A method, system and product for drawing a high-temporal and spatial resolution population distribution map of a city

Abstract

The present application relates to a kind of urban high space-time resolution population distribution map drawing method, system and product, belong to population distribution drawing field, method includes: obtaining data set;Environment population data is as the spatial downscaling weight layer of working time;Build empty attribute fishing net;The sum of the corresponding environment population pixel value of each unit in fishing net is calculated;The weight of each pixel in environment population data is calculated;Each pixel of baidu heat map data of working time is spatially downscaling;Data correction is carried out to obtain real population value, recorded as first grid layer;Luojia 1-01 night light image is normalized;Determine new grid layer;According to the spatial downscaling process of working time, baidu heat map data of sleep time is spatially downscaling, and second grid layer is obtained;First grid layer and second grid layer are combined, and urban high space-time resolution population distribution map is obtained.The above scheme in the present application can realize the accurate drawing of urban population distribution map.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of population distribution mapping, and in particular to a method and system for mapping high spatio-temporal resolution population distribution in cities and products thereof. BACKGROUND

[0002] Rapid urbanization worldwide not only leads to an increase in impervious surfaces, but also is accompanied by a large influx of population into cities in search of more job opportunities and better living benefits. Therefore, the influx of population brings new challenges to urban planning and environmental issues, such as regional growth imbalance, disaster response, water resource shortage, severe traffic congestion and carbon-induced air pollution, especially in metropolises like Beijing which are in line with international standards.

[0003] In the past few decades, many methods have been developed, such as simple spatial interpolation, linear statistical models based on zonal density (Dasymetric) and machine learning models, to downscale population census data from a global scale to grid cells by using multi-source auxiliary data. It has been proven that Dasymetric Mapping is a very effective spatial downscaling method and is widely used to generate grid population density maps. The core idea is to generate a weight layer according to auxiliary data and use the weight layer to decompose coarse resolution variables (such as population) to finer resolution. Widely used auxiliary data are satellite-based remote sensing products, such as night-time light (NTL) images and land use / land cover (LULC). In recent years, emerging geospatial big data, such as points of interest (POI) and building volumes, have also been used as auxiliary data to decompose population census data, which provides new opportunities for generating more accurate grid population density maps. In addition, there are many high-quality and freely available global grid population density maps, such as the LandScan global population database (1 km) and the WorldPop global population product (100 m), which are also made by combining Dasymetric methods with multi-source auxiliary data. However, population is a dynamic variable in time, and the main shift in its distribution occurs in the daily cycle, resulting in rapid changes in population density. Therefore, the time resolution of these products is limited and cannot accurately reflect the dynamic distribution of the population. In addition, although the commonly used auxiliary data can successfully allocate population census data to space, the static nature of the input data set obscures the specific time to which the population distribution refers. Therefore, more dynamic data sources are needed to reflect the short-term and specific-time spatial redistribution of population caused by human mobility.

[0004] In recent years, the rapid development of mobile devices and the abundance of location-based services (LBS) big data enable researchers to analyze human flow patterns and map population spatial distribution with finer temporal resolution. Currently, Baidu heat map data and mobile phone signaling data are the most popular dynamic LBS data, which have accurate spatio-temporal information and can be used to study dynamic population distribution and improve the temporal resolution of grid population density maps. For these two types of data, mobile phone signaling data is the most promising spatio-temporal population data source because it has a very high penetration rate worldwide. According to relevant statistics, in developed countries, the number of mobile phone users has exceeded the total population, and the current penetration rate has reached 121%, while in developing countries, the penetration rate is as high as 90% and is still rising. However, under the current legal framework, due to privacy concerns and the lack of business models, operators are reluctant to release their data. Therefore, the acquisition of this dataset is still limited in geographical coverage, so it is still difficult to map short-term population maps in a large area. Baidu heat map data is widely used in population flow pattern research because it can be publicly accessed and has a high penetration rate, which makes it possible to map dynamic population distribution in a large area. However, the highest spatial resolution of Baidu heat map data available at the city scale is currently 500 meters, which is not sufficient to capture subtle changes in population density within the city. Therefore, it is crucial to develop a spatial downscaling method for Baidu heat map data to map high spatio-temporal resolution population distribution in a large scale. SUMMARY

[0005] The purpose of the present application is to provide a city high spatio-temporal resolution population distribution mapping method, system and product, which integrates environmental population data, night light data and building volume data, and innovatively proposes a spatial downscaling framework for Baidu heat map data during working hours and sleep time, and maps the population distribution of a city with high spatio-temporal resolution.

