A method and system for modeling the distribution of artificial light at night at an urban scale

Abstract

This invention belongs to the field of image generation technology and discloses a method and system for modeling the distribution of artificial light at night at an urban scale. The invention acquires nighttime light remote sensing image data, land use type vector data, ground elevation data, and building morphology vector data of the study area; it uses the land use type vector data to perform super-resolution enhancement processing on the nighttime light remote sensing image data; based on the building morphology vector data and ground elevation data, it generates fixed-resolution three-dimensional point cloud data within the study area; based on the above data, it obtains horizontal and vertical radiance values; and it assigns the radiance values ​​to the three-dimensional point cloud to visualize the distribution of artificial light at night at an urban scale. This invention enables three-dimensional modeling of nighttime light field distribution at an urban scale.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of image generation, and more particularly relates to a city-scale three-dimensional night artificial light distribution modeling method and system. BACKGROUND

[0002] In modern society, night artificial light has become an indispensable part of urban landscape, which not only reflects the intensity and distribution of human activities, but also has a profound impact on ecosystems, human health, and urban safety. With the continuous advancement of urbanization, the distribution pattern and intensity of night artificial light are changing rapidly in both horizontal and vertical directions of the city, which poses new challenges and demands for urban planning, energy management, and environmental protection. In order to solve these problems, it is necessary to propose an effective method to estimate the distribution of night light at the city scale in three dimensions, which has important practical significance for improving the scientificity and effectiveness of urban lighting management in real urban space and promoting sustainable development.

[0003] Currently, most of the modeling research on urban night artificial light uses two-dimensional night light remote sensing images (such as DMSP / OLS, VIIRS, and SDGSAT-1) as data sources, and further analyzes the radiation characteristics of different urban functional areas. In addition, some studies use super-resolution enhancement technology on night light remote sensing images to generate multispectral data with higher spatial accuracy, thereby improving the spatial resolution of night light modeling. In recent years, some studies have begun to explore the estimation of night artificial light distribution in three-dimensional space. Self-report questionnaires have been used in various studies, but this method is limited by subjective expression, lacks accurate description of light intensity distribution, and is difficult to quantitatively express the research results. In addition, some studies use machine learning technology to train collected night street image samples to predict night light in urban areas, but there are inevitable problems such as large training samples and difficult data collection. Another more direct method is to use an illuminometer to measure light intensity, but this measurement is labor-intensive and still limited to individual building models, lacking universal applicability.

[0004] In summary of previous studies, there is currently no suitable method to model the three-dimensional distribution of night light field at the city scale. Traditional night light plane modeling methods can only reflect the horizontal distribution of night light, and cannot effectively capture the changes in building facade light radiation, thereby hindering people's understanding of the distribution of night artificial light in real three-dimensional scenes; and the research on the vertical direction of night light radiation distribution is limited by the difficulty of data collection and field measurement, and is only limited to a few blocks or single buildings, which cannot model the three-dimensional night light distribution at the city scale as a whole, thereby limiting the scope of research. SUMMARY

[0005] The present application provides a kind of urban scale three-dimensional night artificial light distribution modeling method and system, solve the problem that prior art cannot carry out three-dimensional modeling to urban scale night light field distribution.

[0006] The present application provides a kind of urban scale three-dimensional night artificial light distribution modeling method, comprising the following steps:

[0007] Obtain the night light remote sensing image data, land use type vector data, ground elevation data and building shape vector data of research area;

[0008] The night light remote sensing image data is carried out super-resolution enhancement processing using the land use type vector data, and super-resolution enhancement processing data is obtained;

[0009] According to the building shape vector data and the ground elevation data, fixed resolution three-dimensional point cloud data is generated in research area;

[0010] Based on the horizontal point cloud data in the three-dimensional point cloud data, the super-resolution enhancement processing data and the building shape vector data, horizontal night light radiation is modeled, and horizontal radiation value is obtained;

[0011] Based on the facade point cloud data in the three-dimensional point cloud data and the horizontal radiation value, the change of illumination intensity with height is simulated using fitting skewness function, and vertical dimension night light radiation is modeled, and vertical radiation value is obtained;

[0012] The horizontal radiation value and the vertical radiation value are distributed to three-dimensional point cloud, and urban size three-dimensional night artificial light distribution is visualized.

[0013] Preferably, the night light remote sensing image data includes high-resolution panchromatic night light remote sensing data with resolution range of 0.1m to 20m, and low-resolution multispectral night light remote sensing data with resolution range of greater than 20m.

[0014] Preferably, when super-resolution enhancement processing is carried out, the unit area radiation value of each land use type at different light wave bands in each multispectral pixel is estimated by least square method, and high-resolution multispectral data with resolution range of 0.1m to 20m is generated by combining the spectral information of multispectral image and the spatial details of panchromatic image.

