-source database is obtained, which includes dimensionless personnel density distribution data, dimensionless traffic capacity data, dimensionless land cover type data, dimensionless blue-green infrastructure area data, dimensionless building density / height data, and dimensionless composite pollution concentration high-resolution distribution data.

[0041] Furthermore, in step 6, the assessment method for the composite pollution exposure risk of the target area is:

[0042] Firstly, based on the dimensionless data corresponding to various factors such as traffic capacity, land cover type, blue-green infrastructure area, and building density / height in the dimensionless multi-source database, the entropy method is used to extract the urban characteristic factors of various factors; then the characteristic factors corresponding to various factors are used as weight coefficients and multiplied by the dimensionless population density distribution data to obtain the population density distribution data weighted by various characteristic factors; then, based on the obtained population density distribution data weighted by various characteristic factors, combined with the high-resolution distribution data of dimensionless composite pollution concentration, the cumulative composite pollutant concentration is calculated; finally, the relative mortality rate calculation method of epidemiology is used to calculate the composite pollution exposure risk of the target area in combination with the cumulative composite pollutant concentration.

[0043] An early warning device for urban composite pollution exposure risk comprises a map display screen, a text display screen and a switch arranged on the front of a shell, a hook arranged on the back of the shell, a wireless data receiver arranged on the side wall of the shell, and a computing unit, a power supply and a memory arranged inside the shell; wherein:

[0044] The wireless data receiver is responsible for sending personnel location data to the backend, and is responsible for receiving the composite pollution exposure risk corresponding to the personnel location calculated by the backend using the above-mentioned urban composite pollution exposure assessment method based on multi-source big data;

[0045] The operator is responsible for performing linear interpolation fitting on the composite pollution exposure risk values ​​corresponding to the received personnel's location in combination with the city grid coordinates, obtaining and outputting a high-resolution composite pollution exposure risk distribution map; and is responsible for comparing the composite pollution exposure risk at the personnel's location with a set risk threshold based on the high-resolution composite pollution exposure risk distribution map, obtaining and outputting a comparison result;

[0046] The map display screen is responsible for displaying the high-resolution composite pollution exposure risk distribution map output by the calculator;

[0047] The text display screen is responsible for displaying the comparison result between the composite pollution exposure risk of the location of the personnel output by the operator and the set risk threshold; when the composite pollution exposure risk of the location of the personnel is greater than the set risk threshold, the words "pollution seriously exceeds the standard" are displayed; when the composite pollution exposure risk of the location of the personnel is equal to the set risk threshold, the words "pollution has exceeded the standard" are displayed; when the composite pollution exposure risk of the location of the personnel is less than the set risk threshold, the words "pollution has not exceeded the standard" are displayed;

[0048] The memory is responsible for storing the set risk threshold, the acquired personnel location data, the received composite pollution exposure risk value corresponding to the personnel location, the fitted high-resolution composite pollution exposure risk distribution map, and the comparison result between the calculated composite pollution exposure risk at the personnel location and the set risk threshold;

[0049] The power supply is responsible for supplying power to the computing unit, the wireless data receiver, the map display screen, the text display screen and the memory;

[0050] The switch is responsible for turning on and off the power supply;

[0051] The hook is convenient for detachable connection with the clothing of the person.

[0052] A computer device includes: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus, and the memory is used to store at least one executable instruction, wherein the executable instruction enables the processor to execute operations corresponding to the above-mentioned urban composite pollution exposure assessment method based on multi-source big data.

[0053] A computer-readable storage medium stores at least one executable instruction, which enables a processor to execute operations corresponding to the above-mentioned urban complex pollution exposure assessment method based on multi-source big data.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. The present invention constructs a multi-source database that effectively integrates high-precision satellite remote sensing data, ground monitoring station data and mobile phone signaling data, and realizes efficient fusion and analysis of multi-source heterogeneous data including population density, traffic capacity, land cover type, blue-green infrastructure area, building density and height, and complex pollutant concentration, solving the problem of complex format and difficulty in lightweight analysis of massive data.

[0056] 2. The present invention combines personnel positioning (mobile phone signaling) data with census data to model and calculate the real-time distribution results of personnel density, and considers different time periods (working days and non-working days, morning rush hour, evening rush hour and other time periods) to differentiate the risks of exposure to compound pollution, thereby providing more accurate and reliable high-resolution personnel density distribution data for urban compound pollution exposure risk assessment.

