A method and system for delineating urban development boundaries based on big data

Abstract

The present invention discloses a method and system for delineating urban development boundaries based on big data, including collecting image data and background data of a development area, extracting features from the image data through a convolutional neural network to obtain regional features; identifying and dividing the regional features according to the ecological coefficient to obtain first zoning data, performing regional analysis on the background data according to ecological sensitivity to obtain second zoning data; matching the first zoning data with the second zoning data to obtain ambiguous areas and consistent data, performing development evaluation according to the background data to obtain development scores, performing regional correction on the ambiguous areas according to the development scores to obtain ambiguous correction data; constructing an urban development boundary delineation model according to the consistent data and the ambiguous correction data, and optimizing the urban development boundary delineation model. The method can improve the accuracy of urban development boundary delineation and can be directly applied to the urban development boundary delineation system.

Description

Technical Field

[0001] The present invention relates to the field of human-computer interaction, and in particular to a method and system for delineating urban development boundaries based on big data. Background Art

[0002] With the acceleration of urbanization and the rapid development of information technology, big data has become an important tool for modern urban planning and management. Although traditional methods for delineating urban development boundaries, such as those based on remote sensing image interpretation and yearbook data, are easy to operate and relatively objective, their limitations and shortcomings are increasingly prominent when dealing with massive, complex and changeable data. Therefore, developing a method for delineating urban development boundaries based on big data is of great significance for improving the scientificity, accuracy and timeliness of urban planning.

[0003] At present, the delineation of urban development boundaries mainly relies on remote sensing image interpretation and yearbook data. Although the remote sensing image interpretation method can intuitively reflect the scope of urban built-up areas, it is insufficient in dealing with the close connection between social activities, economic activities and urban boundaries, and cannot fully reflect the dynamics and complexity of urban development. Although yearbook data can provide relatively comprehensive social and economic information, the data update speed is slow and the timeliness is poor, which makes it difficult to meet the needs of modern urban planning for fast and accurate information acquisition.

[0004] The rapid development of big data technology has provided a new solution for the delineation of urban development boundaries. Big data technology has the characteristics of large data volume, fast processing speed, and diverse data types. It can realize the rapid analysis and processing of massive data, use a variety of data to optimize boundary delineation, and improve the accuracy of boundary delineation. Summary of the invention

[0005] The purpose of the present invention is to provide a method for delineating urban development boundaries based on big data.

[0006] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0007] The present invention comprises the following steps:

[0008] Collecting image data and background data of the development area, preprocessing the image data and the background data, and extracting features from the image data through a convolutional neural network to obtain regional features;

[0009] Identify and divide the regional features according to the ecological coefficient to obtain first zoning data, and perform regional analysis on the background data according to the ecological sensitivity to obtain second zoning data;

[0010] Matching the first zoning data and the second zoning data to obtain an ambiguous region and consistent data, performing development evaluation according to the background data to obtain a development score, and performing regional correction on the ambiguous region according to the development score to obtain ambiguous correction data;

[0011] A town development boundary delineation model is constructed according to the consistent data and the discrepancy data, the town development boundary delineation model is optimized, the data to be delineated is input into the town development boundary delineation model, and the delineation result is output.

[0012] Furthermore, the method of identifying and dividing the regional characteristics according to the ecological coefficient to obtain the first regional division data includes:

[0013] Calculate the ecological coefficient of regional characteristics:

[0014]

[0015] The ecological coefficient of the regional characteristics in the ath region is The number of regional features is The number of features of the zth region in the ath region is n a,z , the zth region in the ath region has the feature k a,z , regional feature k a,z The probability of occurrence is p(k a,z ), the ecological contribution of the zth regional characteristic in the ath region is The ecological integrity rate of the ath region is β a ;

[0016] Areas with an ecological coefficient less than 0.27 are designated as urban development areas, areas with an ecological coefficient greater than 0.27 and less than 0.573 are designated as basic farmland areas, and areas with an ecological coefficient greater than 0.573 and less than 1 are designated as ecological protection areas;

[0017] Input the regional features into the convolutional neural network classification model according to the region, and calculate the convolution layer output:

[0018]