[0015] Preferably, the horizontal point cloud data in the three-dimensional point cloud data is obtained by discretizing the city horizontal surface into grid units through the ground elevation data and generating horizontal points at the center of each grid;The facade point cloud data in the three-dimensional point cloud data is obtained by discretizing the building facade into vertical grid units through the building shape vector data and generating facade points at the center of each grid.

[0016] Preferably, to obtain the horizontal radiation value, the following sub-steps are performed for each point in the horizontal point cloud data in the three-dimensional point cloud data:

[0017] obtain the radiation value of the land use type to which the point belongs in each band, and obtain the comprehensive weight corresponding to the point;

[0018] multiply the radiation value of the land use type to which the point belongs in each band by the comprehensive weight corresponding to the point, and take the product as the horizontal radiation value corresponding to the point.

[0019] The comprehensive weight is the product of an area weight and a height difference weight; the area weight is calculated as the ratio of the area of the horizontal plane to which the point belongs to the total area of the land use type to which the city surface type belongs in the pixel; and the height difference weight is calculated as the ratio of the height difference of the horizontal plane to which the point belongs relative to surrounding buildings.

[0020] Preferably, the height difference weight is calculated using the following formula:

[0021]

[0022] In the formula, represents the height difference weight, represents the average height of surrounding buildings, represents the height of the i-th building in the target pixel k represents the height of the i-th building in the target pixel m represents the height of the i-th building in the target pixel represents a preset minimum value.

[0023] Preferably, the fitting skewness function uses a positively skewed distribution function.

[0024] Preferably, to obtain the vertical radiation value, a positively skewed distribution function of facade radiation values is constructed based on the horizontal radiation value, the radiation value of the facade point cloud at the ground is initialized as the horizontal radiation value, and then the radiation values of the point cloud at different heights are adjusted according to the positively skewed distribution function.

[0025] Preferably, when visualizing the city-size three-dimensional nighttime artificial light distribution, the radiation values are color-coded and displayed.

[0026] In another aspect, the present application provides a city-size three-dimensional nighttime artificial light distribution modeling system, comprising:

[0027] a data acquisition unit configured to acquire nighttime light remote sensing image data, land use type vector data, ground elevation data, and building shape vector data of a study area;

[0028] a super-resolution enhancement processing unit configured to perform super-resolution enhancement processing on the night light remote sensing image data using the land use type vector data to obtain super-resolution enhancement processing data;

[0029] a three-dimensional point cloud generation unit configured to generate three-dimensional point cloud data of a fixed resolution in a study area according to the building shape vector data and the ground elevation data;

[0030] a horizontal radiation value generation unit configured to model horizontal night light radiation based on horizontal point cloud data in the three-dimensional point cloud data, the super-resolution enhancement processing data, and the building shape vector data to obtain a horizontal radiation value;

[0031] a vertical radiation value generation unit configured to model vertical night light radiation by simulating changes in light intensity with height using a fitting skewness function based on facade point cloud data in the three-dimensional point cloud data and the horizontal radiation value to obtain a vertical radiation value;

[0032] a visualization unit configured to assign the horizontal radiation value and the vertical radiation value to the three-dimensional point cloud to visualize the urban-scale three-dimensional night artificial light distribution;

[0033] The urban-scale three-dimensional night artificial light distribution modeling system is configured to perform the steps of the urban-scale three-dimensional night artificial light distribution modeling method described above.

[0034] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0035] The application provides a city-scale three-dimensional night artificial light distribution modeling method, which comprises the following steps: acquiring night light remote sensing image data, land use type vector data, ground elevation data and building shape vector data of a research area; performing super-resolution enhancement processing on the night light remote sensing image data by using the land use type vector data, so as to obtain super-resolution enhancement processing data; generating three-dimensional point cloud data with a fixed resolution in the research area according to the building shape vector data and the ground elevation data; modeling horizontal night light radiation based on horizontal point cloud data in the three-dimensional point cloud data, the super-resolution enhancement processing data and the building shape vector data, so as to obtain a horizontal radiation value; modeling vertical night light radiation by simulating the change of illumination intensity with height by using a fitting skewness function based on facade point cloud data in the three-dimensional point cloud data and the horizontal radiation value, so as to obtain a vertical radiation value; and distributing the horizontal radiation value and the vertical radiation value to the three-dimensional point cloud, so as to visualize the city-scale three-dimensional night artificial light distribution. Compared with the prior art which mainly models the artificial night light distribution on a two-dimensional plane, the application further develops to a three-dimensional space, can simulate the light distribution at different heights in a city environment, and provides a more detailed and real city light model. The application uses night light remote sensing data as a data source, the data source is stable and rich, is suitable for single-type city area analysis, and can be widely applied to different cities and different land use type areas, and has good applicability and flexibility. The application uses the super-resolution enhancement technology based on the prior knowledge of the land use type, greatly improves the spatial resolution of the night light remote sensing image, and significantly improves the precision of the city night light distribution simulation. The fitting skewness function is used to simulate the change of the illumination intensity with the height, and the influence of the city shape is combined, so that the application can automatically and finely estimate the night light radiation values of different vertical surfaces of the city. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 A flowchart of a city-scale three-dimensional night artificial light distribution modeling method provided for the embodiment 1 of the application;