[0057] 3. The present invention proposes a composite pollution exposure risk assessment method with weighted assignment of multiple factors (traffic capacity, land cover type, blue-green infrastructure area, and building density and height), extracts urban characteristic factors and calculates the cumulative composite pollution concentration, then introduces the epidemiological relative risk mortality calculation method, and constructs an assessment model to scientifically and effectively quantify the exposure risks of different types of pollution.

[0058] 4. The present invention has developed a portable device that can be used for real-time early warning of complex pollution exposure risks. It can provide dynamic early warning of pollution exposure risks at any location within the city. It not only provides an effective tool for urban residents to travel and take relevant protective measures, but also provides a valuable source of data for urban planners and decision makers, while providing important guidance for future urban development and public health policy formulation.

[0059] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. The specific implementation of the present invention is given in detail by the following embodiments and their accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0061] Figure 1 A schematic diagram of traffic capacity data, land cover type data, blue-green infrastructure area data, and building density distribution data of the study area in an embodiment of the present invention;

[0062] Figure 2 A schematic diagram of distribution information of ground monitoring stations in a research area in an embodiment of the present invention;

[0063] Figure 3 This is a schematic diagram of the population density distribution result after weighted processing according to the urban characteristic factors of the research area in an embodiment of the present invention;

[0064] Figure 4 This is a schematic diagram of the pollution concentration results obtained based on the cumulative composite pollution exposure assessment model in an embodiment of the present invention;

[0065] Figure 5 This is a schematic diagram of the quantitative assessment results of the combined pollution exposure risk obtained in an embodiment of the present invention;

[0066] Figure 6 It is a front schematic diagram of a portable device for real-time early warning of composite pollution exposure risk according to the present invention;

[0067] Figure 7 It is a schematic diagram of the back side of the portable device for real-time early warning of composite pollution exposure risk of the present invention;

[0068] Figure 8 It is a schematic diagram of the interior of the portable device for real-time early warning of combined pollution exposure risk of the present invention. DETAILED DESCRIPTION

[0069] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings so that the purpose, features and advantages of the invention can be more clearly understood. It should be understood that the embodiments shown in the accompanying drawings are not intended to limit the scope of the present invention, but are only intended to illustrate the essential spirit of the technical solution of the present invention.

[0070] In the following description, certain specific details are set forth for the purpose of illustrating various disclosed embodiments to provide a thorough understanding of the various disclosed embodiments. However, those skilled in the relevant art will recognize that the embodiments may be practiced without one or more of these specific details. In other cases, well-known devices, structures, and techniques associated with the present application may not be shown or described in detail to avoid unnecessarily obscuring the description of the embodiments.

[0071] Unless the context requires otherwise, throughout the specification and claims, the word "comprise" and variations such as "include" and "have" should be construed in an open, inclusive sense, ie, should be interpreted as "including, but not limited to."

[0072] References throughout the specification to "one embodiment" or "an embodiment" indicate that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of "in one embodiment" or "in an embodiment" in various places throughout the specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any manner in one or more embodiments.

[0073] As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. It should be noted that the term "or" is generally employed in its sense including "and / or" unless the context clearly dictates otherwise.

[0074] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0075] The present invention provides an urban composite pollution exposure assessment method based on multi-source big data, which mainly integrates a model for identifying the real-time distribution of personnel density based on mobile phone signaling data and census data, an algorithm for extracting traffic capacity, land cover type, blue-green infrastructure area, and building density / height based on high-resolution land satellite remote sensing data, a high-resolution interpolation method for composite pollution concentration combining high-precision satellite images with ground monitoring station data, and an urban composite pollution exposure risk assessment method with multi-factor weighted assignment, etc.

[0076] The present invention first establishes a model for identifying the real-time distribution of personnel density based on mobile phone signaling data and population census data. The model first grids the city map of the selected target area based on the fishing net division function of ArcGIS, and uses the minimum grid unit obtained after grid division as the grid unit map for personnel density distribution calculation and pollution exposure assessment, and the minimum grid unit size is 1km×1km. The model then uses the personnel positioning (mobile phone signaling) data and population census data of the target area as influencing factors, calculates the distance from each grid unit map to the influencing factors, and classifies the distance data according to 1 / 4 standard deviation (determines the number of classifications) and quantiles (determines the range of each classification), and then assigns corresponding weights to each classified distance data according to the importance of different influencing factors (importance ranking: personnel positioning data> population census data), and finally calculates the population density distribution data of each unit grid map in the target area according to each distance data and its corresponding weight coefficient. The calculation formula is as follows:

[0077]

[0078] In formula (1), P i represents the population density of the i-th city grid cell, ω j represents the weight of the jth influencing factor, f(dij) represents the distance function from the ith grid unit to the jth influencing factor, and n represents the total number of factors.

