Urban and rural planning management method and system based on artificial intelligence

Abstract

The invention discloses an urban and rural planning management method and system based on artificial intelligence, and belongs to the technical field of urban and rural planning management, and the method comprises the steps of data collection, data preprocessing, urban expansion prediction, air quality prediction and urban and rural planning management. According to the invention, urban and rural planning management is carried out by combining two aspects of urban expansion prediction and air quality prediction, and the planning management dimension is deepened; according to the method, city expansion prediction is carried out by adopting a convolutional long-short-term parallel bidirectional gating circulation network, space and time level features of city expansion are better extracted, the analysis capability of complex data is improved through bidirectional gating circulation, and the overall accuracy of the method is improved; according to the method, air quality prediction is carried out by adopting the time sequence convolution bidirectional long-short-term memory network combined with improved parameter optimization, short-term air change characteristics can be extracted, long-term trend air quality changes can be analyzed, the model performance is optimized by adopting an improved whale optimization search algorithm, and the efficiency and accuracy of urban and rural planning management are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban and rural planning management, and specifically refers to an urban and rural planning management method and system based on artificial intelligence. Background Art

[0002] Artificial intelligence-based urban and rural planning management uses artificial intelligence technology to intelligently process, analyze and predict various data related to urban and rural planning, aiming to provide a scientific basis for urban and rural planning decisions, optimize resource allocation, improve management efficiency, and promote the sustainable development of urban and rural areas.

[0003] However, in the existing intelligent urban and rural planning management methods, there is a technical problem that urban and rural management planning needs to consider both the urban expansion at the macro level and the specific urban and rural data at the micro level. The existing methods mostly characterize urban and rural planning from the macro level through changes in urban area, but lack micro-level considerations, as well as the combined analysis, prediction and management of the two, and it is difficult to meet the actual needs of urban and rural planning management; in the existing urban expansion prediction methods, there is a technical problem that urban expansion prediction involves complex information of two different dimensions, image and time series data, and the existing methods have low parallel processing capabilities for image features and time series features, and the existing technology is difficult to meet the dual needs of urban and rural planning construction for computing efficiency and computing accuracy; in the existing air quality prediction methods, there is a technical problem that the changes in air quality at the border between urban and rural areas involve both short-term high-frequency changes and long-term trend changes, while traditional methods focus on trend change analysis and easily ignore short-term air quality fluctuations, resulting in insufficient depth of air quality analysis and difficulty in providing support for urban and rural planning analysis at the micro level. Summary of the invention

[0004] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an urban and rural planning management method and system based on artificial intelligence. In view of the existing intelligent urban and rural planning management methods, there is a problem that urban and rural management planning needs to consider both the macro-level urban expansion and the micro-level specific urban and rural data. The existing methods mostly characterize urban and rural planning from a macro level through changes in urban area, but lack micro-level considerations, as well as the combined analysis, prediction and management of the two, which makes it difficult to meet the actual needs of urban and rural planning management. The present solution creatively adopts the combination of urban expansion prediction and air quality prediction at the border between urban and rural areas to carry out urban and rural planning management. Through the coordinated analysis of macro expansion and micro air indicators, the dimension of planning management is effectively deepened, and more intelligent, comprehensive and comprehensive technical support is provided for urban and rural planning management. In view of the fact that in the existing urban expansion prediction methods, urban expansion prediction involves complex information of two different dimensions, image and time series data, and the existing methods have low parallel processing capabilities for image features and time series features, the existing technology is difficult to meet the dual needs of urban and rural planning construction for computing efficiency and computing accuracy. Technical issues: This solution creatively adopts a convolutional long-short-term parallel bidirectional gated recurrent network to predict urban expansion. Through the design of parallel structured convolution and long-short-term memory, it effectively extracts the urban expansion characteristics at the spatial and temporal levels. By using the two characteristics as the input of the bidirectional gated recurrent network, the bidirectional structure effectively improves the analysis ability of complex data and the overall accuracy of the method. In the existing air quality prediction methods, there is a technical problem that the air quality changes at the border between urban and rural areas involve both short-term high-frequency changes and long-term trend changes, while the traditional methods focus on trend change analysis and easily ignore short-term air quality fluctuations, resulting in insufficient depth of air quality analysis and difficulty in providing support for urban and rural planning analysis at the micro level. This solution creatively adopts a time-series convolutional bidirectional long-short-term memory network combined with improved parameter optimization for air quality prediction. It extracts short-term air change characteristics through time-series convolution and processes long-term trend air quality changes through bidirectional long-short-term memory. It also adopts an improved whale optimization search algorithm to provide more performance optimization for the parameters of complex mixed models, which improves the efficiency and accuracy of urban and rural planning management as a whole.