[0037] Figure 2 A normal skewness distribution curve used for modeling facade radiation values in the city-scale three-dimensional night artificial light distribution modeling method provided for the embodiment 1 of the application;

[0038] Figure 3 Night light distribution mode diagrams of four land use type areas of a city obtained by the city-scale three-dimensional night artificial light distribution modeling method provided for the embodiment 1 of the application; wherein, Figure 3 (a) in the figure is a night light distribution mode diagram of an education area, Figure 3 (b) in the figure is a night light distribution mode diagram of a commercial area, Figure 3 (c) in the figure is a night light distribution mode diagram of an industrial area,Figure 3 (d) in the diagram represents the nighttime light distribution pattern of the residential area. Detailed Implementation

[0039] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0040] Example 1:

[0041] Example 1 provides a method for modeling the distribution of artificial light at night at a city scale in three dimensions. See [link to example]. Figure 1 This includes the following steps:

[0042] Step 1: Acquire nighttime light remote sensing image data, land use type vector data, ground elevation data, and building morphology vector data of the study area;

[0043] Step 2: Use the land use type vector data to perform super-resolution enhancement processing on the nighttime light remote sensing image data to obtain super-resolution enhanced data;

[0044] Step 3: Based on the building shape vector data and the ground elevation data, generate fixed-resolution three-dimensional point cloud data within the study area;

[0045] Step 4: Based on the horizontal point cloud data in the three-dimensional point cloud data, the super-resolution enhancement processing data, and the building shape vector data, model the horizontal nighttime light radiation to obtain the horizontal radiation value;

[0046] Step 5: Based on the facade point cloud data and the horizontal radiation value in the three-dimensional point cloud data, use the fitted skewness function to simulate the change of light intensity with height, model the night light radiation in the vertical dimension, and obtain the vertical radiation value.

[0047] Step 6: Assign the horizontal and vertical radiation values ​​to the three-dimensional point cloud to visualize the distribution of artificial light at night at a city size.

[0048] It should be noted that the order of steps 2 and 3 above can be adjusted; that is, 3D point cloud data can be generated first, and then super-resolution enhancement data can be obtained. The order of steps defined in Example 1 is an example provided to clearly illustrate the solution of the present invention. The steps in Example 1 will be further explained below based on the above-described step order.

[0049] In step 1, the nighttime light remote sensing image data includes high-resolution panchromatic nighttime light remote sensing data with a resolution range of 0.1m to 20m, and low-resolution multispectral nighttime light remote sensing data with a resolution range greater than 20m.

[0050] In step 2, when performing the super-resolution enhancement processing, a data fusion method based on land use types is adopted, the unit area radiation value of each land use type at different light wave bands in each multi-spectral pixel is estimated by a least square method, and high-resolution multi-spectral data with a resolution ranging from 0.1 m to 20 m is generated by combining the spectral information of the multi-spectral image and the spatial details of the panchromatic image.

[0051] In step 2, the super-resolution enhancement processing is performed to improve the spatial resolution of the noctilucent data, thereby improving the modeling accuracy.

[0052] In step 3, the three-dimensional point cloud data generated is used to simulate the urban surface morphology, and the three-dimensional point cloud data includes horizontal point cloud data and facade point cloud data.

[0053] In step 3, the horizontal point cloud data in the three-dimensional point cloud data is discretized into grid cells by the ground elevation data, and horizontal points are generated at the center of each grid to obtain the horizontal point cloud data; the facade point cloud data in the three-dimensional point cloud data is discretized into vertical grid cells by the building morphology vector data, and facade points are generated at the center of each grid to obtain the facade point cloud data.

[0054] In step 4, according to the super-resolution enhancement processing data (including land use types) obtained in step 2, the building morphology vector data (including building heights) obtained in step 1, and the horizontal point cloud data in the three-dimensional point cloud data obtained in step 3, the horizontal noctilucent radiation is modeled to obtain the horizontal radiation value.

