Urban green space extraction method based on prior knowledge and remote sensing data

Abstract

The application discloses a kind of urban green space extraction methods based on prior knowledge and remote sensing data, method includes: obtaining remote sensing data and pre-processing to obtain multispectral image data, select several preset bands of multispectral image data, set the prior knowledge threshold value corresponding to each band one by one, the pixel point of band is compared with corresponding prior knowledge threshold value, according to the comparison result, pixel point is initialized as first value or second value, obtain several preliminary hierarchical data, based on the normalized difference vegetation index, after the integrated data is corrected, the modified integrated data is obtained;According to the distribution result of multiple ground object types obtained by the modified integrated data and the preset classification rule. According to the result of prior knowledge extraction, the quality is high, the robustness is stronger, can quickly divide urban green space and other ground object types, reduce workload, can be widely applied in data processing technical field.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method for extracting urban green space based on prior knowledge and remote sensing data. Background Technology

[0002] Urban green space is an important indicator of urban infrastructure. Utilizing the spectral characteristics of green areas in remote sensing data, the extent and location of urban green spaces can be quickly extracted, providing crucial reference for urban planners. Traditional remote sensing methods can be used to extract urban green areas. Vegetation indices are methods that assess vegetation cover by calculating the ratios or differences between different bands in remote sensing data. Commonly used vegetation indices include the Normalized Difference Vegetation Index (NDVI) and the Green Light Vegetation Index (GVI). By calculating these indices and setting appropriate thresholds, green areas in cities can be extracted.

[0003] Traditional remote sensing methods cannot retain information on other urban land cover types when extracting urban green space features, and lack detailed classification of other land cover types. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method for extracting urban green space based on prior knowledge and remote sensing data, which can automatically extract green features, water bodies and impermeable surfaces, and solves the problem that other urban feature types cannot be preserved when extracting urban green features.

[0005] In a first aspect, embodiments of the present invention provide a method for extracting urban green space based on prior knowledge and remote sensing data, comprising the following steps:

[0006] Acquire remote sensing data and preprocess the remote sensing data to obtain multispectral image data;

[0007] Several preset bands are selected based on the multispectral image data;

[0008] Based on several preset bands and prior knowledge, a prior knowledge threshold corresponding to each preset band is set. The pixel of each preset band is compared with the corresponding prior knowledge threshold to obtain a comparison result. According to the comparison result, the pixel is initialized to a first value or a second value. After initialization, several preliminary layered data are obtained. Among them, one preset band corresponds to one preliminary layered data, and one pixel has a corresponding first value or second value in different initial layered data.

[0009] The integrated data is obtained by integrating several of the preliminary hierarchical data.

[0010] The normalized vegetation index is calculated by selecting vegetation bands based on the multispectral image data, and the integrated data is then corrected based on the normalized vegetation index to obtain corrected integrated data.

[0011] The distribution results of multiple land cover types are obtained based on the modified and integrated data and the preset classification rules.

[0012] Optionally, selecting several preset bands based on the multispectral image data specifically includes:

[0013] Input the multispectral image data into a preset selection model;

[0014] The selection model selects the green band, red edge band, water vapor absorption band, shortwave infrared 1 band, and shortwave infrared 2 band from the multispectral image data.

[0015] Optionally, the step of setting prior knowledge thresholds corresponding to each of the preset bands based on several preset bands and prior knowledge specifically includes:

[0016] A prior knowledge threshold table is generated based on several preset bands and prior knowledge.

[0017] The prior knowledge threshold is set according to the prior knowledge threshold table, corresponding to each preset band.

[0018] Optionally, the step of comparing each pixel of the preset band with the corresponding prior knowledge threshold to obtain a comparison result, and initializing the pixel to a first value or a second value according to the comparison result, specifically includes:

[0019] Obtain the surface reflectance data of the pixel in several preset wavelength bands:

[0020] Based on the surface reflectance data of the pixels in the same preset band and the prior knowledge threshold;

[0021] Compare the surface reflectance data under the same preset band with the prior knowledge threshold;

[0022] The pixel is initialized to a first value or a second value according to the comparison result and the preset comparison rule.