[0079] The present invention secondly constructs a traffic capacity, land cover type, blue-green infrastructure area, and building density / height extraction algorithm based on high-resolution land satellite remote sensing data. By acquiring high-resolution land satellite remote sensing data of the target area, the traffic capacity distribution data, land cover type distribution data, blue-green infrastructure area data, and building density / height data of each unit grid map in the target area are calculated respectively.

[0080] 1) The calculation method of traffic capacity data is:

[0081] According to the characteristics of vehicles in the image, texture features are extracted from the high-resolution land satellite remote sensing data (satellite images) of the target area, and the local binary pattern (LBP) is calculated to obtain the texture feature image. On this basis, the support vector machine is used to classify the vehicles, and the traffic capacity data of each unit grid map in the target area is calculated according to different vehicle types.

[0082] 2) The calculation method of land cover type distribution data is:

[0083] First, GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data are obtained through high-resolution land satellite remote sensing data (satellite images) in the target area; then the obtained GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data are preprocessed including radiation calibration, geometric correction, and projection conversion; then based on the multi-scale segmentation principle, the preprocessed GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data are image segmented to extract features including spectrum, texture, and shape; finally, the random forest algorithm is used for feature selection, and the CART algorithm is used for training and classification to obtain the land cover type distribution data of each unit grid map in the target area;

[0084] 3) The calculation method for blue and green infrastructure area data is:

[0085] Using the NDVI and MNDWI calculation formulas, the appropriate judgment threshold is determined according to the spectral curve characteristics and experiments, the blue-green infrastructure is extracted based on the high-resolution land satellite remote sensing data (satellite images) of the target area, and the blue-green infrastructure area data of each unit grid map in the target area is calculated;

[0086] The calculation formulas for NDVI and MNDWI are:

[0087]

[0088] In formula (2) and formula (3), NDVI is the normalized vegetation index, MNDWI is the improved normalized difference water index, NIR is the reflectance value of the near infrared band, R is the reflectance value of the red band, MIR is the reflectance value of the mid-infrared band, and Green is the reflectance value of the green band;

[0089] 4) The calculation method for building density / height data is:

[0090] Through the imaging geometry model of buildings and shadows, a single high-resolution satellite remote sensing image in the high-resolution land satellite remote sensing data of the target area is used to calculate the building density / height data of each unit grid map in the target area.

[0091] The present invention also proposes a high-resolution interpolation method for composite pollution concentrations that combines high-precision satellite images with ground monitoring station data. This method fully considers the impact mechanism of satellite images and ground monitoring station data on composite air pollution concentrations. Through satellite remote sensing technology, the PM2.5, O3, and NO2 concentration fields of the target area are extracted. Based on the divided grid unit map, the diffusion distance of adjacent grid unit maps in the concentration field is calculated to obtain the diffusion distance surface, and then the shortest path distance between each grid unit map and the ground monitoring station in the study area is calculated. Inverse distance weighted interpolation is then performed in combination with pollution monitoring data, thereby achieving high-resolution interpolation of different types of pollutant concentrations. Finally, high-resolution distribution data of composite pollution concentrations is obtained based on the distribution results of different types of pollution concentrations.

[0092] The formula for calculating the shortest path distance between the grid cell map and the ground monitoring station is:

[0093] W i = cos(D i -DM k )×V i -sgn[cos(D i -DM k )] (4);

[0094] In formula (4), W i represents the shortest path distance between the i-th grid cell and the ground monitoring station, D i represents the position of the i-th grid cell, DM k represents the location of the kth ground monitoring station, V i represents the pollution concentration of the i-th grid cell;

[0095] The calculation formula of the pollution concentration after interpolation is:

[0096]

[0097] In formula (5), Z R represents the pollution concentration of the Rth category after interpolation, VM k represents the pollution concentration at the kth ground monitoring station, and α is the distance attenuation parameter.

[0098] Based on the calculated population density distribution data, traffic capacity data, land cover type data, blue-green infrastructure area data, building density / height data and composite pollution concentration high-resolution distribution data, the present invention uses a general multi-source relational data fusion (MSF) framework to construct a multi-source database. First, feature extraction is performed on attribute data including population density distribution data, traffic capacity data, land cover type data, blue-green infrastructure area data, building density / height data and composite pollution concentration high-resolution distribution data to measure the similarity between each attribute data, solve the similarity distance and form a distance matrix, and then find the best matching result based on the obtained distance matrix, so that the sum of the distances between all attribute data matching attributes is minimized, that is, a preliminarily screened multi-source database is obtained.