[0055] In step 4, to obtain the horizontal radiation value, the following sub-steps are performed for each point in the horizontal point cloud data in the three-dimensional point cloud data:

[0056] Obtain the radiation value of the land use type to which the point belongs at each wave band, and obtain the comprehensive weight corresponding to the point;

[0057] The product of the radiation value of the land use type to which the point belongs at each wave band and the comprehensive weight corresponding to the point is taken as the horizontal radiation value corresponding to the point.

[0058] The comprehensive weight is the product of an area weight and a height difference weight; the ratio of the area of the horizontal plane to which the point belongs to the total area of the land use type of the urban surface type in the pixel is calculated, and the ratio is taken as the area weight; the height difference ratio of the horizontal plane to which the point belongs relative to the surrounding buildings is calculated, and the ratio is taken as the height difference weight.

[0059] Specifically, the height difference weight is calculated by the following formula:

[0060]

[0061] In the formula, Indicates the weight of the height difference. Indicates the average height of the surrounding buildings. Indicates target pixel k Inner m The height of the building This represents the preset minimum value.

[0062] That is, if the height of the horizontal plane to which the point belongs is less than the average height of the surrounding buildings, then the height difference weight is applied. for( ) / Otherwise, the height difference weighting This is a preset minimum value ε, which can be 10. -6 This indicates that the building is not affected by radiation from surrounding buildings.

[0063] It should be noted that in step 4 of this invention, there are no requirements on the order of obtaining the radiation value of the land use type of the point in each band and obtaining the comprehensive weight corresponding to the point, nor are there any requirements on the order of obtaining the area weight and the elevation difference weight. In order to better understand this invention, step 4 will be illustrated below by taking the example of obtaining the radiation value first and then the comprehensive weight, and obtaining the area weight first and then the elevation difference weight.

[0064] In step 4, the following sub-steps are performed for each point in the initial horizontal point cloud data:

[0065] Step 4.1: Obtain the radiation values ​​of the land use type to which the point belongs in Step 2 in each band;

[0066] Step 4.2: Calculate the ratio of the horizontal surface area of ​​the point to the total area of ​​the land use type of the city surface within the pixel, and use this ratio as the area weight. , used to characterize the theoretical proportion of light radiation that can be obtained by dividing the horizontal surface by area when finely distributing horizontal radiation within a pixel;

[0067] Step 4.3: Calculate the ratio of the height difference between the horizontal plane to which the point belongs and the surrounding buildings, and use this ratio as the height difference weight. , used to characterize the theoretical proportion of light radiation that can be obtained from the horizontal surface divided according to the urban elevation difference when finely distributing horizontal radiation within a pixel;

[0068] Step 4.4, calculate the area weights from steps 4.2 and 4.3. and height difference weight The product of these factors is used as the overall weight. ,Right now This is used to comprehensively characterize the influence of urban morphology on the horizontal surface where the point is located within the pixel;

[0069] Step 4.5, calculating the product of the radiation value of the land use type where the point is located in each band and the comprehensive weight obtained in step 4.4 as the horizontal radiation value, which is used to represent the theoretical light radiation that the horizontal surface to which the point belongs can obtain under the comprehensive influence of urban morphology when the horizontal radiation is finely distributed within the pixel.

[0070] In step 5, the fitting skewness function can adopt a positively skewed distribution function to simulate the trend that the radiation value of the facade point cloud increases rapidly and then decreases slowly with height, and a schematic diagram of the positively skewed distribution curve is shown in Figure 2

[0071] To obtain the vertical radiation value, a positively skewed distribution function of facade radiation value is constructed based on the horizontal radiation value, the radiation value of the facade point cloud at the ground is initialized as the horizontal radiation value, and then the radiation value of the point cloud at different heights is adjusted according to the positively skewed distribution function.

[0072] In step 6, when visualizing the three-dimensional artificial light distribution of urban size at night, the radiation value is color graded and displayed.

[0073] The application will be further illustrated with specific parameters.

[0074] Step 1, obtaining night light remote sensing image data (including night light remote sensing multispectral image data, night light remote sensing panchromatic image data), land use type vector data, ground elevation data and building morphology vector data of the research area, and performing initial processing on the original data.

[0075] Preferably, data of a unified year is adopted, and the specific operation of this step is as follows:

[0076] Obtain night light remote sensing multispectral image data (40m resolution) and panchromatic image data (10m resolution) from SDGSAT-1 satellite images, extract building morphology vector data from 3D-GloBFP data set, obtain ground elevation data, and obtain land use type vector data from EULUC-China data set, which includes 12 different land use types, and perform standardization processing on the obtained data.

[0077] Project all the above data to the WGS84 coordinate system, resample the image data, align the grid grids of the multispectral and panchromatic images, and ensure that each 40m pixel includes exactly 16 10m panchromatic pixels.

