An urban planning decision-making support system based on remote sensing mapping

Abstract

The present invention provides an urban planning decision-making support system based on remote sensing mapping, which relates to the field of data processing technology and includes: a decision-making module for constructing a three-dimensional virtual model of a city according to measurement and analysis results; in the three-dimensional virtual model of the city, corresponding parameters are set for various types of objects; an initial population of planning schemes is generated, and each planning scheme in the population of planning schemes corresponds to a set of parameter combinations in the three-dimensional virtual model of the city; a fitness function for evaluating the pros and cons of each planning scheme is defined; according to the fitness function, the corresponding planning scheme is determined from the current population; and crossover and mutation operations are performed on the planning schemes to generate new planning schemes until a preset number of iterations is reached to obtain the final urban planning scheme. The present invention shortens the decision-making cycle of urban planning and improves decision-making efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an urban planning decision-making support system based on remote sensing mapping. Background Art

[0002] Remote sensing data has the characteristics of wide coverage, large amount of information, and short update cycle, and can reflect the changes of urban surface information in real time and accurately. However, how to extract useful urban land feature information from massive remote sensing data and carry out scientific and reasonable urban planning based on this information is the main challenge currently faced.

[0003] In order to solve this problem, the combination of remote sensing mapping technology and computer-aided decision-making systems has been explored to achieve intelligent management of urban planning. However, most of the existing urban planning decision-making systems have the following problems:

[0004] For example, the data processing capabilities are insufficient and large-scale remote sensing data cannot be processed efficiently; the information extraction accuracy is not high, resulting in deviations between the planning scheme and the actual situation; the decision-making algorithm is not intelligent enough and it is difficult to comprehensively consider multiple planning goals and give the optimal solution. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide an urban planning decision-making auxiliary system based on remote sensing mapping, which shortens the decision-making cycle of urban planning and improves the decision-making efficiency.

[0006] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0007] In a first aspect, an urban planning decision support system based on remote sensing mapping comprises:

[0008] A data acquisition module, used to acquire remote sensing data of urban areas, wherein the remote sensing data includes urban surface information;

[0009] A data preprocessing module, used for preprocessing the remote sensing data to obtain preprocessed remote sensing data;

[0010] An information extraction module is used to extract urban feature data based on pre-processed remote sensing data;

[0011] The measurement and analysis module is used to measure and analyze urban feature data, as well as detect and measure changes in remote sensing data based on time series to generate measurement and analysis results;

[0012] The decision-making module is used to construct a three-dimensional virtual model of the city based on the measurement and analysis results; in the three-dimensional virtual model of the city, corresponding parameters are set for various types of land features; an initial population of planning schemes is generated, and each planning scheme in the planning scheme population corresponds to a set of parameter combinations in the three-dimensional virtual model of the city; a fitness function for evaluating the pros and cons of each planning scheme is defined; according to the fitness function, the corresponding planning scheme is determined from the current population; crossover and mutation operations are performed on the planning schemes to generate new planning schemes, until a preset number of iterations is reached to obtain a final urban planning scheme.

[0013] Furthermore, urban feature data is measured and analyzed, and remote sensing data change detection and measurement based on time series are performed to generate measurement and analysis results, including:

[0014] Using the spectrum and texture characteristics of remote sensing data, the remote sensing data are divided into different types of objects to generate object classification data;

[0015] According to the classification data of land objects, the number of pixels of each type of land object in the image is calculated; according to the spatial resolution information of remote sensing data, the number of pixels is converted into the actual surface area to obtain the actual area of ​​each type of land object;

[0016] Utilize the relevant height data in the remote sensing data to extract the height information of various objects; Based on the height information of various objects and the spatial position of the objects, calculate the distance between different objects;

[0017] For three-dimensional features, calculate their volume by combining their actual area and height information;

[0018] Perform pixel-by-pixel difference calculation on the preprocessed multi-period remote sensing data to generate difference data; convert the difference data into binary change data, in which the changed area is marked as a specific value and the unchanged area is marked as another value;

[0019] Based on the binary change data, the change information of various types of land features is extracted and counted, including the area of ​​the change area and the type of change; based on the land feature classification data and the binary change data, the measurement and analysis results are generated.

[0020] Furthermore, according to the object classification data, the number of pixels of each type of object in the image is calculated, including:

[0021] For each type of ground feature, traverse the entire classification image and count the number of pixels of each type, including: Suppose the classification image is , the threshold is , then each type Number of pixels The formula is:

[0022] ;

[0023] in, Indicates the type of feature The number of pixels; Represents the total pixel set of the image; , and is the weight coefficient; is the average spectral value of adjacent pixels; represents the indicator function, when hour, , otherwise 0; Indicates the type of feature The weight of Represents a function representing the spectral value of a pixel Is it greater than the set threshold? , the formula is defined as:

[0024] .

[0025] Furthermore, the preprocessed multi-period remote sensing data is subjected to pixel-by-pixel difference calculation to generate difference data, including:

[0026] pass The preprocessed multi-period remote sensing data is subjected to pixel-by-pixel difference calculation to generate difference data; Indicates at time Location The Band spectral value; Indicates at time Location The Band spectral value; Indicates at location The Band difference value; represents the spatial weight factor; represents the weight coefficient; Indicates the number of bands, Indicates the index of different bands; among them:

[0027] ;

[0028] in, Indicates location With neighboring pixels The Euclidean distance between .