[0099] Next, based on the multi-source database obtained through preliminary screening, a data cleaning method based on dynamically configurable rules and a range normalization method are used to achieve in-depth analysis of multi-source heterogeneous data, including multi-source data cleaning and dimensionless processing, thereby obtaining a dimensionless multi-source database.

[0100] The specific steps of implementing multi-source data cleaning using a data cleaning method based on dynamically configurable rules are as follows:

[0101] Fill missing values, select the median of similar data as the filling value, and perform standard deviation test on noise data, further check data consistency, that is, check whether the data conforms to the predefined format, and finally implement redundant data detection based on similarity calculation;

[0102] The formula for calculating standard deviation is:

[0103]

[0104] In formula (6), σ is the standard deviation, X i is the value of the data, μ is the mean;

[0105] The similarity calculation formula is:

[0106]

[0107] In formula (7), Sim is the similarity between two data records, X ik and X jk are the attribute values ​​of two k types of data;

[0108] The range normalization method is used to realize dimensionless processing of multi-source data, that is, the original data is mapped to the interval [0,1] through linear transformation, so that data of different magnitudes can be compared and weighted. The specific steps are as follows:

[0109] First, find the maximum and minimum values ​​in each category of data, and then perform dimensionless processing according to the following formula:

[0110]

[0111] In formula (8), X* ik represents the data processed by the range method, A k Represents the maximum value in each category of data, B k Represents the minimum value in each category of data.

[0112] Based on the above data analysis process, a dimensionless multi-source database is obtained, which includes dimensionless personnel density distribution data, dimensionless traffic capacity data, dimensionless land cover type data, dimensionless blue-green infrastructure area data, dimensionless building density / height data, and dimensionless composite pollution concentration high-resolution distribution data.

[0113] Based on the fusion and analysis of multi-source heterogeneous data, the present invention also proposes a multi-factor weighted assignment urban complex pollution exposure risk assessment method. This method extracts urban characteristic factors such as traffic capacity, land cover type, blue-green infrastructure area, and building density / height according to a dimensionless multi-source database, and then calculates the population density distribution data of weighted processing of various characteristic factors. Combined with the dimensionless high-resolution distribution data of complex pollution concentration, the cumulative complex pollutant concentration is calculated, and finally the complex pollution exposure risk of the target area is calculated based on this.

[0114] The assessment method for the combined pollution exposure risk in the target area is:

[0115] First, based on the dimensionless data corresponding to various factors such as traffic capacity, land cover type, blue-green infrastructure area, and building density / height in the dimensionless multi-source database, the entropy method is used to extract the urban characteristic factors of various factors. The calculation formula is as follows:

[0116]

[0117] g j =1-e j (11);

[0118]

[0119] In formula (9) to formula (13), P ij is the proportion of the i-th data in the j-th factor, Z ij is dimensionless data, n is the number of data samples, e j is the entropy value of the jth factor, g j is the coefficient of variation of the jth factor, ω jis the weight of the jth factor, m is the number of factors, and F is the city characteristic factor value; based on the above steps, the indicator weight is determined by calculating the entropy value of each factor, thereby extracting the city characteristic factor.

[0120] Then, the characteristic factors corresponding to various factors are used as weight coefficients and multiplied by the dimensionless population density distribution data to obtain the population density distribution data weighted by various characteristic factors. The calculation formula is as follows:

[0121] pop = TR × F (14);

[0122] In formula (14), pop represents the weighted population density distribution data, TR is the dimensionless population density data, and F is the city characteristic factor value.

[0123] Next, based on the obtained population density distribution data weighted by various characteristic factors and combined with the dimensionless high-resolution distribution data of composite pollution concentration, the cumulative composite pollutant concentration is calculated. The calculation formula is as follows:

[0124]

[0125] In equations (15) and (16), PWE represents the cumulative composite pollution concentration, PC mn represents the concentration of the mth type of pollutant in the nth grid cell, pop dn represents the weighted personnel density data of the nth grid cell on weekdays, pop wn Represents the weighted personnel density data of the nth grid on non-working days, D dm and D wm They represent the number of working days and non-working days with the concentration of the mth type of pollutant higher than the concentration threshold, pop c Represents the average weighted population density data, d d and d w represent the number of working days and non-working days respectively, and d represents the total number of days.