[0023] Optionally, the step of initializing the pixel to a first value or a second value according to the comparison result and a preset comparison rule specifically includes:

[0024] The comparison results of the pixel points in the green band with the first prior threshold, the red edge band with the second prior threshold, the water vapor absorption band with the third prior threshold, the shortwave infrared 1 band with the fourth prior threshold, and the shortwave infrared 2 band with the fifth prior threshold are obtained respectively.

[0025] Pixels with values ​​greater than the first prior threshold are initialized to a first value, and pixels with values ​​less than the first prior threshold are initialized to a second value, thus obtaining the first preliminary layering data.

[0026] Pixels with values ​​greater than the second prior threshold are initialized to a second value, and pixels with values ​​less than the second prior threshold are initialized to a first value, to obtain second preliminary layering data;

[0027] Pixels with values ​​greater than the third prior threshold are initialized to a second value, and pixels with values ​​less than the third prior threshold are initialized to a first value, thus obtaining the third preliminary layering data.

[0028] Pixels with values ​​greater than the fourth prior threshold are initialized to a second value, and pixels with values ​​less than the fourth prior threshold are initialized to a first value, thus obtaining the fourth preliminary layering data.

[0029] Pixels with values ​​greater than the fifth prior threshold are initialized to a first value, and pixels with values ​​less than the fifth prior threshold are initialized to a second value, thus obtaining the fifth preliminary layering data.

[0030] Optionally, the process of integrating the preliminary hierarchical data to obtain integrated data specifically includes:

[0031] The integrated data is obtained by superimposing several of the initial layered data, and the superposition formula is as follows:

[0032] Y1 = x1 + x2 + x3 + x4 + x5

[0033] In the formula, Y1 is the integrated data, and the data size ranges from 0 to 5; x1, x2, x3, x4, and x5 are the preliminary layered data.

[0034] Optionally, the step of selecting vegetation bands based on the multispectral image data to calculate the normalized vegetation index, and then correcting the integrated data based on the normalized vegetation index to obtain corrected integrated data, specifically includes:

[0035] Obtain the red band and near-infrared band from the multispectral image data;

[0036] The normalized vegetation index is calculated based on the red band and the near-infrared band.

[0037] The intermediate parameters are calculated based on the normalized vegetation index of the pixel and the preset intermediate parameter calculation formula.

[0038] The intermediate parameters are used to correct the integrated data to obtain the corrected integrated data.

[0039] Secondly, embodiments of the present invention provide an urban green space extraction system based on prior knowledge and remote sensing data, comprising:

[0040] The first module is used to acquire remote sensing data and preprocess the remote sensing data to obtain multispectral image data.

[0041] The second module is used to select several preset bands based on the multispectral image data;

[0042] The third module is used to set a threshold for prior knowledge corresponding to each of the preset bands based on several preset bands and prior knowledge.

[0043] The fourth module is used to compare the pixel of each preset band with the corresponding prior knowledge threshold to obtain the comparison result;

[0044] The fifth module is used to initialize the pixel to a first value or a second value according to the comparison result, and after initialization, a number of preliminary layer data are obtained. Among them, one preset band corresponds to one preliminary layer data, and one pixel has a corresponding first value or second value in different initial layer data.

[0045] The sixth module is used to integrate the aforementioned preliminary hierarchical data to obtain integrated data;

[0046] The seventh module is used to select vegetation bands based on the multispectral image data to calculate the normalized vegetation index.

[0047] The eighth module is used to correct the integrated data based on the normalized vegetation index to obtain corrected integrated data.

[0048] The ninth module is used to obtain the distribution results of multiple land cover types based on the corrected and integrated data and the preset classification rules.

[0049] Thirdly, embodiments of the present invention provide an urban green space extraction device based on prior knowledge and remote sensing data, comprising:

[0050] At least one processor;

[0051] At least one memory for storing at least one program;

[0052] When the at least one program is executed by the at least one processor, the at least one processor performs the method as described above.

[0053] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to perform the method described above.