[0078] Step 2, performing super-resolution enhancement processing using low-resolution multispectral image, high-resolution panchromatic image and land use type vector data.

[0079] The above step 2 further includes the following sub-steps:​​

[0080] Step 2.1, calculate the spectral index, combine the spectral information of 40m lower resolution multispectral data and the spatial detail information of 10m high resolution panchromatic data to generate high resolution multispectral data initially.

[0081] In the embodiment, the specific operation of this step is as follows:

[0082] Obtain the 40m resolution multispectral image of SDGSAT-1 satellite data, 10m resolution panchromatic image, and urban land use type vector data.

[0083] Using the 40m resolution multispectral image, according to the radiation values of red, green, and blue bands of each 40m pixel , calculate the relative contribution of each band to the overall brightness to obtain the spectral index , the calculation formula is as follows:

[0084]

[0085] In the formula, respectively represent the radiation values of red, green, and blue bands of any 40m pixel, respectively refer to the red, green, and blue bands of multispectral data, and the spectral index is effective for any 10m pixel in the 40m multispectral pixel.

[0086] Combined with the 10m resolution panchromatic image, estimate the radiation values of each band according to the spectral index to obtain the preliminary 10m high resolution multispectral data:

[0087]

[0088] In the formula, refers to high resolution multispectral data, refers to 10m resolution panchromatic data, refers to the radiation value, refers to the red, green, and blue bands of the preliminary high resolution multispectral data. Therefore is the radiation value of the panchromatic grid data after radiation calibration, refers to the spectral weight calculated above, is the estimated radiation value of the high resolution multispectral data in the red, green, and blue bands respectively.

[0089] Step 2.2, the area belonging to the same land use type has similar unit area radiation value, and the unit area radiation value of each land use type is calculated according to this use based on the least square method.

[0090] In the embodiment, the specific operation of this step is as follows:

[0091] Combining the LULC data within the 40m x 40m pixel, it is assumed that the areas of the same land use type have similar radiance values per unit area. For each 40m x 40m pixel, the following equation set is established:

[0092]

[0093] where, is the area matrix of the 12 land use types in each panchromatic pixel, with the dimension of 16 x 12, is the column vector of the radiance values per unit area of the 12 land use types in each band (R, G, B), is the column vector of the preliminary estimated radiance values of the 16 10m high-resolution multispectral pixels.

[0094] Subsequently, the equation set is solved by the least square method to obtain the optimal radiance values per unit area of each land use type in each band (R, G, B).

[0095] Step 2.3, revising the radiance values per unit area of the land use types within the 10m high-resolution multispectral pixels using the initial 10m panchromatic data radiance values.

[0096] In the embodiment, the specific operation of this step is as follows:

[0097] Using the known radiance values per unit area of the land use types, the radiance values of each 10m pixel of the high-resolution multispectral data are recalculated, and then, according to the revised high-resolution multispectral data, the spectral indices are recalculated, and based on the 10m PAN radiance values and the updated spectral indices, the radiance values of each band are redistributed:

[0098]

[0099]

[0100]

[0101] where, denotes any pixel of the high-resolution multispectral data, denotes the recalculated high-resolution multispectral radiance value at the k pixel, denotes the red, green, and blue bands, respectively, is the area weight of each of the 12 main land use types within the k pixel, is the land use type radiance value per unit area calculated by the least square method in step 2.2, is the updated spectral index, ​​It is the high-resolution multispectral radiometric value of the k-th pixel after correction by panchromatic data radiometric values.

[0102] Step 2.4: Recalculate the final unit area radiation value for each land use type based on the corrected high-resolution multispectral radiation value in Step 2.3.

[0103] In this embodiment, this step is specifically performed as follows:

[0104] The final radiation value per unit area for each land use type is calculated using the following formula:

[0105]

[0106] In the formula, k Refers to any pixel in high-resolution multispectral data. It is one of the 12 land use types. In the first k The area weights of the 12 main land use types within a pixel. The radiation value per unit area for the land use type is calculated using the least squares method in step 2.2. For the first k High-resolution multispectral radiometric values ​​of pixels after correction of panchromatic data radiometric values.

[0107] Using the above method, the radiation ratio between each band and land use type can be maintained while taking into account the physical characteristics of the radiation distribution of land use types, ensuring consistency with the original panchromatic data.

[0108] Step 3: Generate fixed-resolution 3D point cloud data based on building shape vector data and ground elevation data.

[0109] In this embodiment, the specific operation of this step is as follows:

[0110] The horizontal plane of the study area is discretized into grid cells with a resolution of 1m, and a horizontal point is generated at the center of each grid cell to form a horizontal point cloud.

[0111] The building facades within the study area are discretized into vertical grid cells with a resolution of 1m, and facade points are generated at the center of each grid to form a facade point cloud.