[0126] Finally, the relative mortality rate calculation method of epidemiology is used to calculate the composite pollution exposure risk of the target area in combination with the cumulative composite pollutant concentration. The calculation formula is as follows:

[0127] RR = β × PWE (17);

[0128]

[0129] In formula (17) and formula (18), RR represents the relative risk rate of death, and β represents the relative mortality rate corresponding to the combined pollution, which is 1.02×10 -1 , Mort represents the exposure risk, y0 represents the baseline mortality rate, and the all-cause mortality rate is 6.14×10-3 , which can be used to assess the comprehensive health impact of air pollution on diseases including cardiovascular, lung cancer, chronic bronchitis, acute bronchitis and asthma attacks.

[0130] See also Figure 6-Figure 8 As shown, according to the urban composite pollution exposure risk assessment method, the coding conversion is performed to form a program that can execute independent operations. The present invention also proposes an early warning device for urban composite pollution exposure risk. The device includes a map display screen 1, a text display screen 2 and a switch 3 arranged on the front of the shell, a hook 5 arranged on the back of the shell, a wireless data receiver 4 arranged on the side wall of the shell, and an operator 6, a power supply 7 and a memory 8 arranged inside the shell. Among them,

[0131] The wireless data receiver 4 is responsible for sending personnel location (mobile phone signaling) data to the background, and is responsible for receiving the composite pollution exposure risk corresponding to the personnel location calculated by the background using the above-mentioned urban composite pollution exposure assessment method based on multi-source big data.

[0132] The operator 6 is responsible for combining the city grid coordinates, performing linear interpolation fitting on the received composite pollution exposure risk values ​​corresponding to the personnel's location, obtaining a high-resolution composite pollution exposure risk distribution map and outputting it; and is responsible for comparing the composite pollution exposure risk at the personnel's location with the set risk threshold based on the high-resolution composite pollution exposure risk distribution map, obtaining a comparison result and outputting it.

[0133] The map display screen 1 is responsible for displaying the high-resolution composite pollution exposure risk distribution map output by the operator 6 .

[0134] The text display screen 2 is responsible for displaying the comparison result between the composite pollution exposure risk of the personnel's location output by the operator 6 and the set risk threshold.

[0135] The memory 8 is responsible for storing the set risk threshold, the acquired personnel location (mobile phone signaling) data, the received composite pollution exposure risk value corresponding to the personnel's location, the fitted high-resolution composite pollution exposure risk distribution map, and the comparison result between the calculated composite pollution exposure risk at the personnel's location and the set risk threshold.

[0136] The power supply 7 is responsible for supplying power to the computing unit 6 , the wireless data receiver 4 , the map display screen 1 , the text display screen 2 and the memory 8 .

[0137] The switch 3 is responsible for switching the power source 7 on and off.

[0138] The hook 5 is convenient for detachable connection with the clothing of the person.

[0139] The working method of the early warning device for urban composite pollution exposure risk of the present invention is as follows:

[0140] According to the personnel positioning (mobile phone signaling) data, the composite pollution exposure risk value corresponding to the personnel's location is input into the operator, and linear interpolation fitting is performed on the urban grid coordinates to obtain a high-resolution composite pollution exposure risk distribution map, which is transmitted to the display device for real-time display.

[0141] According to different time periods (morning peak, evening peak, other time periods) on working days and non-working days, corresponding composite pollution risk thresholds are set and input into the memory.

[0142] Based on the high-resolution composite pollution exposure risk distribution map, the composite pollution exposure risk at the personnel's location is compared with the set risk threshold in the calculator, and the comparison results are saved.

[0143] The comparison results are input into the display device. When the composite pollution exposure risk at the person's location is greater than the set risk threshold, "pollution seriously exceeds the standard" is displayed; when the composite pollution exposure risk at the person's location is equal to the set risk threshold, "pollution exceeds the standard" is displayed; when the composite pollution exposure risk at the person's location is less than the set risk threshold, "pollution does not exceed the standard" is displayed.

[0144] The following is a specific example of the urban composite pollution exposure assessment and early warning based on Nanjing City to specifically illustrate the urban composite pollution exposure assessment method and early warning device based on multi-source big data of the present invention, but the implementation mode of the present invention is not limited to this.

[0145] The types of composite pollutants involved in this embodiment include PM2.5, O3, NO2 (unit: μg / m 3 ), conduct pollution exposure risk assessment and dynamic early warning at an hourly frequency, from May 27, 2021 to May 31, 2021. The specific implementation process is as follows.

[0146] 1. The central area of ​​Nanjing and some surrounding areas were selected as the study area. The area was gridded using the ArcGIS fishnet function to achieve a minimum grid unit size of 1km×1km, which was used as a grid unit map for population density distribution calculation and pollution exposure assessment.