[0006] To achieve the above-mentioned purpose, the present application provides the following solutions:

[0007] In a first aspect, the present application provides a city high spatio-temporal resolution population distribution mapping method, which comprises:

[0008] obtaining a dataset; the dataset comprises geographic spatial big data, remote sensing data, population data, verification data and basic geographic data; the geographic spatial big data comprises Baidu heat map and building volume; the remote sensing data comprises Luo Ga 1-01 night light image and NPP-VIIRS night light image; the population data comprises environmental population data and population census data; the verification data comprises mobile phone signaling data;

[0009] The environmental population data is taken as a spatial downscaling weight layer of the working time;

[0010] The empty attribute fishing net is constructed;

[0011] The sum of the environmental population pixel values corresponding to each unit in the fishing net is calculated;

[0012] The weight of each pixel in the environmental population data is calculated based on the sum of the environmental population pixel values;

[0013] Each pixel of the Baidu heat map data of the working time is spatially down-scaled based on the weight of each pixel;

[0014] The Baidu heat map data of the working time after spatial downscaling is corrected by using the census data to obtain a real population value, which is recorded as a first grid layer;

[0015] The Lujia 1-01 night light image is normalized;

[0016] The corresponding pixel values of the building volume data and the normalized Lujia 1-01 NTL data are added to obtain a new grid layer;

[0017] The new grid layer is taken as a weight layer;

[0018] The Baidu heat map data of the sleep time is spatially down-scaled according to the spatial downscaling process of the working time to obtain a second grid layer;

[0019] The first grid layer and the second grid layer are combined to obtain a city high-temporal resolution population distribution map.

[0020] Optionally, the sum of the environmental population pixel values corresponding to each unit in the fishing net is calculated by using the following formula:

[0021]

[0022] wherein, P j is the pixel value of the environmental population data, S i is the sum of the 25 environmental population pixel values corresponding to the i-th unit in the fishing net.

[0023] Optionally, the weight of each pixel in the environmental population data is calculated based on the sum of the environmental population pixel values by using the following formula:

[0024]

[0025] wherein, P j is the pixel value of the environmental population data, S i is the sum of the 25 environmental population pixel values corresponding to the i-th unit in the fishing net, Wj is a weight value.

[0026] Optionally, the spatial downscaling of each pixel of the work time's heat map data based on the weight of each pixel is specifically implemented by the following formula:

[0027] H ij = H i × W j

[0028] wherein, H i is the i-th pixel value of the heat map data, H ij is the pixel value of the heat map data after spatial downscaling, W j is the weight.

[0029] Optionally, the work time's heat map data after spatial downscaling is corrected by using the population census data to obtain the real population value, which is specifically implemented by the following formula:

[0030]

[0031] wherein, H ij is the pixel value of the heat map data after spatial downscaling, S is the sum of all pixel values of the heat map data after spatial downscaling, H' ij is the real population value, and C represents the total number of urban permanent population.

[0032] Optionally, the normalization processing of the Luojia 1-01 night light image is specifically implemented by the following formula:

[0033]

[0034] wherein, RL i ' is the normalized value of the i-th pixel of the grid layer, RL i represents the original value of the i-th pixel of the grid layer, RL max represents the maximum value of the grid layer, and RL min is the minimum value of the grid layer.

[0035] Optionally, the corresponding pixel values of the building volume data and the normalized Luojia 1-01 NTL data are added to obtain a new grid layer, which is specifically implemented by the following formula:

[0036]

[0037] wherein, BV i is the i-th pixel value of the building volume data, NTL j is the i-th pixel value of the Luojia 1-01 NTL data, and N i is the i-th pixel value of the new grid layer.

[0038] In a second aspect, based on the method described above, the present application further provides a high spatio-temporal resolution population distribution mapping system for a city, comprising:

[0039] a data set acquisition module configured to acquire a data set, wherein the data set comprises geographic spatial big data, remote sensing data, population data, verification data, and basic geographic data, the geographic spatial big data comprises Baidu heat map data and building volume data, the remote sensing data comprises Lujia 1-01 night light image data and NPP-VIIRS night light image data, the population data comprises environmental population data and census data, and the verification data comprises mobile phone signaling data;

[0040] a first weight layer determination module configured to determine the environmental population data as a spatial downscaling weight layer for working hours;

[0041] a fishing net construction module configured to construct an empty attribute fishing net;

[0042] a sum of environmental population pixel value calculation module configured to calculate the sum of environmental population pixel values corresponding to each unit in the fishing net;

[0043] a weight calculation module configured to calculate the weight of each pixel in the environmental population data based on the sum of environmental population pixel values;