[0054] The implementation of this invention provides the following beneficial effects: This invention provides a method for extracting urban green space based on prior knowledge and remote sensing data, comprising: acquiring remote sensing data; preprocessing the remote sensing data to obtain multispectral image data; selecting several preset bands based on the multispectral image data; setting prior knowledge thresholds corresponding to each preset band based on the preset bands and prior knowledge; comparing the pixels of each preset band with the corresponding prior knowledge threshold to obtain a comparison result; initializing the pixels to a first value or a second value according to the comparison result; obtaining several preliminary layered data after initialization, wherein one preset band corresponds to one preliminary layered data, and one pixel has a corresponding first value or second value in different initial layered data; integrating the several preliminary layered data to obtain integrated data; selecting vegetation bands based on the multispectral image data to calculate the normalized vegetation index (NVI); correcting the integrated data based on the NVI to obtain corrected integrated data; and obtaining the distribution results of multiple land cover types according to the corrected integrated data and preset classification rules. After acquiring remote sensing data and extracting the corresponding bands, prior knowledge is used to obtain the prior thresholds for pixels in the corresponding bands. The pixels are then initialized to a first or second value, resulting in preliminary stratification data for pixels in different bands. This preliminary stratification data is then integrated and corrected using the normalized vegetation index (NVI) to obtain corrected and integrated data. Finally, based on preset classification rules, the distribution results of multiple land cover types are obtained. While extracting urban green spaces, other urban land cover types are automatically stratified, effectively distinguishing the scope of urban green spaces. Furthermore, the results extracted based on prior knowledge are of high quality and more robust, enabling rapid urban green space delineation and reducing workload. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the steps of a method for extracting urban green space based on prior knowledge and remote sensing data provided in an embodiment of the present invention;

[0056] Figure 2 This is a flowchart illustrating another method for extracting urban green space based on prior knowledge and remote sensing data provided in an embodiment of the present invention.

[0057] Figure 3This is a true-color image of the study area provided in an embodiment of the present invention;

[0058] Figure 4 This is an extraction map of green space, water body, and impermeable surface in the study area provided in an embodiment of the present invention;

[0059] Figure 5 This is a structural block diagram of an urban green space extraction system based on prior knowledge and remote sensing data provided in an embodiment of the present invention;

[0060] Figure 6 This is a structural block diagram of an urban green space extraction device based on prior knowledge and remote sensing data provided in an embodiment of the present invention. Detailed Implementation

[0061] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0062] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0063] In the description of this invention, "several" means one or more, "multiple" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0064] In the description of this invention, unless otherwise explicitly defined, terms such as "set up," "install," and "connect" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0065] like Figure 1 As shown in the figure, this embodiment of the invention provides a method for extracting urban green space based on prior knowledge and remote sensing data, which includes the following steps.

[0066] S100. Acquire remote sensing data and preprocess the remote sensing data to obtain multispectral image data.

[0067] The remote sensing data used were from Sentinel-2 / Sentinel-1 satellites.

[0068] Reference Figure 2 In one specific embodiment, the remote sensing data source uses second-level data from Sentinel-2 imagery. Using first-level data necessitates image preprocessing, including atmospheric correction and geometric correction. Preprocessing the Sentinel-2 remote sensing data yields directly usable Sentinel-2 multispectral imagery. (Refer to...) Figure 3 Taking a specific area as the study area, its pixel size is 1022x1256. Remote sensing data has advantages such as large scale, multi-temporal, rapid, and efficient monitoring of spatial distribution changes, and can obtain accurate land cover information; however, when extracting green space from remote sensing data, it is usually impossible to simultaneously stratify other land cover types. In this embodiment, Sentinel-2 multispectral image data is processed using prior knowledge to obtain multi-layer data. After correcting the multi-layer data, the distribution results of multiple land cover types are obtained through preset classification rules. While extracting urban green space, other urban land cover types are automatically stratified, which can effectively distinguish the scope of urban green space.

[0069] S200. Select several preset bands based on the multispectral image data.

[0070] Specifically, after acquiring Sentinel-2 multispectral image data, multiple corresponding bands are extracted from the multispectral image data. By extracting the bands in advance, the memory usage of terminals such as computers and servers during processing is reduced, resulting in faster processing speed.

[0071] Optionally, selecting several preset bands based on the multispectral image data specifically includes:

[0072] S210. Input the multispectral image data into a preset selection model;

[0073] S220, The selection model selects the green band, red edge band, water vapor absorption band, shortwave infrared 1 band, and shortwave infrared 2 band from the multispectral image data.