[0029] Furthermore, the calculation formula of the fitness function is:

[0030] ;

[0031] in, Indicates Area of ​​land use type; Indicates Floor area ratio of the land use type; represents the unused land area; represents weight; represents the summation symbol; Indicates The service area of ​​each transportation node; Indicates The traffic capacity of each traffic node; represents the average travel distance in the planning area; It represents the maximum possible travel distance in the planning area; Indicates The number of public facilities of this type; Indicates The area of ​​the area where the public facilities are located; Indicates the number of types of public facilities; , and is the weight coefficient.

[0032] Furthermore, the planning schemes are crossover and mutation operations are performed to generate new planning schemes, including:

[0033] Encode the planning scheme into binary code;

[0034] Generate an initial population of planning solutions, which includes multiple randomly generated planning solutions, each of which is a point in the solution space;

[0035] By evaluating the fitness value of each planning scheme, the corresponding parent scheme is selected from the current population;

[0036] Select two parent solutions A and B and determine an intersection point or intersection area;

[0037] Exchange the genetic information of the parent plans A and B after the crossover point or crossover area to generate two new offspring plans C and D. Offspring plan C contains the first part of the genetic information of the parent A and the last part of the genetic information of the parent B. Select an offspring plan C;

[0038] Determine one or more mutation points, flip the bits in the binary code to change the value of the mutation point to generate a new mutated plan, and repeat the selection, crossover and mutation operations to generate a new planning plan.

[0039] Furthermore, urban feature data include spatial distribution and attribute information of buildings, roads, green spaces, and water bodies.

[0040] In a second aspect, a method for assisting decision-making in urban planning based on remote sensing mapping is provided, the method comprising:

[0041] Acquiring remote sensing data of an urban area, wherein the remote sensing data includes urban surface information;

[0042] Preprocessing the remote sensing data to obtain preprocessed remote sensing data;

[0043] Extract urban feature data based on pre-processed remote sensing data;

[0044] Conduct measurement and analysis of urban feature data, as well as remote sensing data change detection and measurement based on time series to generate measurement and analysis results;

[0045] A three-dimensional virtual model of the city is constructed according to the measurement and analysis results; corresponding parameters are set for various types of land features in the three-dimensional virtual model of the city; an initial population of planning schemes is generated, and each planning scheme in the population of planning schemes corresponds to a set of parameter combinations in the three-dimensional virtual model of the city; a fitness function for evaluating the pros and cons of each planning scheme is defined; according to the fitness function, the corresponding planning scheme is determined from the current population; crossover and mutation operations are performed on the planning schemes to generate new planning schemes, until a preset number of iterations is reached to obtain a final urban planning scheme.

[0046] According to a third aspect, a computing device includes:

[0047] one or more processors;

[0048] The storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method described.

[0049] In a fourth aspect, a computer-readable storage medium stores a program, and when the program is executed by a processor, the method described is implemented.

[0050] The above scheme of the present invention includes at least the following beneficial effects.

[0051] The data acquisition module can quickly and comprehensively acquire remote sensing data of urban areas, including high-resolution urban surface information. The data preprocessing module effectively preprocesses remote sensing data, such as denoising, correction, and registration, significantly improving the accuracy and availability of data. The information extraction module uses image processing and pattern recognition technology to accurately extract urban land object data, such as buildings, roads, green spaces, water bodies, etc., from preprocessed remote sensing data, providing detailed land object information for urban planning.

[0052] The measurement and analysis module can conduct detailed measurement and analysis of urban land object data; the decision-making module realizes the intuitive display and simulation analysis of the urban spatial structure by constructing a three-dimensional virtual model of the city, and sets corresponding parameters for various types of land objects in the model, making the planning scheme more in line with the actual situation.

[0053] By using intelligent optimization technologies such as genetic algorithms, the initial population of planning schemes is generated and the fitness function is defined to evaluate the pros and cons of each planning scheme. In the iterative process, the planning schemes are crossover and mutated to continuously generate new and better planning schemes until the preset number of iterations is reached, thus obtaining the final urban planning scheme. This method can comprehensively consider multiple planning objectives and give a near-optimal solution, greatly shortening the decision-making cycle of urban planning and improving decision-making efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a schematic diagram of an urban planning decision support system based on remote sensing mapping provided by an embodiment of the present invention.

[0055] Figure 2 The present invention provides a flowchart of a method for assisting decision-making in urban planning based on remote sensing mapping. DETAILED DESCRIPTION

[0056] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0057] like Figure 1 As shown, an embodiment of the present invention proposes an urban planning decision support system based on remote sensing mapping, comprising:

[0058] A data acquisition module, used to acquire remote sensing data of urban areas, wherein the remote sensing data includes urban surface information;

[0059] A data preprocessing module, used for preprocessing the remote sensing data to obtain preprocessed remote sensing data;

[0060] An information extraction module is used to extract urban feature data based on the pre-processed remote sensing data. The urban feature data includes the spatial distribution and attribute information of buildings, roads, green spaces, and water bodies.