[0112] The horizontal point cloud and the elevation point cloud are merged to form a complete 3D point cloud model of the city with a resolution of 1m. Each point in the point cloud contains the following attributes: ID number, building number gid (where the point belongs to the road surface with gid=0), height, and spatial coordinates.

[0113] Step 4, modeling horizontal night light radiation based on horizontal point cloud data in the three-dimensional point cloud data, the super-resolution enhancement processing data and the building shape vector data, to obtain horizontal radiation value.

[0114] That is, step 4 takes the high-resolution multi-spectral data corrected in step 2 as the initial night light radiation indicator value, and combines building height and land use type to finely estimate horizontal night light radiation.

[0115] The above step 4 further includes the following sub-steps for each point in the three-dimensional point cloud:

[0116] Step 4.1, obtaining the estimated radiation value of the land use type where the point is located in each waveband calculated by step 2, and recording the point id number;

[0117] Step 4.2, calculating the ratio of the area of the horizontal plane (building roof or ground) where the point belongs to the total area of the land use type where the city surface type belongs in the pixel as the area weight , which is used to represent the proportion of theoretical light radiation that can be allocated to the horizontal surface according to area division when finely allocating horizontal radiation in the pixel;

[0118] Step 4.3, calculating the height difference ratio of the horizontal plane where the point belongs to the surrounding buildings as the height difference weight , which is used to represent the proportion of theoretical light radiation that can be allocated to the horizontal surface according to city height difference when finely allocating horizontal radiation in the pixel;

[0119] In the embodiment, the specific operation of this step is as follows:

[0120] The height difference weight is calculated as follows:

[0121]

[0122] The above formula shows that if the height of the horizontal plane where the point is located is less than the average height of the surrounding buildings, the horizontal plane will be affected by the facade light radiation of the surrounding buildings, and the lower the height, the greater the impact, that is, the height difference weight is greater; otherwise, the height difference weight takes a minimum value , which is 10 -6 in this example, indicating that the building is not affected by the facade radiation of the surrounding buildings due to its height.

[0123] Step 4.4, calculating the product of the area weight and the height difference weight in step 4.2 and step 4.3 as the comprehensive weight , that is , for representing the influence of the urban form on the horizontal surface where the point is located in the pixel.

[0124] Step 4.5, calculating the product of the radiation value of the land use type where the point is located in each band obtained in step 4.1 and the comprehensive weight obtained in step 4.4, for finally representing the theoretical light radiation that can be allocated to the horizontal surface where the point belongs under the influence of the urban form and the land use type to which the horizontal surface belongs when the horizontal radiation is allocated in the pixel:

[0125]

[0126] wherein, is the horizontal radiation value calculated for the point, is the radiation value of the land use type where the point is located in each band obtained in step 4.1, is the comprehensive weight obtained in step 4.4, is the total number of horizontal points included in the horizontal surface to which the point belongs.

[0127] Step 5, based on the facade point cloud data in the three-dimensional point cloud data and the horizontal radiation value, modeling the vertical dimension night light radiation by simulating the change of the light intensity with the height by fitting the skewness function, to obtain the vertical radiation value.

[0128] That is, step 5 simulates the change of the light intensity with the height by fitting the skewness function, to estimate the urban vertical dimension night light radiation.

[0129] In the embodiment, the specific operation of the present step is as follows:

[0130] As shown in the embodiment, the skewness distribution function is expressed by using the expression of the chi-square distribution function, for fitting the trend that the radiation value of the facade point cloud first rapidly rises and then slowly falls with the increase of the height, and the function calculation formula is: Figure 2

[0131]

[0132] wherein, h is the vertical point height, is the light radiation value of the vertical point at the height h = 11, for ensuring that the vertical point cloud reaches the peak value at the height of 9 m, and the slope parameter k .

[0133] Wherein, the chi-square distribution function is an example of fitting the skewness function, in addition, the present application can also use the exponential function, the power function, and various combinations of their mathematics.

[0134] ​​​The radiation value of the bottom of the building facade is initialized to be equal to the average radiation value of the horizontal point cloud within a radius of 0.5 m, and the vertical point cloud radiation value at any height of the building is:

[0135] .

[0136] wherein, is equal to the average radiation value of the horizontal point cloud within a radius of 0.5 m, is the vertical point cloud radiation value at the height of the building. h

[0137] Step 6, distribute the horizontal and vertical radiation values to the three-dimensional point cloud, and visualize the urban size three-dimensional nighttime artificial light distribution.

[0138] That is, step 6 distributes the calculated radiation values to the three-dimensional point cloud according to the horizontal and vertical radiation modeling results and visualizes them.