[0074] Specifically, a selection model is established in advance based on prior knowledge, and the corresponding green band, red edge band, water vapor absorption band, shortwave infrared 1 band, and shortwave infrared 2 band in the multispectral image data are selected through the established selection model.

[0075] S300. Based on several preset bands and prior green space knowledge, set prior knowledge thresholds corresponding to each preset band.

[0076] Specifically, multiple preset bands are obtained through prior knowledge acquired from advance experiments, and a prior knowledge threshold corresponding to each band is set. The prior knowledge threshold is a threshold that can quickly extract classification results, determined through a large number of experiments.

[0077] S400. Compare the pixel points of each preset band with the corresponding prior knowledge threshold to obtain the comparison result.

[0078] Specifically, the relationship between reflectance and prior knowledge threshold is obtained by comparing the reflectance of pixels in a preset band with their corresponding prior knowledge threshold.

[0079] S500. Based on the comparison result, the pixel is initialized to a first value or a second value. After initialization, several preliminary layered data are obtained. One preset band corresponds to one preliminary layered data. One pixel has a corresponding first value or second value in different initial layered data.

[0080] Specifically, the prior knowledge is acquired in advance based on a large amount of experimental data. This prior knowledge, summarized from the experimental results, enables the accurate extraction of different land cover types. First, the corresponding pixels of the extracted bands are compared with thresholds acquired based on the prior knowledge. Based on the surface reflectance data of the pixels and the corresponding thresholds for different bands, the pixels in those bands are initialized to either a first or second value. This initialization yields preliminary layering data for the pixels in that band, resulting in multiple preliminary layering data for multiple bands, with each band having one set of preliminary layering data.

[0081] Optionally, the step of setting prior knowledge thresholds corresponding to each of the preset bands based on several preset bands and prior knowledge specifically includes:

[0082] S310. Generate a prior knowledge threshold table based on several preset bands and prior knowledge;

[0083] S320. Set the prior knowledge threshold corresponding to each preset band according to the prior knowledge threshold table.

[0084] Specifically, a priori knowledge threshold table is pre-generated based on prior knowledge, corresponding to the threshold for each extracted band. This table is stored, and when a corresponding band is extracted, the threshold corresponding to that band is automatically obtained from the prior knowledge threshold table. If a threshold corresponding to a certain band cannot be found in the prior knowledge table, an error is automatically reported, and several preset bands are extracted again, and the threshold is obtained again. If the threshold still cannot be obtained after multiple attempts, an error report is generated for easy review and repair.

[0085] Optionally, the step of comparing each pixel of the preset band with the corresponding prior knowledge threshold to obtain a comparison result, and initializing the pixel to a first value or a second value according to the comparison result, specifically includes:

[0086] S330. Obtain the surface reflectance data of the pixel in several preset bands;

[0087] S340. Based on the surface reflectance data of the pixel points in the same preset band and the prior knowledge threshold;

[0088] S350. Compare the surface reflectance data under the same preset band with the prior knowledge threshold.

[0089] S360. Based on the comparison result and the preset comparison rules, initialize the pixel to a first value or a second value.

[0090] Specifically, each band corresponds to the same pixel, and multiple bands represent different spectral images of the same area, i.e., acquiring different spectral images of the same area. The surface reflectance data of each pixel in each band is acquired. The surface reflectance data of a pixel in a certain band is compared with the corresponding threshold for that band. The threshold values ​​differ for different bands. After comparison, the pixel is initialized to a first or second value based on the surface reflectance data and the threshold value.

[0091] Optionally, the step of initializing the pixel to a first value or a second value according to the comparison result and a preset comparison rule specifically includes:

[0092] S361. Obtain the comparison results of the pixel point in the green band with the first prior threshold, the red edge band with the second prior threshold, the water vapor absorption band with the third prior threshold, the shortwave infrared 1 band with the fourth prior threshold, and the shortwave infrared 2 band with the fifth prior threshold, respectively.

[0093] S362. Initialize the pixels that are greater than the first prior threshold to a first value, and initialize the pixels that are less than the first prior threshold to a second value to obtain the first preliminary layering data.

[0094] S363. Initialize the pixels that are greater than the second prior threshold to a second value, and initialize the pixels that are less than the second prior threshold to a first value to obtain the second preliminary layering data;

[0095] S364. Initialize the pixels that are greater than the third prior threshold to the second value, and initialize the pixels that are less than the third prior threshold to the first value to obtain the third preliminary layering data.