[0061] The measurement and analysis module is used to measure and analyze urban feature data, as well as detect and measure changes in remote sensing data based on time series to generate measurement and analysis results;

[0062] The decision-making module is used to construct a three-dimensional virtual model of the city based on the measurement and analysis results; in the three-dimensional virtual model of the city, corresponding parameters are set for various types of land features; an initial population of planning schemes is generated, and each planning scheme in the planning scheme population corresponds to a set of parameter combinations in the three-dimensional virtual model of the city; a fitness function for evaluating the pros and cons of each planning scheme is defined; according to the fitness function, the corresponding planning scheme is determined from the current population; crossover and mutation operations are performed on the planning schemes to generate new planning schemes, until a preset number of iterations is reached to obtain a final urban planning scheme.

[0063] In an embodiment of the present invention, the data acquisition module can quickly and comprehensively acquire remote sensing data of urban areas, including high-resolution urban surface information, so that planners can more accurately understand the current situation and development trend of the city. The data preprocessing module preprocesses the remote sensing data, such as denoising, correction, and registration, and the module significantly improves the accuracy and availability of the data. The preprocessed remote sensing data is clearer and more accurate, and can more realistically reflect the urban surface information, laying a solid foundation for subsequent information extraction and measurement analysis, and reducing errors and uncertainties caused by data quality problems. The information extraction module uses image processing and pattern recognition technology to accurately extract urban land object data from the preprocessed remote sensing data, including the spatial distribution and attribute information of categories such as buildings, roads, green spaces, and water bodies, so that planners can more accurately understand the spatial structure of the city and the distribution of various types of land objects. The measurement and analysis module can perform detailed measurement and analysis on urban land object data, including the area, perimeter, shape and other characteristics of the land objects, as well as the spatial relationship between the land objects. At the same time, based on the detection and measurement of changes in remote sensing data in time series, this module can also reveal the dynamic trends and laws of urban development, such as urban expansion and reduction of green space, so that planners can grasp the direction and speed of urban development more accurately.

[0064] The decision-making module realizes the intuitive display and simulation analysis of the urban spatial structure by constructing a three-dimensional virtual model of the city. In the model, corresponding parameters are set for various types of landforms to make the planning scheme more in line with the actual situation. By generating an initial population of planning schemes and defining a fitness function to evaluate the pros and cons of each planning scheme, the module can comprehensively consider multiple planning goals and give an approximately optimal solution. In the iterative process, the planning schemes are cross-cut and mutated to continuously generate new and better planning schemes until the preset number of iterations is reached, thereby obtaining the final urban planning scheme. This method significantly improves the scientificity and rationality of urban planning, making the planning scheme more in line with the development needs and actual situation of the city. At the same time, it also greatly shortens the decision-making cycle of urban planning and improves the efficiency and quality of decision-making.

[0065] In another preferred embodiment of the present invention, the specific implementation process of the data acquisition module is as follows:

[0066] According to the specific needs of urban planning, determine the type, resolution, phase and other key parameters of the required remote sensing data, including the scope of the urban area to be acquired, the time range of the data, and the level of detail of the data. According to data requirements, select appropriate remote sensing data sources, including satellite remote sensing data (such as Landsat, Sentinel-2, etc.), UAV remote sensing data or aerial photography data, etc., and consider factors such as data availability, cost, resolution and coverage when selecting.

[0067] Receive remote sensing data transmitted from the data source through receiving equipment or network platform. The received data is format converted, compressed and stored in a secure storage medium. Perform quality checks on the received remote sensing data, including checking the integrity, clarity, resolution, etc. of the data. Re-acquire data with quality problems or take other remedial measures.

[0068] In another preferred embodiment of the present invention, the specific implementation process of the data preprocessing module is as follows:

[0069] The remote sensing data is geometrically corrected to eliminate image distortion and deformation caused by factors such as sensors, platforms or atmosphere. This includes operations such as image registration and projection conversion to ensure the spatial accuracy of the data. Filtering and smoothing algorithms are used to remove noise and interference information in remote sensing data to improve the signal-to-noise ratio and clarity of the data. The remote sensing data is radiometrically corrected to eliminate radiation errors caused by factors such as sensor response and solar radiation angle. This includes operations such as radiation calibration and atmospheric correction to ensure the radiation accuracy of the data. According to the specific needs of urban planning, the remote sensing data is cropped to retain only the urban area of ​​interest. At the same time, if the data is composed of multiple images, image stitching processing is also required to ensure the integrity and continuity of the data. The pre-processed remote sensing data is converted into a format suitable for subsequent information extraction and processing, such as TIFF, JPEG, PNG, etc.

[0070] In another preferred embodiment of the present invention, the specific implementation process of the information extraction module is as follows:

[0071] Load the preprocessed remote sensing data, which is a multi-band (or multi-channel) image containing different spectral information of the urban surface; select specific bands for segmentation or perform band combination as needed to enhance the contrast between the target objects (such as buildings, roads, green spaces, water bodies, etc.) and the background; for each selected band or band combination, analyze its histogram, that is, the distribution of pixel values ​​(or brightness values), and the histogram shows the number of pixels with different brightness values ​​in the image; select a threshold value based on the spectral characteristics of the object and the distribution of the histogram, which is a single value (for binary segmentation); apply the threshold value to the remote sensing data, compare each pixel, and if the pixel value is higher (or lower) than the threshold value, classify it as If a certain area is classified as a target feature, it is classified as the background; otherwise, it is classified as the background; small isolated areas or noise points are removed, which can be achieved by setting the minimum area size or applying morphological operations (such as dilation, erosion, opening operation, closing operation, etc.); the segmentation boundaries are smoothed to reduce the jagged edges caused by pixel-level classification, which can be achieved by applying edge smoothing algorithms; the accuracy of the segmentation results is verified by visual inspection or using reference data (such as field survey data, high-resolution images, etc.); based on the verification results, the threshold selection and post-processing steps are adjusted to improve the segmentation quality, which requires multiple iterations and experiments to obtain the best segmentation effect; the segmentation results are saved in an appropriate format (such as vector layer, classified raster, etc.).