[0019] The convolutional layer output of the ath region is The activation function is γ, the output bias is Y, the convolution kernel size is H, the classification boundary vector is ρ, the penalty coefficient is η, and the numerical error of the z-th region feature is ε z , the diagonal matrix is ​​∑, the ecological correlation matrix of regional characteristics in the ath region is S, transposed to T, the ecological correlation matrix of regional characteristics in different regions is R, and the ecological correlation matrix of different regions in the first y columns is S y , the first y columns of the diagonal matrix are ∑ y The ecological correlation matrix in the first y columns is Ry , the inverse matrix of the diagonal matrix is ​​∑ -1 ;

[0020] Calculate the maximum pooling function value:

[0021]

[0022] The maximum pooling function value of the ath region is W a , the moving length is λ, and the pooling size is h;

[0023] Calculation area category:

[0024]

[0025] The number of regions is q, and the output of the fully connected layer of the ath region is x a , the adjustment coefficient is δ, and the category of the ath area is χ a , output the first zoning data of the area category.

[0026] Furthermore, the method of performing regional analysis on the background data according to ecological sensitivity to obtain second zoning data includes:

[0027] Calculate the correlation between background data and ecology:

[0028]

[0029] Among them The correlation of background data is No. The support of background data is No. The importance of background data is The number of background data is N, The probability of background data appearing in the ecosystem is

[0030] The background data with a correlation greater than 0.413 is used as the key data, and the ecological sensitivity of the key data is calculated:

[0031]

[0032] The ecological sensitivity of the ath key data is The number of key data is In the i-th region The sensitivity weight of key data is The i-th region The background data is The i-th region The external interference of key data is The probability of key data appearing after adding external interference is The ecological restoration value of the rth key data in the i-th region is

[0033] The randomly divided area was divided into unit areas, and the average ecological sensitivity of the unit areas was calculated;

[0034] Unit areas with an average ecological sensitivity of less than 0.195 are designated as ecological protection areas, unit areas with an average ecological coefficient greater than 0.195 and less than 0.479 are designated as basic farmland areas, and unit areas with an average ecological coefficient greater than 0.479 and less than 1 are designated as urban development areas;

[0035] The average ecological sensitivity in the unit areas is compared, the adjacent unit areas are randomly combined according to the principle of the best average ecological sensitivity, and the adjacent areas of the same type are merged and output as the second zoning data.

[0036] Furthermore, the method for performing development evaluation according to the background data to obtain a development score includes:

[0037] Input the background data into the environmental carrying capacity model based on machine learning algorithm to obtain the environmental carrying capacity;

[0038] The development score is calculated based on the environmental carrying capacity and background data within the unit area:

[0039]

[0040] The development score of the τth unit interval is The number of background data is N, and the information entropy of the kth background data in the τth unit interval is The carrying capacity weight of the τth unit interval is x τ , the carrying capacity of the τth unit interval is The first adjustment parameter is θ, the second adjustment parameter is ε, and the conflict coefficient is The coefficient of variation is

[0041] Furthermore, the method of performing regional correction on the ambiguous region according to the development score to obtain ambiguous correction data includes:

[0042] The correction model based on the neural network algorithm for the ambiguous area will be developed and the correction data will be output.

[0043] Furthermore, a method for constructing a town development boundary delineation model based on the consistent data and the discrepancy data includes:

[0044] The objective function is constructed based on the consistent data and the discrepancy data, and the expression is:

[0045]

[0046] The loss function is The jth consistent data is The amount of consistent data is The value of the jth consistent data is No. The calibration data is The number of calibration data is No. The value of each calibration data is

[0047] The urban development boundary delineation models include random forest algorithm, grey correlation algorithm, supervised clustering algorithm, urban development prediction algorithm, and fuzzy comprehensive evaluation algorithm;

[0048] The random forest algorithm divides the input data into training data and test data in a ratio of 5:1;

[0049] The grey correlation algorithm calculates the development correlation in the training data, and sums the development correlation in different regions to obtain the divided development score.

[0050] The supervised clustering algorithm performs interval clustering based on the partition development scores to obtain ecological intervals;

[0051] The urban development prediction algorithm makes a comprehensive development prediction based on the ecological intervals after the unit intervals are divided, and obtains the comprehensive development of the unit intervals;

[0052] The fuzzy comprehensive evaluation algorithm evaluates the comprehensive development of the unit interval, obtains the target value based on the development evaluation, merges the target values ​​greater than 0.73, and outputs the merged boundaries as the predicted demarcation boundaries;

[0053] The parameters of the urban development boundary delineation model are adjusted according to the error between the predicted delineation boundary and the actual delineation boundary.