[0096] S365. Initialize the pixels that are greater than the fourth prior threshold to the second value, and initialize the pixels that are less than the fourth prior threshold to the first value to obtain the fourth preliminary layering data.

[0097] S366. Initialize the pixels that are greater than the fifth prior threshold to a first value, and initialize the pixels that are less than the fifth prior threshold to a second value, to obtain the fifth preliminary layering data.

[0098] In one specific embodiment, refer to Figure 2 Based on prior knowledge, the following bands were selected from Sentinel-2 multispectral imagery: Band 3 (green band), Band 9 (red edge band), Band 10 (water vapor absorption band), Band 11 (shortwave infrared 1 band), and Band 12 (shortwave infrared 2 band) to establish extraction rules. In Band 3, pixels greater than 800 are assigned a value of 1, otherwise 0. This results in a 1022x1256 matrix x1. In Band 9, pixels less than 1500 are assigned a value of 1, otherwise 0. This results in a 1022x1256 matrix x2. In Band 10, pixels less than 5000 are assigned a value of 1, otherwise 0. This results in a 1022x1256 matrix x3. In Band 11, pixels less than 6000 are assigned a value of 1, otherwise 0. This results in a 1022x1256 matrix x4. In band 12, pixels greater than 1200 are equal to 1, otherwise equal to 0. This results in a matrix of size 1022x1256 x 5. Specifically, this can be represented as:

[0099]

[0100] In the formula, x1 represents the initial layering data, and band3 represents the third band of the Sentinel-2 image. The prior knowledge corresponding to the 3rd band is set to 800.

[0101]

[0102] In the formula, x2 represents the initial layering data, and band9 is the 9th band of the Sentinel-2 image. The prior knowledge corresponding to band 9 is set to 1500.

[0103]

[0104] In the formula, x3 represents the initial layering data, and band10 represents the 10th band of the Sentinel-2 image. The prior knowledge corresponding to band 10 is set to 5000.

[0105]

[0106] In the formula, x4 represents the initial layering data, and band11 represents the 11th band of the Sentinel-2 image. The prior knowledge corresponding to band 11 is set to 6000.

[0107]

[0108] In the formula, x5 represents the initial layering data, and band12 represents the 12th band of the Sentinel-2 image. The prior knowledge corresponding to band 12 is set to 1200. After the above layering, five layers of preliminary layered data are obtained, which are the layered data of pixels in the same region in different bands, namely x1, x2, x3, x4, and x5.

[0109] S600, integrate the aforementioned preliminary layered data to obtain integrated data.

[0110] Specifically, the acquired preliminary layered data are integrated to obtain the superimposed value of the preliminary layered data of the same area in several preset bands, which facilitates a clearer analysis of the attributes and land cover types of pixel areas.

[0111] In one specific embodiment, the initial hierarchical data is integrated to obtain a matrix of size 1022x1256 with values ​​ranging from 0 to 5:

[0112] Y1 = x1 + x2 + x3 + x4 + x5

[0113] In the formula, Y1 is the integrated data, with a data size ranging from 0 to 5; x1, x2, x3, x4, and x5 are all preliminary stratified data.

[0114] S700. Based on the multispectral image data, select vegetation bands to calculate the normalized vegetation index.

[0115] Specifically, the normalized vegetation index (NDI) is calculated by selecting the red band and near-infrared band from multispectral image data. The specific selection can be made in advance using the selection model mentioned above. After selection, the NDI is calculated by acquiring the surface reflectance data of the red band and the near-infrared band and using the preset calculation rules.

[0116] S800. Corrected integrated data is obtained by correcting the integrated data based on the normalized vegetation index.

[0117] Specifically, the integrated data is corrected by combining it with the normalized vegetation index (NVI) to obtain the corrected integrated data. The NVI is calculated based on the red and near-infrared bands in the multispectral image data.

[0118] Optionally, the step of selecting vegetation bands based on the multispectral image data to calculate the normalized vegetation index, and then correcting the integrated data based on the normalized vegetation index to obtain corrected integrated data, specifically includes:

[0119] S710. Obtain the red band and near-infrared band from the multispectral image data;

[0120] S720. Calculate the normalized vegetation index based on the red band and the near-infrared band.