[0147] 2. Use personnel positioning (mobile phone signaling) data and population census data as influencing factors to calculate the distance from each grid unit to the influencing factors, classify the distance data according to the 1 / 4 standard deviation and quantile, and then assign weights to each classification based on the importance of different influencing factors (importance ranking: personnel positioning data > population census data), and calculate the population density distribution data based on the distance data and the weight coefficient.

[0148] 3. Based on high-resolution land satellite remote sensing data, the traffic capacity distribution data in the study area are obtained through the local binary pattern algorithm and support vector machine. The land cover type distribution data are extracted using the random forest and CART algorithms. The area distribution data of blue and green infrastructure are extracted by combining the NDVI and MNDWI calculation formulas. The building density distribution data is calculated based on the geometric model of building shadows and imaging, such as Figure 1 shown.

[0149] 4. Extract the PM2.5, O3, and NO2 concentration fields in the study area through satellite remote sensing technology, calculate the diffusion distances of adjacent grid cells in the concentration field based on the grid map, and further calculate the distances between the grid cells and the ground monitoring stations in the study area (such as Figure 2 The shortest path distance between two objects (as shown) is calculated by combining distance data with pollution monitoring data to achieve high-resolution interpolation of different types of pollution concentrations.

[0150] 5. Based on the above-mentioned high-resolution data such as population density distribution, traffic capacity, land cover type, blue-green infrastructure area data, building density, pollutant concentration, etc., a multi-source database is constructed, and through the data cleaning method with dynamic configurable rules and the range normalization method, in-depth analysis of multi-source heterogeneous data (including data cleaning and dimensionless processing) is achieved to obtain a dimensionless multi-source database.

[0151] 6. Based on the dimensionless multi-source database, the entropy method is used to extract the characteristic factors of influencing factors such as transportation, land, blue-green infrastructure and building density (0.12, 0.17, 0.55 and 0.16 respectively). The characteristic factors corresponding to each influencing factor are used as weight coefficients and multiplied by the dimensionless population density distribution data to obtain the weighted population density distribution data, such as Figure 3 As shown, Figure 3 This is the result of the weighted population density distribution (from May 27, 2021 to May 31, 2021, at 10:00 every day).

[0152] 7. Based on the weighted personnel density distribution data and the dimensionless composite pollution concentration high-resolution distribution data, the cumulative composite pollution concentration is calculated, such as Figure 4 shown.

[0153] 8. Based on the relative mortality calculation method, the cumulative composite pollution concentration is combined to obtain the composite pollution exposure risk in the study area, such as Figure 5 As shown, Figure 5 Quantify the assessment results for exposure risk.

[0154] 9. Input the composite pollution exposure risk assessment results into Figure 6In the computing unit of the portable early warning device shown, a high-resolution composite pollution exposure risk map is obtained (displayed by a screen display device), and the corresponding composite pollution exposure risk is matched according to the location of the personnel and compared with the set risk threshold. When it is greater than the set risk threshold, it will display "pollution seriously exceeds the standard"; when it is equal to the set risk threshold, it will display "pollution has exceeded the standard"; when it is less than the set risk threshold, it will display "pollution has not exceeded the standard".

[0155] The present invention also provides a computer device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other through the communication bus, and the memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute operations corresponding to the above-mentioned urban complex pollution exposure assessment method based on multi-source big data.

[0156] The present invention also provides a computer-readable storage medium, in which at least one executable instruction is stored, and the executable instruction enables a processor to execute operations corresponding to the above-mentioned urban complex pollution exposure assessment method based on multi-source big data.

[0157] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for urban composite pollution exposure assessment based on multi-source big data, characterized in that: include: Step 1) Select the target area and divide its city map into grids, and use the smallest grid unit obtained after grid division as the grid unit map for personnel density distribution calculation and pollution exposure assessment; Step 2) obtaining personnel location data and population census data in the target area, and using them to calculate population density distribution data for each grid unit map in the target area; Step 3) obtaining high-resolution land satellite remote sensing data of the target area, and using it to calculate the traffic capacity distribution data, land cover type distribution data, blue-green infrastructure area data and building density / height data of each unit grid map in the target area; Step 4) obtaining ground monitoring station data of the target area, and combining it with high-resolution land satellite remote sensing data to calculate high-resolution distribution data of composite pollution concentration for each unit grid map in the target area; Step 5) Based on the calculated population density distribution data, traffic capacity data, land cover type data, blue-green infrastructure area data, building density / height data, and high-resolution distribution data of composite pollution concentration, a multi-source database is constructed using a multi-source relational data fusion framework, and multi-source heterogeneous data is deeply analyzed to obtain a dimensionless multi-source database; Step 6) According to the dimensionless multi-source database, extract the urban characteristic factors of traffic capacity, land cover type, blue-green infrastructure area, building density / height, and then calculate the population density distribution data of various characteristic factors weighted by this factor, and then calculate the cumulative concentration of composite pollutants by combining the dimensionless high-resolution distribution data of composite pollution concentration, and finally calculate the composite pollution exposure risk of the target area.