[0044] a first spatial downscaling module configured to perform spatial downscaling on each pixel of the Baidu heat map data for working hours based on the weight of each pixel;

[0045] a real population value calculation module configured to correct the spatially down-scaled Baidu heat map data for working hours by using the census data to obtain real population values, denoted as a first grid layer;

[0046] a normalization module configured to perform normalization processing on the Lujia 1-01 night light image data;

[0047] a new grid layer determination module configured to add the corresponding pixel values of the building volume data and the normalized Lujia 1-01 NTL data to obtain a new grid layer;

[0048] a second weight layer determination module configured to determine the new grid layer as a weight layer;

[0049] a second downscaling module configured to perform spatial downscaling on the Baidu heat map data for sleeping hours according to the spatial downscaling process for working hours to obtain a second grid layer;

[0050] The urban high-temporal and high-spatial resolution population distribution map determination module is configured to combine the first grid layer and the second grid layer to obtain an urban high-temporal and high-spatial resolution population distribution map.

[0051] In a third aspect, the present application provides an electronic device comprising a memory and a processor, wherein the memory is configured to store a computer program, and the processor is configured to execute the computer program to enable the electronic device to perform the urban high-temporal and high-spatial resolution population distribution map drawing method.

[0052] In a fourth aspect, the present application provides a computer readable storage medium storing a computer program, wherein the computer program is configured to be executed by a processor to implement the urban high-temporal and high-spatial resolution population distribution map drawing method.

[0053] According to the embodiments of the present application, the following technical effects are achieved.

[0054] The above-mentioned scheme in the present application integrates environmental population data, night light data and building volume data, and innovatively proposes a spatial downscaling framework for Baidu heat map data in working hours and sleeping hours, and draws a population distribution map of a city at a high-temporal and high-spatial resolution (i.e., every hour, 100 meters), thereby greatly improving the accuracy of the spatial downscaling framework in working hours and sleeping hours. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0056] Figure 1 The present application is an urban high-temporal and high-spatial resolution population distribution map drawing method flowchart.

[0057] Figure 2 The present application is a spatial downscaling framework for Baidu heat map data in working hours and sleeping hours.

[0058] Figure 3 The present application is a scatter plot of Baidu heat map population density and mobile signaling population density.

[0059] Figure 4 The present application is an urban high-temporal and high-spatial resolution population distribution map drawing system structure schematic diagram. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of the present application.

[0061] The purpose of the present application is to provide a city high-temporal and high-spatial resolution population distribution mapping method, system and product, which integrates environmental population data, night light data and building volume data, and innovatively proposes a spatial downscaling framework for Baidu heat map data in working hours and sleeping hours, and draws a population distribution map of a city at high-temporal and high-spatial resolution.

[0062] In order to make the above-mentioned purposes, characteristics and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0063] The present application proposes a spatial downscaling framework for Baidu heat map data to draw dynamic population distribution (see Figure 2 ). This spatial downscaling framework is divided into three parts: first, the related data sets are preprocessed, second, the weight layers are generated in working hours and sleeping hours respectively, and finally, the results are verified. The specific scheme is as follows:

[0064] Figure 1 The flowchart of the city high-temporal and high-spatial resolution population distribution mapping method of the present application is shown in Figure 2 , the spatial downscaling framework for Baidu heat map data in working hours and sleeping hours of the present application is shown in Figure 1 and Figure 2 , the method in the present application includes:

[0065] Step 1: Obtain a data set; the data set includes: geographic spatial big data, remote sensing data, population data, verification data and basic geographic data; the geographic spatial big data includes: Baidu heat map and building volume; the remote sensing data includes: Luo Ga 1-01 night light image and NPP-VIIRS night light image; the population data includes: environmental population data and population census data; the verification data includes: mobile phone signaling data.

[0066] The data used in the present application mainly includes geographic spatial big data, remote sensing data, population data, verification data and basic geographic data. Table 1 lists 8 types of data used in the present application. In order to ensure the consistency of spatial position, all the data in the present application are re-projected into WGS-1984-UTM-Zone-50N coordinate system.

[0067] Table 1. List of datasets and sources used in this study

[0068]

[0069]

[0070] Wherein, the introduction of Baidu heat map data is as follows:

[0071] In 2011, Baidu launched a big data visualization product (i.e., Baidu heat map). As a big data application with hundreds of millions of users, Baidu heat map is based on the location information of users accessing Baidu products (such as Baidu map, Baidu search, etc.), calculates the human flow heat value in different time and different areas, and displays it on Baidu map after density analysis processing. Therefore, Baidu heat map can greatly reflect the accurate crowd heat in the region and is widely used in the study of dynamic population distribution.