[0072] Load an image processed by a threshold segmentation algorithm, which has segmented the urban surface into different areas or objects, such as buildings, roads, green spaces, water bodies, etc.

[0073] Determine the feature type, including:

[0074] Shape features: including the boundary shape, aspect ratio, area, perimeter, etc. of the feature.

[0075] Size feature: refers to the area or volume of the object, which can be calculated based on the number of pixels or actual size.

[0076] Texture features: describe the details and patterns of the surface of objects, such as roughness, smoothness, contrast, etc.

[0077] Color features: For color images, the average color, color histogram, color distribution, etc. of the objects can be extracted.

[0078] Use boundary tracing algorithms (such as Moore-Neighbor Tracing Algorithm) to identify the boundaries of features and then calculate the perimeter and shape parameters. For closed areas, use region filling algorithms to determine the area of ​​features and calculate shape descriptors such as aspect ratio and compactness (ratio of area to perimeter). Calculate the number of pixels that make up the feature, which can be used as an approximate indicator of size. If the geographic reference information of the image (such as resolution and scale) is known, the number of pixels can be converted into actual area or volume.

[0079] Calculate the gray-level co-occurrence matrix of the image and extract texture features such as contrast, energy, homogeneity, etc.; apply wavelet transform to decompose the image and extract high-frequency and low-frequency components to describe the details and overall structure of the texture; use the Laws texture filter to convolve the image and extract texture energy features; count the number of pixels of different colors in the image to form a color histogram, and calculate the low-order moments of the color (such as mean, variance, skewness) to describe the distribution of color; convert the image from RGB color space to other color spaces (such as HSV, Lab) to extract more distinctive color features.

[0080] In a preferred embodiment of the present invention, urban feature data are measured and analyzed, and remote sensing data change detection and measurement based on time series are performed to generate measurement and analysis results, including:

[0081] Using the spectral and texture characteristics of remote sensing data, the remote sensing data is divided into different types of objects to generate object classification data, including:

[0082] Acquire remote sensing data from satellites, drones or other remote sensing platforms. These data contain multiple bands (such as visible light, infrared, thermal infrared, etc.) and are stored in raster or image format; perform geometric correction on remote sensing data to eliminate image deformation caused by factors such as sensors, platforms or the curvature of the earth's surface; perform radiation correction to eliminate image brightness distortion caused by factors such as atmospheric absorption, scattering or sensor response; crop remote sensing data according to the scope of the study area to reduce the amount of data and focus on the area of ​​interest. If the data has different spatial resolutions, resampling is required so that all bands have the same pixel size and alignment; select bands from remote sensing data that can reflect the differences in ground objects, including visible light bands (used to distinguish vegetation, water bodies and soil, etc.), infrared bands (used to detect thermal radiation and vegetation health), etc.; calculate some spectral indices, such as the normalized difference vegetation index ( NDVI), water index (WI), etc., to enhance the spectral differences between objects; use texture analysis methods (such as gray level co-occurrence matrix) to extract texture features of remote sensing data, which can reflect the surface roughness, repetitive patterns, etc. of the objects; select an appropriate unsupervised classification algorithm, such as the K-means clustering algorithm, and set initial parameters for the K-means clustering algorithm, such as the number of clusters (K value), number of iterations, convergence conditions, etc.; input the preprocessed remote sensing data and the extracted features into the unsupervised classification algorithm to perform classification operations. The algorithm will divide the pixels into different object categories based on the similarity of the data; analyze the sources of errors in the classification results, such as confusion between spectral and texture features, limitations of the classification algorithm, etc.; adjust the parameters of the classification algorithm according to the evaluation results to improve the classification accuracy; output the final object classification results in a raster data format (such as GeoTIFF), where the value of each pixel represents a different object category.

[0083] According to the classification data of the objects, the number of pixels of each type of objects in the image is calculated; according to the spatial resolution information of the remote sensing data, the number of pixels is converted into the actual surface area to obtain the actual area of ​​each type of objects, which specifically includes: determining the spatial resolution of the remote sensing data, which refers to the actual ground area represented by each pixel in the remote sensing image; the spatial resolution is measured in meters or feet, which represents the actual size of each pixel on the ground (such as 10 meters resolution means that each pixel represents an area of ​​10 meters × 10 meters on the ground); through the classification results of the remote sensing data, the number of pixels occupied by each type of object can be known, and these classification results can be obtained by the K-means clustering method (such as K-means clustering), in which each pixel is assigned a type of object label; according to the spatial resolution of the remote sensing data, the actual ground area represented by a single pixel can be calculated. For example, if the spatial resolution is 10 meters, the area represented by each pixel is 10 meters × 10 meters = 100 square meters; for each type of object, the number of pixels occupied by it is multiplied by the area of ​​a single pixel to obtain the actual area of ​​that type of object.