[0121] S730. The intermediate parameters are calculated based on the normalized vegetation index of the pixel and the preset intermediate parameter calculation formula.

[0122] S740. The intermediate parameters are used to correct the integrated data to obtain corrected integrated data.

[0123] Specifically, the red and near-infrared bands are extracted from the multispectral image data beforehand. Then, the normalized vegetation index (NVI) is calculated for the pixels in the red and near-infrared bands. Intermediate parameters are obtained based on the NVI of the pixels, and these intermediate parameters are used to correct and integrate the data, resulting in corrected and integrated data. The specific calculation formula is shown below:

[0124] The Normalized Difference Vegetation Index (NDVI) is calculated using the following formula:

[0125]

[0126] In the formula, band8 is the 8th band (near-infrared band) of the sentinel-2 image; band4 is the 4th band (red band) of the sentinel-2 image.

[0127] In the NDVI matrix, we set pixels greater than 0.3 to 6 and pixels less than 0.3 to 0, resulting in a matrix of intermediate parameters:

[0128]

[0129] In the formula, Y2 is an intermediate parameter, which also represents the range of vegetation.

[0130] The process of using NDVI to correct the integrated data can be represented as follows:

[0131] Y = Y1 + Y2

[0132] In the formula, Y is the final stratified data, and its value ranges from 0 to 11.

[0133] S900. Based on the corrected and integrated data and the preset classification rules, the distribution results of multiple land cover types are obtained.

[0134] Specifically, the final score data is stratified and assigned values. Pixels with Y values ​​ranging from 1 to 3 represent water bodies; pixels with Y values ​​of 1 and 2 represent inland water bodies, and pixels with Y value of 3 represent ocean water bodies. Pixels with Y values ​​ranging from 4 to 5 represent impermeable surfaces; pixels with Y value of 4 represent bare soil and concrete roads, and pixels with Y value of 5 represent buildings. Pixels with Y values ​​ranging from 6 to 11 represent green spaces. After these steps, the distribution results of green spaces, water bodies, and impermeable surfaces (buildings, bare soil, and concrete roads) are obtained, as follows: Figure 4 As shown. In Figure 4 In the diagram, brown represents green areas, blue represents water bodies, light blue represents buildings, and yellow represents bare soil and concrete roads.

[0135] The implementation of this invention provides the following beneficial effects: This invention provides a method for extracting urban green space based on prior knowledge and remote sensing data, comprising: acquiring remote sensing data; preprocessing the remote sensing data to obtain multispectral image data; selecting several preset bands based on the multispectral image data; setting prior knowledge thresholds corresponding to each preset band based on the preset bands and prior knowledge; comparing the pixels of each preset band with the corresponding prior knowledge threshold to obtain a comparison result; initializing the pixels to a first value or a second value according to the comparison result; obtaining several preliminary layered data after initialization, wherein one preset band corresponds to one preliminary layered data, and one pixel has a corresponding first value or second value in different initial layered data; integrating the several preliminary layered data to obtain integrated data; selecting vegetation bands based on the multispectral image data to calculate the normalized vegetation index (NVI); correcting the integrated data based on the NVI to obtain corrected integrated data; and obtaining the distribution results of multiple land cover types according to the corrected integrated data and preset classification rules. After acquiring remote sensing data and extracting the corresponding bands, prior knowledge is used to obtain the prior thresholds for pixels in the corresponding bands. The pixels are then initialized to a first or second value, resulting in preliminary stratification data for pixels in different bands. This preliminary stratification data is then integrated and corrected using the normalized vegetation index (NVI) to obtain corrected and integrated data. Finally, based on preset classification rules, the distribution results of multiple land cover types are obtained. While extracting urban green spaces, other urban land cover types are automatically stratified, effectively distinguishing the scope of urban green spaces. Furthermore, the results extracted based on prior knowledge are of high quality and more robust, enabling rapid urban green space delineation and reducing workload.

[0136] By fully leveraging the advantages of multispectral remote sensing data, the process of distinguishing between green spaces, water bodies, and impermeable surfaces has been simplified. This ensures the complete extraction of green spaces while also extracting additional information on water bodies and impermeable surfaces. It can be applied to the rapid extraction of large-scale urban green spaces in engineering projects, reducing manual workload, improving extraction efficiency, and ensuring the accuracy of subsequent calculations of urban greening rates.