[0072] Baidu heat map data is derived from Baidu map (http: / / map.baidu.com (accessed on August 17, 2022)). Taking Beijing as an example, this study obtained Baidu heat map data of 24 time periods (0:00-1:00, etc.) in Beijing on August 17, 2022 (Wednesday) using Baidu's application programming interface, with a spatial resolution of 500 meters. The heat value of each vector point represents the total number of Baidu signal responses in the time period and the spatial range to which it belongs. First, a 500x500 meter cell size fishnet covering the entire Beijing was created in ArcGIS 10.6. Then, we assigned the heat value of each vector point to the corresponding fishnet. Finally, we generated 24 raster layers with a spatial resolution of 500 meters using the fishnet with heat value information, corresponding to the 24 time periods. Gender

[0073] The introduction of mobile phone signaling data is as follows:

[0074] Mobile phone signaling data is based on the interaction between mobile phone users and base stations to determine the spatial location of users at different times. Mobile phone signaling data is generated continuously as long as the mobile phone is turned on. Therefore, mobile phone signaling data is the most ideal data source for studying population flow patterns.

[0075] ​Take Beijing as an example, the present application obtains the mobile phone signaling data of four time periods (0:00-1:00, 9:00-10:00, 15:00-16:00 and 21:00-22:00) on August 17, 2022 in Beijing from China Mobile operator, China Unicom operator and China Telecom operator, with a spatial resolution of 200 meters and no privacy problem. However, since the operator only provides data of 5172 geographic points, the present application only uses these data to verify the spatial downscaling framework proposed by the present application. The signaling value of each vector point represents the total number of mobile phone signaling responses in the time period and spatial range it belongs to. First, an empty attribute fishnet with a 200x200 meter cell size covering the whole Beijing is created in ArcGIS10.6. Then, the signaling value of each vector point is assigned to the corresponding fishnet. Finally, four raster layers with a spatial resolution of 200 meters are generated using the fishnet with mobile phone signaling value information, corresponding to the four time periods respectively.

[0076] The introduction of remote sensing data is as follows:

[0077] The first edition product of the National Polar-Orbiting Partnership Visible Infrared Imaging Radiometer Suite (NPP-VIIRS) NTL 2018 September monthly composite data is derived from the Earth Observation Group (https: / / eogdata.mines.edu / products / vnl / (visited on July 3, 2022)), with a spatial resolution of 500 meters. Luojia 1-01 is a new generation of NTL remote sensing satellite launched on June 2, 2018. It is a sun-synchronous satellite with the ability to cover the Earth within 15 days. Luojia 1-01 is equipped with a more sensitive complementary metal oxide semiconductor sensor with 14-bit quantization function, making it superior to NPP-VIIRS

[55] . In this study, the Luojia 1-01 NTL data on September 6, 2018 is derived from Wuhan University (http: / / 59.175.109.173:8888 / (visited on June 17, 2022)), with a spatial resolution of 130 meters.

[0078] The positioning accuracy of the LoKa 1-01 NTL data we obtained was lower than its spatial resolution, and the image offset at some locations reached 1611 meters, which had a negative impact on fine population mapping. Therefore, through 12 pairs of geometric control points selected from the NTL data and Google Maps, the LoKa 1-01 NTL data was geometrically corrected using Google Maps. After geometric correction, 12 pairs of control points were randomly selected from the corrected NTL data and Google Maps for accuracy evaluation. Through evaluation, it was found that the average positioning error was 16.3 meters, which was less than the spatial resolution of the LoKa 1-01 NTL data. The pixel value (DN) in the original NTL data cannot effectively describe the brightness degree of the light; therefore, formula (1) is used for radiation calibration. Since the LoKa 1-01 NTL data has considerable background noise, which can be misleading. The NPP-VIIRS NTL data eliminates the pollution of cloud cover, moonlight and other factors; therefore, the DN value of 0 in the NPP-VIIRS NTL data is used in the present invention to mask the LoKa 1-01 NTL data to eliminate the relevant noise. Finally, the LoKa 1-01 NTL data is resampled to a spatial resolution of 100 meters using the nearest neighbor method to avoid changing any pixel value in the resampling process.

[0079]

[0080] where L is the radiance (W·m -2 ·sr -1 ·μm -1 ) of a pixel in the LoKa 1-01 image, and DN is the gray value of a pixel in the LoKa 1-01 image.