[0084] Utilize the relevant height data in the remote sensing data to extract the height information of various types of objects; Based on the height information of various types of objects, combined with the spatial position of the objects, calculate the distance between different objects, specifically including: the acquired remote sensing data contains height information, which comes from stereo image pairs (such as stereo images taken by satellites or drones), laser radar (LiDAR) data or digital surface models (DSM), etc.; perform necessary preprocessing on remote sensing data, including denoising, correction, and registration, to ensure data quality; If stereo image pairs are used, generate disparity maps through stereo matching algorithms, and then convert them into height maps; For LiDAR data, extract ground and elevation information through point cloud processing software, if already With DSM, directly extract the height information therein; spatially superimpose the feature classification data and height data previously obtained by the K-means clustering method; assign corresponding height values ​​to each feature category according to the classification results, and perform statistical analysis on the height information of each feature category, such as calculating the average height, maximum height, minimum height, etc.; place all features in the same coordinate system, and for each feature category or single feature, determine its center point or representative point as the basis for calculating the distance; in two-dimensional or three-dimensional space, use the Euclidean distance formula to calculate the distance between the center points of two features; traverse all feature categories or single features, calculate the distance between them, and generate a distance matrix.

[0085] For three-dimensional objects, their volume is calculated by combining their actual area and height information, specifically including: obtaining the actual area information of the object through classification and area calculation of remote sensing data; extracting the height information of the object through the height data in the remote sensing data (such as digital surface model DSM or laser radar LiDAR data); matching data so that the area information and height information are for the same object and in the same spatial reference system; for three-dimensional shapes (such as cylinders, cones, etc.), the volume can be calculated directly using geometric formulas; for objects with complex shapes, polyhedrons (such as triangulated networks) can be used to approximate their surfaces, and then the volume of the object can be estimated by calculating the volume of the polyhedron.

[0086] Perform pixel-by-pixel difference calculation on the preprocessed multi-period remote sensing data to generate difference data; convert the difference data into binary change data, in which the changed area is marked as a specific value and the unchanged area is marked as another value, specifically including: traversing each element in the difference data (such as each pixel in the image), comparing the element value with the threshold; if the element value is greater than or equal to the threshold, mark the element as a changed area (such as assigning a value of 1); if the element value is less than the threshold, mark the element as an unchanged area (such as assigning a value of 0); create a new array according to the size of the difference data to store the binarization result, and fill the corresponding position in the output array with 1 (changed) or 0 (unchanged) according to the comparison result of the previous step;

[0087] Based on the binary change data, extract and count the change information of various types of features, including the area and change type of the change area; generate measurement and analysis results based on the feature classification data and the binary change data, including: overlay analysis of the binary change data and the feature classification data, which can be completed in GIS software through raster calculation or spatial connection; for each change area (pixel marked as 1), find its corresponding feature type, record the feature type before and after the change to determine the change type (for example, from forest to city); if the data is in raster format, the area of ​​the change area can be obtained by calculating the number of pixels marked as changed and multiplying it by the pixel size; if the data is in vector format, the area calculation tool in the GIS software can be used directly; classify and count the change information according to the feature type, count the change area of ​​each feature type, count the area and number of each change type (such as from forest to city, from farmland to water, etc.), and summarize the statistical results into tables or charts for further analysis and visualization.

[0088] In the embodiment of the present invention, the K-means clustering method can automatically classify according to the characteristics of the data, reduce manual intervention, improve classification efficiency, and classify the objects in the entire remote sensing image, providing comprehensive object distribution information; through the conversion of the number of pixels and the spatial resolution, the accurate area of ​​each type of object is obtained, providing accurate data for urban planning, land use, etc.; the area data supports quantitative analysis of urban objects, such as the proportion of objects, the density of object distribution, etc.; the extraction of height information enables the three-dimensional spatial distribution of urban objects to be presented; the distance calculation between objects helps to understand the urban spatial structure, such as the distance between buildings, the width of roads, etc., providing a spatial reference for urban planning; the volume data provides an important basis for the calculation of the floor area ratio of urban buildings, the evaluation of building energy consumption, etc., and through volume calculation, the utilization efficiency and potential of the three-dimensional space of the city can be analyzed; through difference calculation and binarization processing, the changes of urban objects, such as new buildings, land cover changes, etc., can be accurately detected, and the binarization data simplifies the subsequent change information extraction and statistical analysis process. The measurement and analysis results integrate information on object classification, area, height, volume, change, etc., and provide comprehensive data support for urban planning and management.

[0089] In a preferred embodiment of the present invention, the number of pixels of each type of ground object in the image is calculated according to the ground object classification data, including:

[0090] For each type of ground feature, traverse the entire classification image and count the number of pixels of each type, including: Suppose the classification image is , the threshold is , then each type Number of pixels The formula is:

[0091] ;

[0092] in, Indicates the type of feature The number of pixels; Represents the total pixel set of the image; , and is the weight coefficient; is the average spectral value of adjacent pixels; represents the indicator function, when hour, , otherwise 0; Indicates the type of feature The weight of Represents a function representing the spectral value of a pixel Is it greater than the set threshold? , the formula is defined as:

[0093] .