[0054] Furthermore, the method for optimizing the town development boundary delineation model includes:

[0055] Get the optimal searcher position and calculate the searcher position:

[0056]

[0057] The position of the b-th searcher in the t+1th iteration is The position of the b-th searcher in the t-th iteration is The optimal position of the searcher in the tth iteration is The spiral constant is v, and the random number between -1 and 1 is ζ;

[0058] Calculate the weight coefficient:

[0059]

[0060] The weight coefficient is θ and the maximum number of iterations is t max , update the coefficient vector, the expression is:

[0061]

[0062] s2=2r2

[0063] The first coefficient vector is s1, the second coefficient vector is s2, and the random numbers from 0 to 1 are r1 and r2 respectively;

[0064] Calculate the search subposition in different situations, the expression is:

[0065]

[0066] The probability between 0 and 1 is P, and the randomly selected search subposition is The updated search sub-position is When P<0.5 and s1<1, the searcher position calculation formula is:

[0067]

[0068] Update the search strategy, the expression is:

[0069]

[0070] The neighborhood position of the b-th searcher in the t-th iteration is C b (t), with the neighborhood position C b (t) builds a neighborhood Z for the radius b (t), calculate the search subposition after the search strategy:

[0071]

[0072] The third random number from 0 to 1 is r3, and the position of the b-th search sub-dimension in the t-th iteration after updating the search strategy is The Euclidean distance between the searcher and the neighboring searcher is D. The position of the bDth searcher in the cth dimension after updating the search strategy in the t+1th iteration is The neighborhood position of the cth dimension at the tth iteration is The random position of the cth dimension at the tth iteration is

[0073] Update the search sub-position, the expression is:

[0074]

[0075] Where The fitness of The updated position of the search sub-dimension c of the b-th search sub-iteration in the t+1th iteration is

[0076] The replacement searcher is iterated continuously until the maximum number of iterations is reached, otherwise the coefficient vector is updated.

[0077] Second, a system for defining urban development boundaries based on big data, including:

[0078] Data acquisition module: used to collect image data and background data of the development area, pre-process the image data and the background data, and extract features from the image data through a convolutional neural network to obtain regional features;

[0079] Regional division module: used to identify and divide the regional characteristics according to the ecological coefficient to obtain the first regional division data, and to perform regional analysis on the background data according to the ecological sensitivity to obtain the second regional division data;

[0080] A region correction module: used for matching the first zoning data and the second zoning data to obtain an ambiguous region and consistent data, performing development evaluation according to the background data to obtain a development score, and performing region correction on the ambiguous region according to the development score to obtain ambiguous data;

[0081] Model building module: used to build a town development boundary delineation model based on the consistent data and the discrepancy data, optimize the town development boundary delineation model, input the data to be delineated into the town development boundary delineation model, and output the delineation result.

[0082] The beneficial effects of the present invention are:

[0083] The present invention is a method and system for delineating urban development boundaries based on big data. Compared with the prior art, the present invention has the following technical effects:

[0084] The present invention can improve the accuracy of urban development boundary delineation through preprocessing, feature extraction, identification and division, regional analysis, acquisition of development score, regional correction, model construction and model optimization steps, thereby improving the precision of urban development boundary delineation, optimizing urban development boundary delineation, greatly saving resources, and improving work efficiency. It can realize automatic delineation of urban development boundaries and carry out regional revision of urban development boundary delineation in real time, which is of great significance to urban development boundary delineation, can adapt to urban development boundary delineation of different standards and different urban development boundary delineation requirements, and has a certain universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 This is a flowchart of the steps of a method for delineating urban development boundaries based on big data according to the present invention. DETAILED DESCRIPTION

[0086] The present invention is further described below by means of specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0087] The present invention provides a method and system for delineating urban development boundaries based on big data, comprising the following steps:

[0088] like Figure 1 As shown, in this embodiment, the following steps are included:

[0089] Collecting image data and background data of the development area, preprocessing the image data and the background data, and extracting features from the image data through a convolutional neural network to obtain regional features;

[0090] In the actual assessment, background data include total population, age structure, gender ratio, population density, GDP, industrial output value, labor force employment rate, number of enterprises, fiscal revenue, air quality, water quality, noise pollution level, noise pollution level, ecological environment, public transportation passenger flow, road congestion level, number of traffic accidents, transportation facilities status, land use type, land area, land use intensity, land use change trend;

[0091] Regional characteristics include vegetation coverage, forests, grasslands, farmlands, animal populations, plant populations, ecosystem structure, water conservation, soil conservation, and carbon sinks;

[0092] Identify and divide the regional features according to the ecological coefficient to obtain first zoning data, and perform regional analysis on the background data according to the ecological sensitivity to obtain second zoning data;

[0093] Matching the first zoning data and the second zoning data to obtain an ambiguous region and consistent data, performing development evaluation according to the background data to obtain a development score, and performing regional correction on the ambiguous region according to the development score to obtain ambiguous correction data;

[0094] A town development boundary delineation model is constructed according to the consistent data and the discrepancy data, the town development boundary delineation model is optimized, the data to be delineated is input into the town development boundary delineation model, and the delineation result is output.

[0095] In this embodiment, the method for obtaining first zoning data by identifying and dividing the regional features according to the ecological coefficient includes:

[0096] Calculate the ecological coefficient of regional characteristics:

[0097]

[0098] The ecological coefficient of the regional characteristics in the ath region is The number of regional features is The number of features of the zth region in the ath region is n a,z , the zth region in the ath region has the feature k a,z , regional feature k a,z The probability of occurrence is p(k a,z ), the ecological contribution of the zth regional characteristic in the ath region is The ecological integrity rate of the ath region is β a ;

[0099] Areas with an ecological coefficient less than 0.27 are designated as urban development areas, areas with an ecological coefficient greater than 0.27 and less than 0.573 are designated as basic farmland areas, and areas with an ecological coefficient greater than 0.573 and less than 1 are designated as ecological protection areas;

[0100] Input the regional features into the convolutional neural network classification model according to the region, and calculate the convolution layer output:

[0101]

[0102] The convolutional layer output of the ath region is The activation function is γ, the output bias is Y, the convolution kernel size is H, the classification boundary vector is ρ, the penalty coefficient is η, and the numerical error of the z-th region feature is ε z , the diagonal matrix is ​​∑, the ecological correlation matrix of regional characteristics in the ath region is S, transposed to T, the ecological correlation matrix of regional characteristics in different regions is R, and the ecological correlation matrix of different regions in the first y columns is S y , the first y columns of the diagonal matrix are ∑ y , the ecological correlation matrix within the first y columns is R y , the inverse matrix of the diagonal matrix is ​​∑ -1 ;

[0103] Calculate the maximum pooling function value:

[0104]

[0105] The maximum pooling function value of the ath region is W a , the moving length is λ, and the pooling size is h;

[0106] Calculation area category:

[0107]

[0108] The number of regions is q, and the output of the fully connected layer of the ath region is x a , the adjustment coefficient is δ, and the category of the ath area is χ a , output the first zoning data of the area category.

[0109] In this embodiment, the method of performing regional analysis on the background data according to ecological sensitivity to obtain the second regionalization data includes:

[0110] Calculate the correlation between background data and ecology:

[0111]

[0112] Among them The correlation of background data is No. The support of background data is No. The importance of background data is The number of background data is N, The probability of background data appearing in the ecosystem is

[0113] The background data with a correlation greater than 0.413 is used as the key data, and the ecological sensitivity of the key data is calculated:

[0114]

[0115] The ecological sensitivity of the ath key data is The number of key data is In the i-th region The sensitivity weight of key data is The i-th region The background data is The i-th region The external interference of key data is The probability of key data appearing after adding external interference is The i-th region The ecological restoration value of key data is

[0116] The randomly divided area was divided into unit areas, and the average ecological sensitivity of the unit areas was calculated;

[0117] Unit areas with an average ecological sensitivity of less than 0.195 are designated as ecological protection areas, unit areas with an average ecological coefficient greater than 0.195 and less than 0.479 are designated as basic farmland areas, and unit areas with an average ecological coefficient greater than 0.479 and less than 1 are designated as urban development areas;

[0118] The average ecological sensitivity in the unit areas is compared, the adjacent unit areas are randomly combined according to the principle of the best average ecological sensitivity, and the adjacent areas of the same type are merged and output as the second zoning data.