[0081] The introduction of the building volume data is as follows:

[0082] The building contour data comes from Baidu Map (http: / / map.baidu.com (visited on February 6, 2022)). First, an empty attribute fishnet with a unit size of 100x100 meters covering the entire Beijing was created in ArcGIS 10.6. Then, an intersection operation was performed between the fishnet and the building contour data. Since the building contour data has area and height information, the building volume of each unit can be calculated. Finally, the fishnet with building volume information was used to generate a raster layer with a spatial resolution of 100 meters.

[0083] The introduction of the environmental population data is as follows:

[0084] Each pixel value in the environmental population data represents the relative size of the probability of the presence of population during working hours [39, 40, 42]. The environmental population data is derived from the study of Bao et al. In this study, a population spatialization model named GXLS-Stacking was built by integrating gradient boosting decision trees (GBDT), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), and support vector regression (SVR) through the ensemble learning algorithm stacking. Then, we combined the socioeconomic data (such as point of interest data, building volume data, and artificial impervious surface data) and natural environmental data that enhance the spatial distribution of population characteristics with the census data to train the model and generate a high-precision grid population density map of Beijing in 2020 with a spatial resolution of 100 meters. The results show that the accuracy of our environmental population data is much higher than that of the WorldPop population dataset.

[0085] The introduction of the basic geographic data and census data is as follows:

[0086] Taking Beijing as an example, the administrative division map of Beijing comes from Tianditu (https: / / www.tianditu.gov.cn / (visited on August 6, 2022)). The census data comes from the Beijing Municipal Government, and the total number of permanent residents in Beijing in 2020 is 21,893,095. The census data is used to correct the Baidu heat map data and the mobile signaling data.

[0087] Steps 2-7 are an introduction to the working hours spatial scale framework:

[0088] Since the sleep time and working hours of different occupations are not consistent, the sleep time and working hours of the regular working day are set to 0:00-7:00 and 7:00-24:00, respectively. An efficient partition density (Dasymetric) method is used to spatially downscale the Baidu heat map data (500 meters) to a spatial resolution of 100 meters. Dasymetric mapping is an auxiliary data-driven method that has been widely used for spatial downscaling. The Dasymetric method introduces the density information of auxiliary variables to redistribute the standardized data to a finer scale distribution. The key step is to define the weight layer, which is usually determined by the existing or assumed relationship between the standardized data and the auxiliary variables. Since each pixel value in the environmental population data represents the relative size of the probability of the presence of population during working hours, the environmental population data is used as the spatial downscaling weight layer for working hours in this invention, and it is assumed that the spatial distribution of the Baidu heat map data (500 meters) at a finer scale (100 meters) is the same as that of the environmental population data (100 meters).

[0089] Step 2: Use the environmental population data as the spatial downscaling weight layer of the working time.

[0090] Step 3: Construct the empty attribute fishnet.

[0091] In this embodiment, Beijing is taken as an example, and an empty attribute fishnet with a unit size of 500*500 meters covering the whole Beijing is created in ArcGIS 10.6.

[0092] Step 4: Calculate the sum of the environmental population pixel values corresponding to each unit in the fishnet.

[0093] The calculation formula is as follows:

[0094]

[0095] wherein, P j is the pixel value of the environmental population data, S i is the sum of the 25 environmental population pixel values corresponding to the i-th unit in the fishnet.

[0096] Step 5: Calculate the weight of each pixel in the environmental population data based on the sum of the environmental population pixel values.

[0097] The calculation formula is as follows:

[0098]

[0099] wherein, P j is the pixel value of the environmental population data, S i is the sum of the 25 environmental population pixel values corresponding to the i-th unit in the fishnet, and W j is the weight value.

[0100] Step 6: Spatially downscale each pixel of the Baidu heat map data of the working time based on the weight of each pixel.

[0101] The calculation formula is as follows:

[0102] H ij = H i * W j (4)

[0103] wherein, H i is the i-th pixel value of the Baidu heat map data, H ij is the pixel value of the Baidu heat map data after spatial downscaling, and W j is the weight.

[0104] Step 7: Correct the Baidu heat map data of the working time after spatial downscaling by using the population census data to obtain the real population value, which is recorded as the first grid layer.

[0105] Since Baidu's heatmap data is sampled data, the extracted heat values ​​do not represent the actual population size, but only the relative size of the population density. It is assumed that Beijing's total population is constant throughout the day, and that daily population inflow and outflow are balanced. Therefore, census data is used to correct the spatially downscaled Baidu heatmap data so that each pixel value represents the actual population. The calculation formula is as follows:

[0106]

[0107] Among them, H ij H′ represents the pixel values ​​of the Baidu heatmap data after spatial downscaling, and S represents the sum of all pixel values ​​of the spatially downscaled Baidu heatmap data; ij C represents the actual population value, which is the total resident population of Beijing in the 2020 census, namely 21,893,095 people. It is used to correct the Baidu heat map data and mobile signaling data after spatial downscaling, so that the sum of all pixel values ​​in the raster layer of these two data sets is consistent with 21,893,095.