[0094] In the embodiment of the present invention, the formula not only takes into account the ground object classification data , and also introduces the spectral value of the pixel ), the average spectral value of adjacent pixels And through the threshold function The spectral values ​​are screened. This multi-dimensional information fusion helps to improve the accuracy of ground object type pixel quantity statistics. , and This method can be adapted to different image data and classification requirements. By introducing the average spectral value of adjacent pixels This method takes into account the spatial relationship between pixels to a certain extent, which helps to reduce the impact of isolated pixels on statistical results and improve the robustness of statistics. The introduction of enables this method to be based on whether the spectral value of the pixel is greater than the set threshold To screen effective pixels, further improve the accuracy and pertinence of statistics.

[0095] In a preferred embodiment of the present invention, pixel-by-pixel difference calculation is performed on the preprocessed multi-period remote sensing data to generate difference data, including:

[0096] pass Perform pixel-by-pixel difference calculation on the preprocessed multi-period remote sensing data to generate difference data;

[0097] in, Indicates at time Location The Band spectral value; Indicates at time Location The Band spectral value; Indicates at location The Band difference value; represents the spatial weight factor; represents the weight coefficient; Indicates the number of bands, Indicates the index of different bands; among them:

[0098] ;

[0099] in, Indicates location With neighboring pixels The Euclidean distance between .

[0100] In the embodiment of the present invention, the embodiment not only considers the difference of spectral values ​​at different time points, but also introduces a spatial weight factor To adjust this difference. The spatial weight factor takes into account both the spatial distance and spectral difference between pixels, making the difference calculation more accurate and able to more accurately reflect the actual situation of the changes in the ground objects. Remote sensing images often have spatial heterogeneity, that is, pixels at different locations may be affected by different factors. By introducing the spatial distance This method can better adapt to the spatial heterogeneity and reduce the error caused by different spatial positions. The difference weights between different pixels can be adjusted dynamically, so that the difference calculation results are more significant in areas with large spectral changes, and relatively smooth in areas with small spectral changes.

[0101] In a preferred embodiment of the present invention, a three-dimensional virtual model of the city is constructed according to the measurement and analysis results; in the three-dimensional virtual model of the city, corresponding parameters are set for various types of land features; an initial population of planning schemes is generated, and each planning scheme in the population of planning schemes corresponds to a set of parameter combinations in the three-dimensional virtual model of the city, specifically including:

[0102] Collect the city's topographic maps, building floor plans, elevation data, and feature classification data; pre-process the data, including denoising, coordinate conversion, and data format unification; use 3D modeling software (such as CityEngine, SketchUp, 3ds Max, etc.) or programming libraries (such as Three.js, Cesium, etc.) to build a 3D model of the city; set corresponding models and textures for different types of features (such as roads, buildings, green spaces, etc.) based on the feature classification data; set adjustable parameters for various features in the city's 3D virtual model, such as the height, width, and color of buildings, the width of roads, and the coverage of green spaces; determine the range of possible values ​​for each feature parameter to form a parameter space, and randomly generate a set of parameter combinations in the parameter space as a planning scheme. Repeat the above process to generate multiple planning schemes to form an initial population of planning schemes.

[0103] In a preferred embodiment of the present invention, the calculation formula of the fitness function is:

[0104] ;

[0105] in, Indicates Area of ​​land use type; Indicates Floor area ratio of the land use type; represents the unused land area; represents weight; represents the summation symbol; Indicates The service area of ​​each transportation node; Indicates The traffic capacity of each traffic node; represents the average travel distance in the planning area; It represents the maximum possible travel distance in the planning area; Indicates The number of public facilities of this type; Indicates The area of ​​the area where the public facilities are located; Indicates the number of types of public facilities; , and is the weight coefficient.

[0106] In the embodiment of the present invention, the fitness function integrates multiple dimensions such as land use efficiency, traffic convenience, and public facilities distribution, and can comprehensively evaluate the pros and cons of planning schemes, thereby helping to formulate more scientific and reasonable urban planning schemes; The land use type and its floor area ratio are partially taken into consideration, and the area and weight of unused land are also taken into consideration, which helps to encourage efficient use of land resources, reduce waste, and optimize land allocation. Partially focusing on the service area and traffic capacity of transportation nodes, as well as the average travel distance within the planning area, will help improve the coverage and efficiency of the transportation network. Partial attention is paid to the distribution density of public facilities, which can ensure that various types of public facilities are evenly distributed in the planning area.

[0107] In a preferred embodiment of the present invention, crossover and mutation operations are performed on the planning scheme to generate a new planning scheme, including:

[0108] Encode the planning scheme into binary code;

[0109] Generate an initial population of planning solutions, which includes multiple randomly generated planning solutions, each of which is a point in the solution space;

[0110] By evaluating the fitness value of each planning scheme, the corresponding parent scheme is selected from the current population;

[0111] Select two parent solutions A and B and determine an intersection point or intersection area;

[0112] Exchange the genetic information of the parent plans A and B after the crossover point or crossover area to generate two new offspring plans C and D. Offspring plan C contains the first part of the genetic information of the parent A and the last part of the genetic information of the parent B. Select an offspring plan C;

[0113] Determine one or more mutation points, flip the bits in the binary code to change the value of the mutation point to generate a new mutated plan, and repeat the selection, crossover and mutation operations to generate a new planning plan.