[0119] In this embodiment, the method for performing development evaluation according to the background data to obtain a development score includes:

[0120] Input the background data into the environmental carrying capacity model based on machine learning algorithm to obtain the environmental carrying capacity;

[0121] The development score is calculated based on the environmental carrying capacity and background data within the unit area:

[0122]

[0123] The development score of the τth unit interval is The number of background data is N, and the information entropy of the κth background data in the τth unit interval is The carrying capacity weight of the τth unit interval is x τ , the carrying capacity of the τth unit interval is The first adjustment parameter is θ, the second adjustment parameter is ε, and the conflict coefficient is The coefficient of variation is

[0124] In this embodiment, the method of performing region correction on the ambiguous region according to the development score to obtain ambiguous correction data includes:

[0125] The correction model based on the neural network algorithm for the ambiguous area will be developed and the correction data will be output.

[0126] In this embodiment, the method for constructing a town development boundary delineation model based on the consistent data and the discrepancy data includes:

[0127] The objective function is constructed based on the consistent data and the discrepancy data, and the expression is:

[0128]

[0129] The loss function is The jth consistent data is The amount of consistent data is The value of the jth consistent data is No. The calibration data is The number of calibration data is No. The value of each calibration data is

[0130] The urban development boundary delineation models include random forest algorithm, grey correlation algorithm, supervised clustering algorithm, urban development prediction algorithm, and fuzzy comprehensive evaluation algorithm;

[0131] The random forest algorithm divides the input data into training data and test data in a ratio of 5:1;

[0132] The grey correlation algorithm calculates the development correlation in the training data, and sums the development correlation in different regions to obtain the divided development score.

[0133] The supervised clustering algorithm performs interval clustering based on the partition development scores to obtain ecological intervals;

[0134] The urban development prediction algorithm makes a comprehensive development prediction based on the ecological intervals after the unit intervals are divided, and obtains the comprehensive development of the unit intervals;

[0135] The fuzzy comprehensive evaluation algorithm evaluates the comprehensive development of the unit interval, obtains the target value based on the development evaluation, merges the target values ​​greater than 0.73, and outputs the merged boundaries as the predicted demarcation boundaries;

[0136] The parameters of the urban development boundary delineation model are adjusted according to the error between the predicted delineation boundary and the actual delineation boundary.

[0137] In this embodiment, the method for optimizing the urban development boundary delineation model includes:

[0138] Get the optimal searcher position and calculate the searcher position:

[0139]

[0140] The position of the b-th searcher in the t+1th iteration is The position of the b-th searcher in the t-th iteration is The optimal position of the searcher in the tth iteration is The spiral constant is ν, and the random number between -1 and 1 is ζ;

[0141] Calculate the weight coefficient:

[0142]

[0143] The weight coefficient is θ and the maximum number of iterations is t max , update the coefficient vector, the expression is:

[0144]

[0145] s2=2r2

[0146] The first coefficient vector is s1, the second coefficient vector is s2, and the random numbers from 0 to 1 are r1 and r2 respectively;

[0147] Calculate the search subposition in different situations, the expression is:

[0148]

[0149] The probability between 0 and 1 is P, and the randomly selected search subposition is The updated search sub-position is When P<0.5 and s1<1, the searcher position calculation formula is:

[0150]

[0151] Update the search strategy, the expression is:

[0152]

[0153] The neighborhood position of the b-th searcher in the t-th iteration is C b (t), with the neighborhood position C b (t) builds a neighborhood Z for the radius b (t), calculate the search subposition after the search strategy:

[0154]

[0155] The third random number from 0 to 1 is r3, and the position of the b-th search sub-dimension in the t-th iteration after updating the search strategy is The Euclidean distance between the searcher and the neighboring searcher is D. The position of the bDth searcher in the cth dimension after updating the search strategy in the t+1th iteration is The neighborhood position of the cth dimension at the tth iteration is The random position of the cth dimension at the tth iteration is

[0156] Update the search sub-position, the expression is:

[0157]

[0158] Where The fitness of The updated position of the search sub-dimension c of the b-th search sub-iteration in the t+1th iteration is

[0159] The replacement searcher is iterated continuously until the maximum number of iterations is reached, otherwise the coefficient vector is updated.