[0108] The following steps introduce the sleep temporal-spatial downscaling framework:

[0109] The spatial downscaling framework for sleep time is almost identical to that for work time, with the only difference being the definition of the weighting layers. Assumptions: (1) During sleep time, larger building volumes can accommodate more people; (2) In areas where building volume data is missing, NTL intensity reflects the residential distribution of the entire area (i.e., the brighter the NTL, the greater the population during sleep time). Therefore, building volume and NTL are used as the weighting layers in the spatial downscaling framework for sleep time.

[0110] Because building volume data is missing to varying degrees in both urban and rural areas, and NTL data can effectively reflect the population distribution of sleep time, we use NTL data to fill in these missing data gaps. However, the blooming effect is inherent to NTL, indicating that an NTL in a small area of ​​urban land will brighten the surrounding area. Therefore, this invention chooses to normalize the NTL data, keeping the relative size between pixel values ​​unchanged, and assigning these values ​​to the corresponding pixels in the building volume data. This both supplements the missing building volume information and avoids affecting the original building volume information due to excessively large pixel values. In this invention, we first normalize the NTL data of Luojia 1-01. Then, we add the corresponding pixel values ​​of the building volume data and the NTL data of Luojia 1-01 to obtain a new raster layer. Finally, we use the new raster layer as a weight layer to spatially downscale the Baidu heatmap data of sleep time according to the spatial downscaling process of working time. The specific steps are as follows:

[0111] Step 8: Normalizing the Lujia 1-01 night light image.

[0112] The specific formula is:

[0113]

[0114] Wherein, RL i is the normalized value of the i-th pixel of the grid layer, RL i represents the original value of the i-th pixel of the grid layer, RL max represents the maximum value of the grid layer, RL min is the minimum value of the grid layer.

[0115] Step 9: Adding the corresponding pixel values of the building volume data and the normalized Lujia 1-01 NTL data to obtain a new grid layer.

[0116] The specific formula is:

[0117]

[0118] Wherein, BV i is the i-th pixel value of the building volume data, NTL i is the i-th pixel value of the Lujia 1-01 NTL data, N i is the i-th pixel value of the new grid layer.

[0119] Step 10: Taking the new grid layer as a weight layer.

[0120] Step 11: Spatially downscaling the Baidu heat map data of sleep time according to the spatial downscaling process of working time to obtain a second grid layer.

[0121] Step 12: Combining the first grid layer and the second grid layer to obtain a city high-temporal and spatial resolution population distribution map.

[0122] The present application combines the 17 time period grid layers with a spatial resolution of 100 meters generated by the working time spatial downscaling framework and the 7 time period grid layers with a spatial resolution of 100 meters generated by the sleep time spatial downscaling framework to obtain 24 time period grid layers with a spatial resolution of 100 meters within a day, i.e. a high-temporal and spatial resolution population distribution map.

[0123] The above method proposed in the present application is verified as follows:

[0124] Accuracy evaluation is an important step to verify the accuracy of the spatial downscaling framework and is also an evaluation standard to judge the result. The present application adopts three widely used precision evaluation indexes, i.e. the determination coefficient (R 2), mean absolute error (MAE) and root mean square error (RMSE). The equations for calculating the above three indicators are as follows.

[0125]

[0126]

[0127]

[0128] wherein yi is the true value, is the predicted value, is the average value of the true value, and n is the total number of sample points.

[0129] A strict accuracy evaluation is performed on the generated high spatio-temporal resolution population distribution map, which can illustrate the reliability of the spatial downscaling framework proposed in the present application. First, an empty attribute fishnet with a 200*200 meter unit size covering the whole Beijing is created in ArcGIS 10.6. Then, the regional summation statistics of the population distribution map in four time periods (i.e. 0:00-1:00, 9:00-10:00, 15:00-16:00, 21:00-22:00) are performed using the 200-meter fishnet, and compared with the 5172 mobile signaling data corrected by the population census data, and finally the accuracy verification results are obtained. Figure 3 The accuracy evaluation results of the Baidu heat map population density in four time periods are shown, which correspond to one population distribution map generated by the spatial downscaling framework in the sleep time and three population distribution maps generated by the spatial downscaling framework in the working time, respectively. In general, the scatter points in the four time periods are distributed near the 1:1 line, which indicates that the population distribution maps in the four time periods have good accuracy, and also reflects the reliability of the spatial downscaling framework proposed in the present application for the Baidu heat map in the working time and the sleep time.