[0114] In an embodiment of the present invention, by exchanging the genetic information of two parent solutions after the intersection or intersection area, a completely new offspring solution can be generated. These offspring solutions inherit the excellent features of the parents and introduce new combinations, which helps to explore different areas of the solution space. By randomly changing certain elements in the planning scheme (such as flipping bits in binary coding), a new solution that is completely different from the parent can be generated. The mutation operation increases the diversity of the population, giving the algorithm the opportunity to jump out of the current optimal solution and further explore other possible better solutions. The crossover and mutation operations in the genetic algorithm are equivalent to a "random search" in the solution space. This search mechanism helps the algorithm avoid falling into a local optimal solution, because even if the current optimal solution performs well in a local area, the crossover and mutation operations may still generate a better solution in the global scope. By selecting parent solutions with higher fitness for crossover and mutation, new solutions that are more likely to inherit excellent features can be generated. These new solutions often have higher fitness values, thereby accelerating the convergence process of the algorithm. The crossover and mutation operations increase the diversity of planning solutions by introducing new combinations and changes. This diversity is crucial to finding the global optimal solution because it allows the algorithm to search in a wider solution space. The crossover and mutation operations in genetic algorithms are performed automatically without human intervention, which enables the algorithm to automatically search the solution space in an intelligent way and find potential better solutions.

[0115] When applied specifically, the following steps are included:

[0116] According to the complexity and sophistication of the planning scheme, determine the length of the binary code, which should be long enough to represent all the key elements in the planning scheme; convert each element in the planning scheme (such as land use type, transportation nodes, public facilities, etc.) into binary form. For example, if a piece of land can be used in three ways (residential, commercial, green space), it can be represented by two binary numbers (such as 00, 01, 10); connect the binary codes of all elements in sequence to form a long binary code string, which represents the entire planning scheme.

[0117] According to the computing resources and time limit, the number of planning schemes in the initial population is determined; multiple planning schemes are randomly generated in the solution space, and each scheme corresponds to a binary code string. These schemes constitute the initial population; according to the goals and constraints of urban planning, a fitness function is defined to evaluate the pros and cons of planning schemes; the binary code string of each planning scheme is substituted into the fitness function to calculate its fitness value, which reflects the degree to which the planning scheme meets the goals and constraints of urban planning.

[0118] According to the size of the fitness value, the planning scheme with higher fitness is selected as the parent. The selection strategies include roulette selection and tournament selection. A certain number of parent schemes are selected from the initial population to form a parent population. Two planning schemes A and B are randomly selected from the parent population as the parents of the crossover.

[0119] A crossover point or crossover region is randomly determined in the binary code string of A and B, and the genetic information (i.e., binary code) of A and B after the crossover point or crossover region is exchanged to generate two new offspring solutions C and D. C contains the front part of A and the back part of B, and D is the opposite; one or more mutation points are randomly determined in the binary code string of offspring solution C; the binary bits at the mutation points are flipped (i.e., 0 becomes 1, 1 becomes 0) to generate a new mutated solution; the new solution generated by the crossover and mutation operations is added to the population to replace the old solution with lower fitness to form a new population, and the above selection, crossover, and mutation operations are repeated until the preset number of iterations is reached or other termination conditions are met (such as the fitness value reaches the preset threshold, the population diversity is lower than a certain level, etc.). After the iteration is completed, the planning solution with the highest fitness is selected from the final population as the optimal solution.

[0120] Through the implementation of the above steps, the genetic algorithm can continuously explore the solution space, generate new planning schemes, and gradually optimize the best solution that meets the urban planning goals and constraints.

[0121] like Figure 2 As shown, an embodiment of the present invention further provides an urban planning decision-making assistance method based on remote sensing mapping, comprising:

[0122] Acquiring remote sensing data of an urban area, wherein the remote sensing data includes urban surface information;

[0123] Preprocessing the remote sensing data to obtain preprocessed remote sensing data;

[0124] Extract urban feature data based on pre-processed remote sensing data;

[0125] Conduct measurement and analysis of urban feature data, as well as remote sensing data change detection and measurement based on time series to generate measurement and analysis results;

[0126] A three-dimensional virtual model of the city is constructed according to the measurement and analysis results; corresponding parameters are set for various types of land features in the three-dimensional virtual model of the city; an initial population of planning schemes is generated, and each planning scheme in the population of planning schemes corresponds to a set of parameter combinations in the three-dimensional virtual model of the city; a fitness function for evaluating the pros and cons of each planning scheme is defined; according to the fitness function, the corresponding planning scheme is determined from the current population; crossover and mutation operations are performed on the planning schemes to generate new planning schemes, until a preset number of iterations is reached to obtain a final urban planning scheme.