[0160] Second, a system for defining urban development boundaries based on big data, including:

[0161] Data acquisition module: used to collect image data and background data of the development area, pre-process the image data and the background data, and extract features from the image data through a convolutional neural network to obtain regional features;

[0162] Regional division module: used to identify and divide the regional characteristics according to the ecological coefficient to obtain the first regional division data, and to perform regional analysis on the background data according to the ecological sensitivity to obtain the second regional division data;

[0163] A region correction module: used for matching the first zoning data and the second zoning data to obtain an ambiguous region and consistent data, performing development evaluation according to the background data to obtain a development score, and performing region correction on the ambiguous region according to the development score to obtain ambiguous data;

[0164] Model building module: used to build a town development boundary delineation model based on the consistent data and the discrepancy data, optimize the town development boundary delineation model, input the data to be delineated into the town development boundary delineation model, and output the delineation result.

[0165] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for delineating urban development boundaries based on big data, characterized in that: The following steps are involved: Collecting image data and background data of the development area, preprocessing the image data and the background data, and extracting features from the image data through a convolutional neural network to obtain regional features; Identifying and dividing the regional features according to the ecological coefficient to obtain first zoning data, and performing regional analysis on the background data according to the ecological sensitivity to obtain second zoning data; comprising: Calculate the ecological coefficient of regional characteristics: The ecological coefficient of the regional characteristics in the ath region is The number of regional features is The number of features of the zth region in the ath region is n a,z , the zth region in the ath region has the feature k a,z , regional feature k a,z The probability of occurrence is p(k a,z ), the ecological contribution of the zth regional characteristic in the ath region is The ecological integrity rate of the ath region is β a ; Areas with an ecological coefficient less than 0.27 are designated as urban development areas, areas with an ecological coefficient greater than 0.27 and less than 0.573 are designated as basic farmland areas, and areas with an ecological coefficient greater than 0.573 and less than 1 are designated as ecological protection areas; Input the regional features into the convolutional neural network classification model according to the region, and calculate the convolution layer output: The convolutional layer output of the ath region is The activation function is γ, the output bias is Y, the convolution kernel size is H, the classification boundary vector is ρ, the penalty coefficient is η, and the numerical error of the z-th region feature is ε z , the diagonal matrix is ​​∑, the ecological correlation matrix of regional characteristics in the ath region is S, transposed to T, the ecological correlation matrix of regional characteristics in different regions is R, and the ecological correlation matrix of different regions in the first y columns is S y , the first y columns of the diagonal matrix are ∑ y , the ecological correlation matrix within the first y columns is R y , the inverse matrix of the diagonal matrix is ​​∑ -1 ; Calculate the maximum pooling function value: The maximum pooling function value of the ath region is W a , the moving length is λ, and the pooling size is h; Calculation area category: The number of regions is q, and the output of the fully connected layer of the ath region is x a , the adjustment coefficient is δ, and the category of the ath area is χ a , output the first zoning data of the area category; Calculate the correlation between background data and ecology: Among them The correlation of background data is No. The support of background data is No. The importance of background data is The number of background data is N, The probability of background data appearing in the ecosystem is The background data with a correlation greater than 0.413 is used as the key data, and the ecological sensitivity of the key data is calculated: The ecological sensitivity of the i-th key data is The number of key data is In the i-th region The sensitivity weight of key data is The i-th region The background data is The i-th region The external interference of key data is The probability of key data appearing after adding external interference is The i-th region The ecological restoration value of key data is The randomly divided area was divided into unit areas, and the average ecological sensitivity of the unit areas was calculated; Unit areas with an average ecological sensitivity of less than 0.195 are designated as ecological protection areas, unit areas with an average ecological coefficient greater than 0.195 and less than 0.479 are designated as basic farmland areas, and unit areas with an average ecological coefficient greater than 0.479 and less than 1 are designated as urban development areas; Compare the average ecological sensitivity in the unit area, randomly combine the adjacent unit areas according to the principle of the best average ecological sensitivity, and merge the adjacent areas of the same type to output as the second division data; Matching the first zoning data and the second zoning data to obtain an ambiguous region and consistent data, performing development evaluation according to the background data to obtain a development score, and performing regional correction on the ambiguous region according to the development score to obtain ambiguous correction data; comprising: Input the background data into the environmental carrying capacity model based on machine learning algorithm to obtain the environmental carrying capacity; The development score is calculated based on the environmental carrying capacity and background data within the unit area: The development score of the τth unit interval is The number of background data is N, and the information entropy of the kth background data in the τth unit interval is The carrying capacity weight of the τth unit interval is x τ , the carrying capacity of the τth unit interval is The first adjustment parameter is θ, the second adjustment parameter is ε, and the conflict coefficient is The coefficient of variation is A town development boundary delineation model is constructed according to the consistent data and the discrepancy data, the town development boundary delineation model is optimized, the data to be delineated is input into the town development boundary delineation model, and the delineation result is output.