[0137] This method is helpful in assessing the distribution, accessibility, and service quality of green spaces in cities, thereby providing urban planners and policymakers with recommendations on green space protection, increasing green space area, and improving green space quality. The method is simple in principle, robust, and easier to promote and apply.

[0138] like Figure 5 As shown, this embodiment of the invention also provides an urban green space extraction system based on prior knowledge and remote sensing data, comprising:

[0139] The first module is used to acquire remote sensing data and preprocess the remote sensing data to obtain multispectral image data.

[0140] The second module is used to select several preset bands based on the multispectral image data;

[0141] The third module is used to set a threshold for prior knowledge corresponding to each of the preset bands based on several preset bands and prior knowledge.

[0142] The fourth module is used to compare the pixel of each preset band with the corresponding prior knowledge threshold to obtain the comparison result;

[0143] The fifth module is used to initialize the pixel to a first value or a second value according to the comparison result, and after initialization, a number of preliminary layer data are obtained. Among them, one preset band corresponds to one preliminary layer data, and one pixel has a corresponding first value or second value in different initial layer data.

[0144] The sixth module is used to integrate the aforementioned preliminary hierarchical data to obtain integrated data;

[0145] The seventh module is used to select vegetation bands based on the multispectral image data to calculate the normalized vegetation index.

[0146] The eighth module is used to correct the integrated data based on the normalized vegetation index to obtain corrected integrated data.

[0147] The ninth module is used to obtain the distribution results of multiple land cover types based on the corrected and integrated data and the preset classification rules.

[0148] It is evident that the content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0149] like Figure 6 As shown, this embodiment of the invention also provides an urban green space extraction device based on prior knowledge and remote sensing data, comprising:

[0150] At least one processor;

[0151] At least one memory for storing at least one program;

[0152] When the at least one program is executed by the at least one processor, the at least one processor performs the method steps described in the above method embodiments.

[0153] It is evident that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented in this device embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0154] Furthermore, this application also discloses a computer program product or computer program stored in a computer-readable storage medium. A processor of a computer device can read the computer program from the computer-readable storage medium, and the processor executes the computer program, causing the computer device to perform the described method. Similarly, the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0155] It is understood that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital information processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data information such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0156] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A method for extracting urban green space based on prior knowledge and remote sensing data, characterized in that, The method comprises: acquiring remote sensing data; preprocessing the remote sensing data to obtain multispectral image data; selecting a plurality of preset wavebands based on the multispectral image data, specifically comprising: inputting the multispectral image data into a preset selection model; the selection model selects a green waveband, a red edge waveband, a water vapor absorption waveband, a short-wave infrared 1 waveband, and a short-wave infrared 2 waveband from the multispectral image data; setting a priori knowledge threshold value corresponding to each of the plurality of preset wavebands based on the plurality of preset wavebands and priori knowledge; comparing each pixel point of the preset waveband with the corresponding priori knowledge threshold value to obtain a comparison result; initializing the pixel point to a first value or a second value according to the comparison result, and obtaining a plurality of preliminary layering data after initialization, wherein one of the preset wavebands corresponds to one of the preliminary layering data, and one of the pixel points has a corresponding first value or second value in different preliminary layering data; integrating the plurality of preliminary layering data to obtain integrated data; selecting a vegetation waveband based on the multispectral image data to calculate a normalized vegetation index; correcting the integrated data based on the normalized vegetation index to obtain corrected integrated data; obtaining the distribution results of a plurality of ground object types according to the corrected integrated data and a preset classification rule.

2. The method of claim 1, wherein, The method further comprises: acquiring ground surface reflectance data of the pixel point in the plurality of preset wavebands; comparing the ground surface reflectance data of the pixel point in the same preset waveband with the priori knowledge threshold value; comparing the ground surface reflectance data of the pixel point in the same preset waveband with the priori knowledge threshold value; initializing the pixel point to a first value or a second value according to the comparison result and a preset comparison rule.