[0005] The technical solution adopted by the present invention is as follows: The present invention provides an urban and rural planning management method based on artificial intelligence, the method comprising the following steps:

[0006] Step S1: data collection;

[0007] Step S2: data preprocessing;

[0008] Step S3: urban expansion prediction;

[0009] Step S4: air quality prediction;

[0010] Step S5: Urban and rural planning management.

[0011] Furthermore, in step S1, the data collection includes the following steps:

[0012] Step S11: collecting original urban expansion data, specifically, collecting urban remote sensing image data, social economic data and historical urban expansion data to obtain original urban expansion data;

[0013] The socio-economic data, including population growth, GDP, economic development level and land prices;

[0014] Step S12: collecting original air quality data, specifically, collecting air quality monitoring data and meteorological data in the urban-rural border area to obtain the original air quality data;

[0015] The air quality monitoring data includes PM2.5, SO 2 and CO 2 Pollutant concentration;

[0016] The meteorological data include temperature, humidity, wind speed, wind direction and precipitation.

[0017] Furthermore, in step S2, the data preprocessing includes the following steps:

[0018] Step S21: pre-processing the original urban expansion data, specifically, performing data cleaning and data standardization on the original urban expansion data, and converting the socio-economic data and historical urban expansion data in the original urban expansion data into time series data, so as to obtain high-quality urban expansion data;

[0019] Step S22: pre-processing the original air quality data, specifically, obtaining high-quality air quality data by performing data cleaning and data standardization on the original air quality data.

[0020] Further, in step S3, the urban expansion prediction is used to predict the urban expansion area, specifically, based on the high-quality urban expansion data, a convolutional long-term and short-term parallel bidirectional gated recurrent network is used to perform urban expansion prediction to obtain the predicted urban expansion area;

[0021] The convolutional long short-term parallel bidirectional gated recurrent network includes a parallel convolutional long short-term memory subnet, a bidirectional gated recurrent subnet and a prediction output layer;

[0022] The parallel convolutional long short-term memory subnetwork includes a parallel convolutional subnetwork and a long short-term memory subnetwork;

[0023] The convolutional subnet is used to extract spatial features from urban remote sensing image data;

[0024] The long short-term memory subnet is used to capture complex temporal dependencies;

[0025] The bidirectional gated recurrent subnetwork is used to capture bidirectional time series information, thereby improving the model prediction performance;

[0026] The urban expansion prediction comprises the following steps:

[0027] Step S31: constructing a parallel convolutional long short-term memory subnet, specifically setting a parallel convolutional subnet and a long short-term memory subnet in the parallel convolutional long short-term memory subnet, processing the high-quality data of urban expansion through the parallel convolutional long short-term memory subnet, and obtaining the spatial characteristics and temporal characteristics of urban expansion, including the following steps:

[0028] Step S311: constructing a convolutional subnet, specifically receiving urban remote sensing image data in the urban expansion high-quality data as convolutional subnet input data, stacking N convolutional pooling layers in the convolutional subnet, and setting a flattening layer after the last convolutional pooling layer, and processing the convolutional subnet input data through the convolutional subnet to obtain the urban expansion spatial features;

[0029] The convolutional pooling layer includes a convolutional layer and a pooling layer, and the calculation formula is:

[0030] ;

[0031] In the formula, h cnn is the convolutional feature map, specifically the output feature map of the convolutional layer, Conv2D(·) is a two-dimensional convolution operation, X is the convolutional layer input, and w cnn is the convolution kernel, b cnn is the convolution bias term, h cp It is the convolutional pooling feature map, specifically the output feature map of the pooling layer, and MaxPooling(·) is the maximum pooling operation;

[0032] Step S312: constructing a long short-term memory subnet, specifically receiving the socioeconomic data and historical urban expansion data in the urban expansion high-quality data as the long short-term memory subnet input data, stacking M long short-term memory layers in the long short-term memory subnet, and processing the long short-term memory subnet input data through the long short-term memory subnet to obtain the urban expansion time series characteristics;

[0033] Step S32: constructing a bidirectional gated recurrent subnetwork, specifically receiving the urban expansion spatial features and the urban expansion temporal features as the bidirectional gated recurrent subnetwork input data, setting a self-attention mechanism and K bidirectional gated recurrent layers in the bidirectional gated recurrent subnetwork, and processing the bidirectional gated recurrent subnetwork input data through the bidirectional gated recurrent subnetwork to obtain the urban expansion bidirectional features;

[0034] Step S33: constructing a prediction output layer, specifically, processing the bidirectional features of urban expansion through a fully connected layer and a linear activation function to obtain a prediction output layer result;

[0035] Step S34: model training, specifically, performing model training by constructing a parallel convolutional long short-term memory subnet, constructing a bidirectional gated recurrent subnet, and constructing a prediction output layer to obtain an urban expansion prediction model;

[0036] Step S35: Calculate the urban expansion prediction result, specifically, perform urban expansion prediction through the urban expansion prediction model to obtain the urban expansion prediction area.