2. The method for delineating urban development boundaries based on big data according to claim 1 is characterized in that: The method for performing regional correction on the ambiguous region according to the development score to obtain ambiguous correction data comprises: The correction model based on the neural network algorithm for the ambiguous area will be developed and the correction data will be output.

3. According to the method for delineating urban development boundaries based on big data as claimed in claim 1, it is characterized in that: The method for constructing a town development boundary delineation model based on the consistent data and the discrepancy data comprises: The objective function is constructed based on the consistent data and the discrepancy data, and the expression is: The loss function is The jth consistent data is The amount of consistent data is The value of the jth consistent data is No. The calibration data is The number of calibration data is No. The value of each calibration data is The urban development boundary delineation models include random forest algorithm, grey correlation algorithm, supervised clustering algorithm, urban development prediction algorithm, and fuzzy comprehensive evaluation algorithm; The random forest algorithm divides the input data into training data and test data in a ratio of 5:1; The grey correlation algorithm calculates the development correlation in the training data, and sums the development correlation in different regions to obtain the divided development score. The supervised clustering algorithm performs interval clustering based on the partition development scores to obtain ecological intervals; The urban development prediction algorithm makes a comprehensive development prediction based on the ecological intervals after the unit intervals are divided, and obtains the comprehensive development of the unit intervals; The fuzzy comprehensive evaluation algorithm evaluates the comprehensive development of the unit interval, obtains the target value based on the development evaluation, merges the target values ​​greater than 0.73, and outputs the merged boundaries as the predicted demarcation boundaries; The parameters of the urban development boundary delineation model are adjusted according to the error between the predicted delineation boundary and the actual delineation boundary.

4. According to the method for delineating urban development boundaries based on big data as claimed in claim 1, it is characterized in that: The method for optimizing the urban development boundary delineation model comprises: Get the optimal searcher position and calculate the searcher position: The position of the b-th searcher in the t+1th iteration is The position of the b-th searcher in the t-th iteration is The optimal position of the searcher in the tth iteration is The spiral constant is ν, and the random number between -1 and 1 is ζ; Calculate the weight coefficient: The weight coefficient is The maximum number of iterations is t max , update the coefficient vector, the expression is: s2＝2r2 The first coefficient vector is s1, the second coefficient vector is s2, and the random numbers from 0 to 1 are r1 and r2 respectively; Calculate the search subposition in different situations, the expression is: The probability between 0 and 1 is P, and the randomly selected search subposition is The updated search sub-position is When P<0.5 and s1<1, the searcher position calculation formula is: Update the search strategy, the expression is: The neighborhood position of the b-th searcher in the t-th iteration is C b (t), with the neighborhood position C b (t) builds a neighborhood Z for the radius b (t), calculate the search subposition after the search strategy: The third random number from 0 to 1 is r3, and the position of the b-th searcher in the c-th dimension after updating the search strategy in the t-th iteration is The Euclidean distance between the searcher and the neighboring searcher is D. The position of the bDth searcher in the cth dimension after updating the search strategy in the t+1th iteration is The neighborhood position of the cth dimension at the tth iteration is The random position of the cth dimension at the tth iteration is Update the search sub-position, the expression is: Where The fitness of The updated position of the search sub-dimension c of the b-th search sub-iteration in the t+1th iteration is The replacement searcher is iterated continuously until the maximum number of iterations is reached, otherwise the coefficient vector is updated.