[0139] In the embodiment, the specific operation of this step is as follows:

[0140] According to the calculation results of steps 4 and 5, the calculated radiation values are stored as new attributes of the points according to the id number recorded in step 4.1.

[0141] For example, the SDGSAT-1 night light image of a certain city in China in 2023 and the EULUC-China national urban land use type data and building shape data are selected. First, 400m×400m research areas are selected in four different land use types (education area, business area, industrial area, residential area) in a certain city, and the following operations are performed on each research area. First, the research area is super-resolution enhanced using the SDGSAT-1 night light image combined with the EULUC-China national urban land use type vector data, and the urban three-dimensional point cloud is generated combined with the building shape vector data. Then, the radiation values of the horizontal point cloud and the vertical point cloud in the research area are modeled, and finally the modeled radiation values are stored as new attributes of the points, and the point cloud radiation values are divided into 10 levels, and the linear color transformation is selected in the Arcgis pro software to display the nighttime light distribution pattern of the four land use type areas in the city from blue to red color classification, ensuring that the model can accurately reflect the characteristics of urban night light distribution, as shown in Figure 3 , wherein, Figure 3 (a) in (a) is the nighttime light distribution pattern of the education area, Figure 3 (b) in (b) is the nighttime light distribution pattern of the business area, Figure 3 (c) in (c) is the nighttime light distribution pattern of the industrial area, Figure 3 (d) in (d) is the nighttime light distribution pattern of the residential area. ​

[0142] Since one of the data sources acquired and utilized by the present application is night light remote sensing image, and the night light remote sensing image obtained by the satellite has the characteristics of wide range and repeatable observation, the horizontal light distribution can be quickly obtained. And based on the research, the vertical light distribution of the building facade is affected by the horizontal light distribution of the ground surface where the building is located. Based on this important conclusion, the present application estimates the vertical light distribution of the building using the horizontal light distribution of the ground surface, and further realizes the modeling of the three-dimensional night light distribution of the city scale as a whole.

[0143] The present application can realize the fine estimation of the city night light radiation by combining multi-source data, super-resolution enhancement technology and light radiation modeling, and the accuracy verification is carried out by using the spatial overlay analysis of the multispectral remote sensing image and the street view image, so as to improve the accuracy, efficiency and reliability of the modeling.

[0144] The present application establishes a high-precision city size three-dimensional night artificial light distribution modeling method based on night light remote sensing image by comprehensively using remote sensing image processing, three-dimensional modeling and radiation modeling technology, which provides a scientific basis for city light pollution assessment and energy consumption analysis.

[0145] Embodiment 2:

[0146] Embodiment 2 provides a city size three-dimensional night artificial light distribution modeling system, comprising:

[0147] A data acquisition unit is configured to acquire night light remote sensing image data, land use type vector data, ground elevation data and building shape vector data of a research area;

[0148] A super-resolution enhancement processing unit is configured to perform super-resolution enhancement processing on the night light remote sensing image data by using the land use type vector data, to obtain super-resolution enhancement processing data;

[0149] A three-dimensional point cloud generation unit is configured to generate three-dimensional point cloud data of a fixed resolution in the research area according to the building shape vector data and the ground elevation data;

[0150] A horizontal radiation value generation unit is configured to model horizontal night light radiation based on horizontal point cloud data in the three-dimensional point cloud data, the super-resolution enhancement processing data and the building shape vector data, to obtain a horizontal radiation value;

[0151] A vertical radiation value generation unit is configured to model vertical night light radiation by simulating the change of illumination intensity with height by using a fitting skewness function based on facade point cloud data in the three-dimensional point cloud data and the horizontal radiation value, to obtain a vertical radiation value;

[0152] a visualization unit for assigning the horizontal and vertical radiance values to a three-dimensional point cloud for visualizing the urban scale three-dimensional night-time artificial light distribution;

[0153] The urban scale three-dimensional night-time artificial light distribution modeling system of embodiment 2 is configured to perform the steps of the urban scale three-dimensional night-time artificial light distribution modeling method of embodiment 1.

[0154] Since the functions of the units of the urban scale three-dimensional night-time artificial light distribution modeling system of embodiment 2 correspond to the steps of the urban scale three-dimensional night-time artificial light distribution modeling method of embodiment 1, the description of embodiment 1 can be referred to for understanding embodiment 2, which will not be repeated here.