[0037] Further, in step S4, the air quality prediction is used to predict the air quality of the urban-rural border area, specifically based on the high-quality air quality data, using the time series convolutional bidirectional long short-term memory network combined with improved parameter optimization to perform air quality prediction, including the following steps:

[0038] Step S41: constructing an improved parameter optimization algorithm, specifically constructing an improved whale optimization algorithm, including the following steps:

[0039] Step S411: Initialize the whale population by chaotic mapping, specifically by initializing each whale position in the search space through Logistic mapping, and the calculation formula is:

[0040] ;

[0041] In the formula, x k+1 is the chaotic value generated by the k+1th mapping, k is the mapping index, c is the Logistic mapping control parameter, x k is the chaotic value generated by the kth mapping, the initial value of which is a random number in the range [0,1], pos i,j is the position of the i-th whale in the j-th dimension, i is the whale index, j is the dimension index, up j is the upper bound of the search space of the jth dimension, low j is the lower bound of the search space of the jth dimension;

[0042] Step S412: Evaluate the fitness of the whale population, specifically, select the model accuracy as the fitness function, calculate the fitness value of each whale position through the fitness function, and select the whale position with the best fitness value as the optimal whale position;

[0043] Step S413: Perform mutation operation on the whale positions with the bottom 5% of fitness value ranking, and the calculation formula is:

[0044] ;

[0045] In the formula, is the position of the ith whale in the jth dimension after mutation, G j This is the best position for whales at present. is the position of the first whale randomly selected in the current iteration in the jth dimension, is the position of the second whale randomly selected in the current iteration in the jth dimension, is a binary random number whose value is 0 or 1, and is used to ensure the randomness of the mutation operation;

[0046] Step S414: updating the whale position, specifically updating the whale position by simulating the whale's prey-seeking behavior and prey-catching behavior, including the following steps:

[0047] Step S4141: Prey-seeking behavior is used to explore the search space over a large area. Specifically, the whale position is updated by simulating the prey-seeking behavior of the whale. The calculation formula is:

[0048] ;

[0049] In the formula, is the position of the ith whale in the jth dimension at the t+1th iteration, t is the iteration index, is the position of the ith whale in the jth dimension at the tth iteration, U i is the step size coefficient of the ith whale, which is used to control the amplitude of the whale position update. is the mean position of all whales in the jth dimension, u is the fitness coefficient, r is a random number in the range [0,1], t max is the maximum number of iterations, is the decay factor, which is used to control the decay rate of the fitness coefficient;

[0050] Step S4142: Prey hunting behavior is used to refine the development of the search space. Specifically, the prey hunting behavior of whales is simulated and the spiral update strategy is adopted to update the whale position. The calculation formula is:

[0051] ;

[0052] In the formula, is the adaptive step size coefficient, is the opposite position of the ith whale in the jth dimension at the tth iteration, cos(·) is the cosine function, is the circumference of a circle, q is the spiral control parameter, is the initial adaptive step coefficient, a is the spiral amplitude constant, which is used to control the spiral update amplitude;

[0053] Step S4143: Selecting a whale position update method, specifically randomly generating a whale behavior factor in the range of [0,1]. When the whale behavior factor is greater than or equal to 0.5, the whale position is updated through the prey-finding behavior described in step S4141. Otherwise, the whale position is updated through the prey-hunting behavior described in step S4142.