5. A system for defining urban development boundaries based on big data, characterized in that: include: Data acquisition module: used to collect image data and background data of the development area, pre-process the image data and the background data, and extract features of the image data through a convolutional neural network to obtain regional features; The regional division module is used to identify and divide the regional characteristics according to the ecological coefficient to obtain the first regional division data, and to perform regional analysis on the background data according to the ecological sensitivity to obtain the second regional division data; including: Calculate the ecological coefficient of regional characteristics: The ecological coefficient of the regional characteristics in the ath region is The number of regional features is The number of features of the zth region in the ath region is n a,z , the zth region in the ath region has the feature k a,z , regional feature k a,z The probability of occurrence is p(k a,z ), the ecological contribution of the zth regional characteristic in the ath region is The ecological integrity rate of the ath region is β a ; Areas with an ecological coefficient less than 0.27 are designated as urban development areas, areas with an ecological coefficient greater than 0.27 and less than 0.573 are designated as basic farmland areas, and areas with an ecological coefficient greater than 0.573 and less than 1 are designated as ecological protection areas; Input the regional features into the convolutional neural network classification model according to the region, and calculate the convolution layer output: The convolutional layer output of the ath region is The activation function is γ, the output bias is Y, the convolution kernel size is H, the classification boundary vector is ρ, the penalty coefficient is η, and the numerical error of the z-th region feature is ε z , the diagonal matrix is ​​Σ, the ecological correlation matrix of regional characteristics in the ath region is S, transposed to T, the ecological correlation matrix of regional characteristics in different regions is R, and the ecological correlation matrix of different regions in the first y columns is S y , the first y columns of the diagonal matrix are ∑ y The ecological correlation matrix in the first y columns is R y , the inverse matrix of the diagonal matrix is ​​∑ -1 ; Calculate the maximum pooling function value: The maximum pooling function value of the ath region is W a , the moving length is λ, and the pooling size is h; Calculation area category: The number of regions is q, and the output of the fully connected layer of the ath region is x a , the adjustment coefficient is δ, and the category of the ath area is χ a , output the first zoning data of the area category; Calculate the correlation between background data and ecology: Among them The correlation of background data is No. The support of background data is No. The importance of background data is The number of background data is N, The probability of background data appearing in the ecosystem is The background data with a correlation greater than 0.413 is used as the key data, and the ecological sensitivity of the key data is calculated: The ecological sensitivity of the i-th key data is The number of key data is In the i-th region The sensitivity weight of key data is The i-th region The background data is The i-th region The external interference of key data is The probability of key data appearing after adding external interference is The i-th region The ecological restoration value of key data is The randomly divided area was divided into unit areas, and the average ecological sensitivity of the unit areas was calculated; Unit areas with an average ecological sensitivity of less than 0.195 are designated as ecological protection areas, unit areas with an average ecological coefficient greater than 0.195 and less than 0.479 are designated as basic farmland areas, and unit areas with an average ecological coefficient greater than 0.479 and less than 1 are designated as urban development areas; Compare the average ecological sensitivity in the unit area, randomly combine the adjacent unit areas according to the principle of the best average ecological sensitivity, and merge the adjacent areas of the same type to output as the second division data; A region correction module is used to match the first zoning data and the second zoning data to obtain an ambiguous region and consistent data, perform development evaluation according to the background data to obtain a development score, and perform region correction on the ambiguous region according to the development score to obtain ambiguous correction data; including: Input the background data into the environmental carrying capacity model based on machine learning algorithm to obtain the environmental carrying capacity; The development score is calculated based on the environmental carrying capacity and background data within the unit area: The development score of the τth unit interval is The number of background data is N, and the information entropy of the κth background data in the τth unit interval is The carrying capacity weight of the τth unit interval is x τ , the carrying capacity of the τth unit interval is The first adjustment parameter is θ, the second adjustment parameter is ε, and the conflict coefficient is The coefficient of variation is Model building module: used to build a town development boundary delineation model based on the consistent data and the discrepancy data, optimize the town development boundary delineation model, input the data to be delineated into the town development boundary delineation model, and output the delineation result.

Images