3. The method of claim 2, wherein, The method further comprises: respectively acquiring a comparison result of the pixel point in the green waveband and a first prior threshold value, a comparison result of the pixel point in the red edge waveband and a second prior threshold value, a comparison result of the pixel point in the water vapor absorption waveband and a third prior threshold value, a comparison result of the pixel point in the short-wave infrared 1 waveband and a fourth prior threshold value, and a comparison result of the pixel point in the short-wave infrared 2 waveband and a fifth prior threshold value; initializing the pixel point greater than the first prior threshold value to a first value, and initializing the pixel point less than the first prior threshold value to a second value to obtain a first preliminary layering data; initializing the pixel point greater than the second prior threshold value to a second value, and initializing the pixel point less than the second prior threshold value to a first value to obtain a second preliminary layering data; initializing the pixel point greater than the third prior threshold value to a second value, and initializing the pixel point less than the third prior threshold value to a first value to obtain a third preliminary layering data; initializing the pixel point greater than the fourth prior threshold value as a second value, and initializing the pixel point less than the fourth prior threshold value as a first value to obtain fourth preliminary hierarchical data; initializing the pixel point greater than the fifth prior threshold value as a first value, and initializing the pixel point less than the fifth prior threshold value as a second value to obtain fifth preliminary hierarchical data.

4. The method of claim 1, wherein, The prior knowledge threshold value corresponding to each of the preset wave bands is set based on the preset wave bands and prior knowledge, and specifically includes: A prior knowledge threshold value table is generated based on the preset wave bands and prior knowledge; The prior knowledge threshold value corresponding to each of the preset wave bands is set according to the prior knowledge threshold value table.

5. The method of claim 1, wherein, The preliminary hierarchical data are integrated to obtain integrated data, and specifically includes: The preliminary hierarchical data are superimposed to obtain the integrated data, and the superimposition formula is as follows: In the formula, is the integrated data, and the data size ranges from 0 to 5; , are all the preliminary hierarchical data.

6. The method of claim 1, wherein, The normalized vegetation index is calculated based on the vegetation wave band selected from the multispectral image data, and the integrated data is modified based on the normalized vegetation index to obtain modified integrated data, and specifically includes: The red wave band and the near-infrared wave band in the multispectral image data are acquired; The normalized vegetation index is calculated according to the red wave band and the near-infrared wave band; An intermediate parameter is calculated according to the normalized vegetation index of the pixel point and a preset intermediate parameter calculation formula; The integrated data is modified by the intermediate parameter to obtain modified integrated data.

7. A system for urban green space extraction based on prior knowledge and remote sensing data, characterized in that, It includes: A first module is configured to acquire remote sensing data, and pre-process the remote sensing data to obtain multispectral image data; A second module is configured to select a plurality of preset wave bands based on the multispectral image data, and specifically includes: inputting the multispectral image data into a preset selection model; the selection model selects a green wave band, a red edge wave band, a water vapor absorption wave band, a short-wave infrared 1 wave band and a short-wave infrared 2 wave band in the multispectral image data; A third module is configured to set a prior knowledge threshold value corresponding to each of the preset wave bands based on the preset wave bands and prior knowledge; A fourth module is configured to compare the pixel point of each of the preset wave bands with the corresponding prior knowledge threshold value to obtain a comparison result; A fifth module is configured to initialize the pixel point as a first value or a second value according to the comparison result, and obtain a plurality of preliminary hierarchical data after initialization, wherein one of the preset wave bands corresponds to one of the preliminary hierarchical data, and one of the pixel points has a corresponding first value or second value in different preliminary hierarchical data; A sixth module is configured to integrate the preliminary hierarchical data to obtain integrated data; A seventh module is configured to select a vegetation wave band based on the multispectral image data to calculate a normalized vegetation index; An eighth module is configured to modify the integrated data based on the normalized vegetation index to obtain modified integrated data; A ninth module is configured to obtain the distribution result of a plurality of ground object types according to the modified integrated data and a preset classification rule. 8.A device for extracting urban green space based on prior knowledge and remote sensing data, characterized in that, It includes: At least one processor; at least one memory storing at least one program; when the at least one program is executed by the at least one processor, the at least one processor is caused to implement the method according to any one of claims 1-6.

9. A computer readable storage medium having stored therein a program that is executable by a processor, characterized in that, The program executable by the processor for performing the method according to any one of claims 1-6 when the program is executed by the processor.

Images