[0127] It should be noted that the system is a system corresponding to the above method, and all implementation methods in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0128] The embodiment of the present invention further provides a computing device, comprising: a processor, a memory storing a computer program, wherein when the computer program is executed by the processor, the method described above is executed. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0129] The embodiment of the present invention also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the method described above. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0130] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An urban planning decision support system based on remote sensing mapping, characterized in that: The system comprises: A data acquisition module, used to acquire remote sensing data of urban areas, wherein the remote sensing data includes urban surface information; A data preprocessing module, used for preprocessing the remote sensing data to obtain preprocessed remote sensing data; An information extraction module is used to extract urban feature data based on pre-processed remote sensing data; The measurement and analysis module is used to measure and analyze urban feature data, as well as detect and measure changes in remote sensing data based on time series to generate measurement and analysis results; The decision-making module is used to construct a three-dimensional virtual model of the city according to the measurement and analysis results; in the three-dimensional virtual model of the city, corresponding parameters are set for various types of land features; an initial population of planning schemes is generated, and each planning scheme in the population of planning schemes corresponds to a set of parameter combinations in the three-dimensional virtual model of the city; a fitness function for evaluating the pros and cons of each planning scheme is defined; according to the fitness function, the corresponding planning scheme is determined from the current population; crossover and mutation operations are performed on the planning schemes to generate new planning schemes, until a preset number of iterations is reached, so as to obtain the final urban planning scheme; Conduct measurement and analysis of urban feature data, as well as remote sensing data change detection and measurement based on time series to generate measurement and analysis results, including: Using the spectrum and texture characteristics of remote sensing data, the remote sensing data are divided into different types of objects to generate object classification data; According to the classification data of land objects, the number of pixels of each type of land object in the image is calculated; according to the spatial resolution information of remote sensing data, the number of pixels is converted into the actual surface area to obtain the actual area of ​​each type of land object; Utilize the relevant height data in the remote sensing data to extract the height information of various objects; Based on the height information of various objects and the spatial position of the objects, calculate the distance between different objects; For three-dimensional features, calculate their volume by combining their actual area and height information; Perform pixel-by-pixel difference calculation on the preprocessed multi-period remote sensing data to generate difference data; convert the difference data into binary change data, in which the changed area is marked as a specific value and the unchanged area is marked as another value; Based on the binary change data, extract and count the change information of various types of land features, including the area and type of change; generate measurement and analysis results based on the land feature classification data and the binary change data; According to the object classification data, the number of pixels of each type of object in the image is calculated, including: For each type of ground feature, traverse the entire classification image and count the number of pixels of each type, including: Suppose the classification image is , the threshold is , then each type Number of pixels The formula is: ； in, Indicates the type of feature The number of pixels; Represents the total pixel set of the image; is the weight coefficient; is the average spectral value of adjacent pixels; represents the indicator function, when hour, , otherwise 0; Indicates the type of feature The weight of Represents a function representing the spectral value of a pixel Is it greater than the set threshold? , the formula is defined as: ； Perform pixel-by-pixel difference calculation on the preprocessed multi-period remote sensing data to generate difference data, including: pass Perform pixel-by-pixel difference calculation on the preprocessed multi-period remote sensing data to generate difference data; in, Indicates at time Location The Band spectral value; Indicates at time Location The Band spectral value; Indicates at location The Band difference value; represents the spatial weight factor; represents the weight coefficient; Indicates the number of bands, Represents the index of different bands; among them, ； in, Indicates location With neighboring pixels The Euclidean distance between The calculation formula of the fitness function is: ； in, Indicates Area of ​​land use type; Indicates Floor area ratio of land use type; represents the unused land area; represents weight; represents the summation symbol; Indicates The service area of ​​each transportation node; Indicates The traffic capacity of each traffic node; represents the average travel distance in the planning area; It represents the maximum possible travel distance in the planning area; Indicates The number of public facilities of this type; Indicates The area of ​​the area where the public facilities are located; Indicates the number of types of public facilities; is the weight coefficient.

2. The urban planning decision support system based on remote sensing mapping according to claim 1 is characterized in that: Perform crossover and mutation operations on planning schemes to generate new planning schemes, including: Encode the planning scheme into binary code; Generate an initial population of planning solutions, which includes multiple randomly generated planning solutions, each of which is a point in the solution space; By evaluating the fitness value of each planning scheme, the corresponding parent scheme is selected from the current population; Select two parent solutions A and B and determine an intersection point or intersection area; Exchange the genetic information of the parent plans A and B after the crossover point or crossover area to generate two new offspring plans C and D. Offspring plan C contains the first part of the genetic information of the parent A and the last part of the genetic information of the parent B. Select an offspring plan C; Determine one or more mutation points, flip the bits in the binary code to change the value of the mutation point to generate a new mutated plan, and repeat the selection, crossover and mutation operations to generate a new planning plan.

3. The urban planning decision support system based on remote sensing mapping according to claim 2 is characterized in that: Urban feature data includes the spatial distribution and attribute information of buildings, roads, green spaces, and water bodies.

4. A method for assisting decision-making in urban planning based on remote sensing mapping, characterized in that: Applicable to the system as claimed in any one of claims 1 to 3, comprising: Acquiring remote sensing data of an urban area, wherein the remote sensing data includes urban surface information; Preprocessing the remote sensing data to obtain preprocessed remote sensing data; Extract urban feature data based on pre-processed remote sensing data; Conduct measurement and analysis of urban feature data, as well as remote sensing data change detection and measurement based on time series to generate measurement and analysis results; A three-dimensional virtual model of the city is constructed according to the measurement and analysis results; corresponding parameters are set for various types of land features in the three-dimensional virtual model of the city; an initial population of planning schemes is generated, and each planning scheme in the population of planning schemes corresponds to a set of parameter combinations in the three-dimensional virtual model of the city; a fitness function for evaluating the pros and cons of each planning scheme is defined; according to the fitness function, the corresponding planning scheme is determined from the current population; crossover and mutation operations are performed on the planning schemes to generate new planning schemes, until a preset number of iterations is reached to obtain a final urban planning scheme.

5. A computing device, characterized in that: include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method according to claim 4.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, which implements the method according to claim 4 when executed by a processor.

Images