2. The urban composite pollution exposure assessment method based on multi-source big data according to claim 1 is characterized in that: In step 1, the ArcGIS-based fishing net division technology is used to realize the grid division of the urban map of the target area, and the minimum grid unit size of the urban map of the target area after grid division is 1km×1km.

3. The urban composite pollution exposure assessment method based on multi-source big data according to claim 1 is characterized in that: In step 2, the population density distribution data of each unit grid map in the target area is calculated as follows: First, the personnel location data and population census data of the target area are used as influencing factors, the distance from each grid unit map to the influencing factors is calculated, and the distance data is classified according to the 1 / 4 standard deviation and quantile; Then, according to the importance of different influencing factors, a corresponding weight is assigned to each classified distance data; Finally, according to each distance data and its corresponding weight coefficient, the population density distribution data of each unit grid map in the target area is calculated.

4. The urban composite pollution exposure assessment method based on multi-source big data according to claim 1 is characterized in that: In step 3, the calculation methods of traffic capacity distribution data, land cover type distribution data, blue-green infrastructure area data and building density / height data of each unit grid map in the target area are as follows: 1) The calculation method of traffic capacity data is: According to the characteristics of vehicles in the image, texture features are extracted from the high-resolution land satellite remote sensing data of the target area, and the local binary pattern is calculated to obtain the texture feature image. On this basis, the support vector machine is used to classify the vehicles, and the traffic capacity data of each unit grid map in the target area is calculated according to different vehicle types. 2) The calculation method of land cover type distribution data is: First, GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data were obtained through high-resolution land satellite remote sensing data of the target area; then, the obtained GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data were preprocessed including radiation calibration, geometric correction, and projection conversion; then, based on the multi-scale segmentation principle, the preprocessed GF-1 wide-width multispectral data, MODIS data, and national forest resources continuous inventory fixed plot data were image segmented to extract features including spectrum, texture, and shape; finally, the random forest algorithm was used for feature selection, and the CART algorithm was used for training and classification to obtain the land cover type distribution data of each unit grid map in the target area; 3) The calculation method for blue and green infrastructure area data is: Using the NDVI and MNDWI calculation formulas, the appropriate judgment threshold is determined according to the spectral curve characteristics and experiments, the blue-green infrastructure is extracted based on the high-resolution land satellite remote sensing data of the target area, and the blue-green infrastructure area data of each unit grid map in the target area is calculated; 4) The calculation method for building density / height data is: Through the imaging geometry model of buildings and shadows, a single high-resolution satellite remote sensing image in the high-resolution land satellite remote sensing data of the target area is used to calculate the building density / height data of each unit grid map in the target area.

5. The urban composite pollution exposure assessment method based on multi-source big data according to claim 1 is characterized in that: In step 4, the calculation method of the high-resolution distribution data of the composite pollution concentration of each unit grid map in the target area is: Through satellite remote sensing technology, the PM2.5, O3, and NO2 concentration fields in the target area are extracted. Based on the divided grid unit map, the diffusion distance of adjacent grid unit maps in the concentration field is calculated to obtain the diffusion distance surface. Then, the shortest path distance between each grid unit map and the ground monitoring station in the study area is calculated. The pollution monitoring data is then combined for inverse distance weighted interpolation to achieve high-resolution interpolation of different types of pollutant concentrations. Finally, based on the distribution results of different types of pollution concentrations, high-resolution distribution data of composite pollution concentrations are obtained.