[0130] Based on the above method in the present application, the present application further provides a system for drawing a high spatio-temporal resolution population distribution map in a city, as shown in Figure 4 The system comprises:

[0131] A data set acquisition module 201 is configured to acquire a data set, wherein the data set comprises geographic spatial big data, remote sensing data, population data, verification data and basic geographic data; the geographic spatial big data comprises a Baidu heat map and a building volume; the remote sensing data comprises a Luo-Ga 1-01 night light image and an NPP-VIIRS night light image; the population data comprises environmental population data and population census data; and the verification data comprises mobile signaling data.

[0132] A first weight layer determination module 202 is configured to determine the environmental population data as a spatial downscaling weight layer in the working time.

[0133] a fishing net construction module 203, configured to construct an empty attribute fishing net;

[0134] a total sum of environment population pixel value calculation module 204, configured to calculate a total sum of environment population pixel values corresponding to each unit in the fishing net;

[0135] a weight calculation module 205, configured to calculate a weight of each pixel in the environment population data based on the total sum of the environment population pixel values;

[0136] a first spatial downscaling module 206, configured to perform spatial downscaling on each pixel of the Baidu heat map data of working time based on the weight of each pixel;

[0137] a real population value calculation module 207, configured to correct the spatially down-scaled Baidu heat map data of working time by using the census data to obtain a real population value, denoted as a first grid layer;

[0138] a normalization module 208, configured to perform normalization processing on the Lucheng 1-01 night light image;

[0139] a new grid layer determination module 209, configured to add corresponding pixel values of the building volume data and the normalized Lucheng 1-01 NTL data to obtain a new grid layer;

[0140] a second weight layer determination module 210, configured to use the new grid layer as a weight layer;

[0141] a second downscaling module 211, configured to perform spatial downscaling on the Baidu heat map data of sleeping time according to the spatial downscaling process of working time to obtain a second grid layer;

[0142] a city high spatio-temporal resolution population distribution map determination module 212, configured to combine the first grid layer and the second grid layer to obtain a city high spatio-temporal resolution population distribution map.

[0143] In addition, the present application also provides an electronic device comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to make the electronic device execute the city high spatio-temporal resolution population distribution map drawing method.

[0144] The present application also provides a computer readable storage medium storing a computer program, wherein the computer program is executed by a processor to implement the city high spatio-temporal resolution population distribution map drawing method.

[0145] The various embodiments described in this specification are presented for the purpose of illustrating the principles of the present application and its best mode of operation. Each of the embodiments described in this specification has been provided for the purpose of illustration only and the various embodiments are not intended to limit the present application in any way unless otherwise specifically indicated. The same parts and / or features of the various embodiments described in this specification can be referenced using the same reference numerals for the ease of understanding of the present application.

[0146] The principles and implementations of the present application have been described in the above embodiments, which are only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, the specific implementation and application range of the present application can be changed according to the idea of the present application. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for mapping urban high spatio-temporal resolution population distribution, characterized in that, The method comprises: acquiring a data set; the data set comprises: geospatial big data, remote sensing data, population data, verification data and basic geographic data; the geospatial big data comprises: Baidu heat map and building volume; the remote sensing data comprises: Luo Ga 1-01 night light image and NPP-VIIRS night light image; the population data comprises: environmental population data and census data; the verification data comprises: mobile phone signaling data; taking the environmental population data as a spatial downscaling weight layer of working hours; constructing an empty attribute fishing net; calculating the sum of environmental population pixel values corresponding to each unit in the fishing net; calculating the weight of each pixel in the environmental population data based on the sum of the environmental population pixel values; spatially downscaling each pixel of the Baidu heat map data of working hours based on the weight of each pixel; correcting the spatially downscaled Baidu heat map data of working hours by using the census data to obtain a real population value, denoted as a first grid layer; normalizing the Luo Ga 1-01 night light image; adding corresponding pixel values of the building volume data and the normalized Luo Ga 1-01 NTL data to obtain a new grid layer; taking the new grid layer as a weight layer; spatially downscaling the Baidu heat map data of sleeping hours according to the spatial downscaling process of working hours to obtain a second grid layer; combining the first grid layer and the second grid layer to obtain a city high-temporal-resolution population distribution map; The sum of environmental population pixel values corresponding to each unit in the fishing net is calculated by the following formula: wherein, is a pixel value of the environmental population data, is a sum of 25 environmental population pixel values corresponding to the th unit in the fishing net. The sum of corresponding pixel values of the building volume data and the normalized Luo Ga 1-01 NTL data is added to obtain a new grid layer by the following formula: in, The first of the building volume data pixel value, For Luojia 1-01 NTL data, the first pixel value, For the first raster layer Each pixel value.