[0155] Finally, it should be noted that the above detailed description is only used to explain the technical solutions of the present application and not to limit the present application. Although the present application has been described in detail with reference to the examples, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A method of modelling urban scale three-dimensional night-time artificial light distribution, characterized in that, The method comprises the following steps: obtaining night light remote sensing image data, land use type vector data, ground elevation data and building shape vector data of a study area; performing super-resolution enhancement processing on the night light remote sensing image data by using the land use type vector data to obtain super-resolution enhancement processing data; generating fixed-resolution three-dimensional point cloud data in the study area according to the building shape vector data and the ground elevation data; horizontal point cloud data in the three-dimensional point cloud data is obtained by discretizing the urban horizontal surface into grid units by the ground elevation data and generating a horizontal point at the center of each grid; facade point cloud data in the three-dimensional point cloud data is obtained by discretizing the building facade into vertical grid units by the building shape vector data and generating a facade point at the center of each grid; modeling horizontal night light radiation based on the horizontal point cloud data in the three-dimensional point cloud data, the super-resolution enhancement processing data and the building shape vector data to obtain horizontal radiation values; to obtain the horizontal radiation values, the following sub-steps are performed for each point in the horizontal point cloud data in the three-dimensional point cloud data: obtaining radiation values of a land use type to which the point belongs in each waveband, and obtaining a comprehensive weight corresponding to the point; multiplying the radiation values of the land use type to which the point belongs in each waveband by the comprehensive weight corresponding to the point to obtain a horizontal radiation value corresponding to the point; wherein the comprehensive weight is the product of an area weight and a height difference weight; calculating a ratio of an area of a horizontal plane to which the point belongs to a total area of a land use type of a city surface type within a pixel, and taking the ratio as the area weight; calculating a height difference ratio of the horizontal plane to which the point belongs relative to surrounding buildings, and taking the ratio as the height difference weight; modeling vertical night light radiation based on the facade point cloud data in the three-dimensional point cloud data and the horizontal radiation values by using a fitted skewness function to simulate changes in illumination intensity with height to obtain vertical radiation values; distributing the horizontal radiation values and the vertical radiation values to the three-dimensional point cloud to visualize city-size three-dimensional artificial light distribution at night.

2. The urban-scale three-dimensional nighttime artificial light distribution modeling method of claim 1, wherein, The night light remote sensing image data comprises high-resolution panchromatic night light remote sensing data with a resolution range of 0.1m to 20m, and low-resolution multispectral night light remote sensing data with a resolution range of greater than 20m.

3. The urban-scale three-dimensional nighttime artificial light distribution modeling method of claim 2, wherein, When performing super-resolution enhancement processing, the unit area radiation value of each land use type at different light wavebands in each multispectral pixel is estimated by a least squares method, and high-resolution multispectral data with a resolution range of 0.1m to 20m is generated in combination with spectral information of the multispectral image and spatial details of the panchromatic image.

4. The urban-scale three-dimensional nighttime artificial light distribution modeling method of claim 1, wherein, The height difference weight is calculated using the following formula: In the formula, represents the height difference weight, represents the average height of surrounding buildings, represents the target pixel k the height of the m building, represents a preset minimum value.

5. The urban-scale three-dimensional nighttime artificial light distribution modeling method of claim 1, wherein, The fitted skewness function uses a positive skewness distribution function.

6. The urban-scale three-dimensional nighttime artificial light distribution modeling method of claim 5, wherein, To obtain the vertical radiation values, a positive skewness distribution function of facade radiation values is constructed based on the horizontal radiation values, the radiation value of the facade point cloud at the ground is initialized as the horizontal radiation value, and then the radiation values of the point cloud at different heights are adjusted according to the positive skewness distribution function.

7. The urban-scale three-dimensional nighttime artificial light distribution modeling method of claim 1, wherein, When visualizing the urban-scale three-dimensional nighttime artificial light distribution, the magnitude of the radiance values is color graded.

8. A system for modeling urban-scale three-dimensional nighttime artificial light distribution, characterized in that, The method comprises the steps of: acquiring nighttime light remote sensing image data, land use type vector data, ground elevation data, and building shape vector data of a study area; performing super-resolution enhancement processing on the nighttime light remote sensing image data using the land use type vector data to obtain super-resolution enhancement processing data; generating three-dimensional point cloud data of a fixed resolution in the study area according to the building shape vector data and the ground elevation data; modeling horizontal nighttime light radiance based on horizontal point cloud data in the three-dimensional point cloud data, the super-resolution enhancement processing data, and the building shape vector data to obtain horizontal radiance values; modeling vertical nighttime light radiance based on facade point cloud data in the three-dimensional point cloud data and the horizontal radiance values using a fitting skewness function to simulate the change of light intensity with height to obtain vertical radiance values; visualizing the urban-scale three-dimensional nighttime artificial light distribution by assigning the horizontal radiance values and the vertical radiance values to the three-dimensional point cloud; The urban-scale three-dimensional nighttime artificial light distribution modeling system is configured to perform the steps of the urban-scale three-dimensional nighttime artificial light distribution modeling method according to any one of claims 1-7.

Images