6. The urban composite pollution exposure assessment method based on multi-source big data according to claim 1 is characterized in that: In step 5, the method for constructing the dimensionless multi-source database is: First, feature extraction is performed on attribute data including population density distribution data, traffic capacity data, land cover type data, blue-green infrastructure area data, building density / height data, and high-resolution distribution data of composite pollution concentration to measure the similarity between attribute data, solve the similarity distance and form a distance matrix, and then find the best matching result based on the obtained distance matrix, so that the sum of the distances between matching attributes of all attribute data is minimized, that is, a preliminarily screened multi-source database is obtained; Then, based on the multi-source database obtained through preliminary screening, a data cleaning method based on dynamically configurable rules and a range normalization method are used to achieve in-depth analysis of multi-source heterogeneous data, including multi-source data cleaning and dimensionless processing, thereby obtaining a dimensionless multi-source database; 1) The specific steps of implementing multi-source data cleaning using a data cleaning method based on dynamically configurable rules are as follows: Fill missing values, select the median of similar data as the filling value, and perform standard deviation test on noise data, further check data consistency, that is, check whether the data conforms to the predefined format, and finally implement redundant data detection based on similarity calculation; 2) The range normalization method is used to realize dimensionless processing of multi-source data, that is, the original data is mapped to the interval [0,1] through linear transformation, so that data of different magnitudes can be compared and weighted; Based on the above data analysis process, a dimensionless multi-source database is obtained, which includes dimensionless personnel density distribution data, dimensionless traffic capacity data, dimensionless land cover type data, dimensionless blue-green infrastructure area data, dimensionless building density / height data, and dimensionless composite pollution concentration high-resolution distribution data.

7. The urban composite pollution exposure assessment method based on multi-source big data according to claim 1 is characterized in that: In step 6, the assessment method for the combined pollution exposure risk of the target area is: Firstly, based on the dimensionless data corresponding to traffic capacity, land cover type, blue-green infrastructure area, and building density / height in the dimensionless multi-source database, the entropy method is used to extract the urban characteristic factors of each factor. Then, the characteristic factors corresponding to various factors are used as weight coefficients and multiplied by the dimensionless population density distribution data, so as to obtain the population density distribution data weighted by various characteristic factors; Then, based on the obtained population density distribution data weighted by various characteristic factors and combined with the dimensionless high-resolution distribution data of composite pollution concentration, the cumulative composite pollutant concentration is calculated; Finally, the relative mortality rate calculation method of epidemiology is used, combined with the cumulative concentration of complex pollutants, to calculate the complex pollution exposure risk of the target area.

8. An early warning device for urban composite pollution exposure risk, characterized in that: The device comprises a map display screen (1), a text display screen (2) and a switch (3) arranged on the front of the housing, a hook (5) arranged on the back of the housing, a wireless data receiver (4) arranged on the side wall of the housing, and a computing unit (6), a power supply (7) and a memory (8) arranged inside the housing; wherein: The wireless data receiver (4) is responsible for sending personnel location data to the background, and is responsible for receiving the composite pollution exposure risk corresponding to the personnel location calculated by the background using the urban composite pollution exposure assessment method based on multi-source big data as described in any one of claims 1 to 7; The operator (6) is responsible for combining the city grid coordinates, performing linear interpolation fitting on the received composite pollution exposure risk values ​​corresponding to the location of the personnel, obtaining a high-resolution composite pollution exposure risk distribution map and outputting it; and is responsible for comparing the composite pollution exposure risk at the location of the personnel with a set risk threshold based on the high-resolution composite pollution exposure risk distribution map, obtaining a comparison result and outputting it; The map display screen (1) is responsible for displaying the high-resolution composite pollution exposure risk distribution map output by the computing unit (6); The text display screen (2) is responsible for displaying the comparison result between the composite pollution exposure risk at the location of the personnel output by the computing unit (6) and the set risk threshold; when the composite pollution exposure risk at the location of the personnel is greater than the set risk threshold, the words "pollution seriously exceeds the standard" are displayed; when the composite pollution exposure risk at the location of the personnel is equal to the set risk threshold, the words "pollution has exceeded the standard" are displayed; when the composite pollution exposure risk at the location of the personnel is less than the set risk threshold, the words "pollution has not exceeded the standard" are displayed; The memory (8) is responsible for storing the set risk threshold, the acquired personnel location data, the received composite pollution exposure risk value corresponding to the personnel location, the fitted high-resolution composite pollution exposure risk distribution map, and the comparison result between the calculated composite pollution exposure risk at the personnel location and the set risk threshold; The power supply (7) is responsible for supplying power to the computing unit (6), the wireless data receiver (4), the map display screen (1), the text display screen (2) and the memory (8); The switch (3) is responsible for turning on and off the power supply (7); The hook (5) is convenient for detachable connection with a person's clothing.

9. A computer device, characterized in that: include: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other through the communication bus, and the memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the urban complex pollution exposure assessment method based on multi-source big data as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer storage medium stores at least one executable instruction, which enables the processor to perform operations corresponding to the urban complex pollution exposure assessment method based on multi-source big data as described in any one of claims 1 to 7.

Images