2. The urban high spatio-temporal resolution population mapping method of claim 1, wherein, The weight of each pixel in the environmental population data is calculated based on the sum of the environmental population pixel values by the following formula: wherein, is a pixel value of the environmental population data, is a sum of 25 environmental population pixel values corresponding to the th unit in the fishing net, is a weight value.

3. The urban high spatio-temporal resolution population mapping method of claim 1, wherein, Each pixel of the Baidu heat map data of working hours is spatially down-scaled based on the weight of each pixel by the following formula: wherein, is the i-th pixel value of the baidu heat map data, is the i-th pixel value of the baidu heat map data, is the i-th pixel value of the baidu heat map data after spatial downscaling, is the weight. 4.The urban high-temporal-resolution population distribution mapping method according to claim 1, wherein, The real population value is obtained by correcting the spatially down-scaled Baidu heat map data of working hours by using the census data by the following formula: wherein, is the pixel value of the baidu heat map data after spatial downscaling, is the sum of all pixel values of the baidu heat map data after spatial downscaling, is the real population value, and C represents the total number of urban permanent population.

5. The urban high spatio-temporal resolution population mapping method of claim 1, wherein, The Luo Ga 1-01 night light image is normalized by the following formula: wherein, is a normalized value of the i-th pixel of the grid layer, is a normalized value of the i-th pixel of the grid layer, is an original value of the i-th pixel of the grid layer, is an original value of the i-th pixel of the grid layer, is a maximum value of the grid layer, is a minimum value of the grid layer.

6. A system for mapping urban high-temporal and high-spatial resolution population distribution, characterized by, The system comprises: a data set acquisition module configured to acquire a data set; the data set comprises: geospatial big data, remote sensing data, population data, verification data and basic geographic data; the geospatial big data comprises: Baidu heat map and building volume; the remote sensing data comprises: Luo Ga 1-01 night light image and NPP-VIIRS night light image; the population data comprises: environmental population data and census data; the verification data comprises: mobile phone signaling data; a first weight layer determination module configured to take the environmental population data as a spatial downscaling weight layer of working hours; a fishing net construction module configured to construct an empty attribute fishing net; an environmental population pixel value sum calculation module configured to calculate the sum of environmental population pixel values corresponding to each unit in the fishing net; an environmental population pixel value sum calculation module configured to calculate the sum of environmental population pixel values corresponding to each unit in the fishing net; a weight calculation module configured to calculate a weight of each pixel in the environment population data based on a sum of the environment population pixel values; a first spatial downscaling module configured to perform spatial downscaling on each pixel of the work time Baidu heat map data based on the weight of each pixel; a real population value calculation module configured to correct the spatially down-scaled work time Baidu heat map data using the census data to obtain a real population value, denoted as a first grid layer; a normalization module configured to perform normalization on the OLI 1-01 night light image; a new grid layer determination module configured to add corresponding pixel values of the building volume data and the normalized OLI 1-01 NTL data to obtain a new grid layer; a second weight layer determination module configured to use the new grid layer as a weight layer; a second downscaling module configured to perform spatial downscaling on the sleep time Baidu heat map data according to the spatial downscaling process of the work time to obtain a second grid layer; a city high spatio-temporal resolution population distribution map determination module configured to combine the first grid layer and the second grid layer to obtain a city high spatio-temporal resolution population distribution map; the sum of the environment population pixel values corresponding to each unit in the fishing net is calculated according to the following formula: wherein, is a pixel value of the environmental population data, is a sum of 25 environmental population pixel values corresponding to the th unit in the fishing net. the adding of the corresponding pixel values of the building volume data and the normalized OLI 1-01 NTL data to obtain the new grid layer is performed according to the following formula: in, The first of the building volume data pixel value, For Luojia 1-01 NTL data, the first pixel value, For the first raster layer Each pixel value.

7. An electronic device, comprising: The electronic device comprises a memory and a processor. The memory is configured to store a computer program. The processor is configured to execute the computer program to enable the electronic device to perform the city high spatio-temporal resolution population distribution map drawing method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer program is stored in the memory and is executed by the processor to implement the city high spatio-temporal resolution population distribution map drawing method according to any one of claims 1-5.

Images