[0054] Step S415: Iteration of the whale population, specifically, iteratively updating the whale position by repeatedly executing steps S412 to S414 until the maximum number of iterations is reached;

[0055] Step 42: Construct a temporal convolutional bidirectional long short-term memory network, including the following steps:

[0056] Step S421: construct a time series convolutional subnet to extract local time series features, specifically, stack L convolutional layers in the time series convolutional subnet, set a causal dilated convolutional layer, a ReLU activation function and a random dropout layer in each convolutional layer, and set a residual connection between each convolutional layer. The air quality high-quality data is processed through the time series convolutional subnet to obtain the air quality time series convolutional features;

[0057] Step S422: constructing a bidirectional long short-term memory subnet to capture global dependencies, specifically setting a forward long short-term memory layer and a backward long short-term memory layer in the bidirectional long short-term memory subnet, and processing the air quality temporal convolution features through the bidirectional long short-term memory subnet to obtain the air quality long-term and short-term features;

[0058] Step S423: constructing an output layer, specifically, processing the bidirectional features of urban expansion through a fully connected layer and a Softmax activation function to obtain an output layer result;

[0059] Step S43: constructing an air quality prediction model, specifically by training a temporal convolutional bidirectional long short-term memory network and using an improved parameter optimization algorithm to tune model parameters to obtain an air quality prediction model;

[0060] Step S44: Calculate the air quality prediction result, specifically, perform air quality prediction through an air quality prediction model to obtain air quality prediction levels, wherein the air quality prediction levels include excellent, good, light pollution, moderate pollution, heavy pollution and severe pollution.

[0061] Further, in step S5, the urban and rural planning management is specifically to generate an urban and rural planning management report by integrating the predicted urban expansion area and the predicted air quality level, with a focus on supervising the urban and rural bordering areas with poor predicted air quality levels;

[0062] The air quality forecast level is poor, which specifically refers to the air quality forecast level reaching light pollution or above.

[0063] The present invention provides an urban and rural planning management system based on artificial intelligence, comprising: a data collection module, a data preprocessing module, an urban expansion prediction module, an air quality prediction module and an urban and rural planning management module;

[0064] The data collection module is used for data collection, and obtains urban expansion raw data and air quality raw data through data collection, and sends the urban expansion raw data and the air quality raw data to the data preprocessing module;

[0065] The data preprocessing module is used for data preprocessing, and obtains urban expansion high-quality data and air quality high-quality data through data preprocessing, and sends the urban expansion high-quality data to the urban expansion prediction module, and sends the air quality high-quality data to the air quality prediction module;

[0066] The urban expansion prediction module is used for urban expansion prediction, obtains the urban expansion prediction area through urban expansion prediction, and sends the urban expansion prediction area to the urban and rural planning management module;

[0067] The air quality prediction module is used for air quality prediction, obtains the air quality prediction level through air quality prediction, and sends the air quality prediction level to the urban and rural planning management module;

[0068] The urban and rural planning management module is used for urban and rural planning management, and generates an urban and rural planning management report by integrating the predicted urban expansion area and the predicted air quality level.

[0069] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0070] (1) In view of the technical problem that the existing intelligent urban and rural planning management methods need to consider both the urban expansion at the macro level and the specific urban and rural data at the micro level, the existing methods mostly characterize urban and rural planning from the macro level through the change of urban area, but lack the consideration of the micro level, as well as the combined analysis, prediction and management of the two, which makes it difficult to meet the actual needs of urban and rural planning management. This solution creatively adopts the combination of urban expansion prediction and air quality prediction at the border between urban and rural areas to carry out urban and rural planning management. Through the coordinated analysis of macro expansion and micro air indicators, it effectively deepens the dimension of planning management and provides smarter, more comprehensive and more comprehensive technical support for urban and rural planning management.

[0071] (2) In view of the technical problem that the existing urban expansion prediction methods involve complex information of two different dimensions, image and time series data, and the existing methods have low parallel processing capabilities for image features and time series features, and the existing technology is difficult to meet the dual requirements of urban and rural planning and construction for computing efficiency and computing accuracy, this solution creatively adopts convolutional long-short-term parallel bidirectional gated recurrent network to predict urban expansion. Through the parallel structure of convolution and long-short-term memory design, the urban expansion characteristics at the spatial and temporal levels are effectively extracted. By using the two features as the input of the bidirectional gated recurrent network respectively, the bidirectional structure effectively improves the analysis capability of complex data and improves the overall accuracy of the method.

[0072] (3) In the existing air quality prediction methods, there is a technical problem that the changes in air quality at the border between urban and rural areas involve both short-term high-frequency changes and long-term trend changes. Traditional methods focus on trend change analysis and tend to ignore short-term air quality fluctuations, resulting in insufficient depth of air quality analysis and difficulty in providing support for urban and rural planning analysis at the micro level. This solution creatively uses a time series convolutional bidirectional long short-term memory network combined with improved parameter optimization for air quality prediction. It extracts short-term air change characteristics through time series convolution and processes long-term trend air quality changes through bidirectional long short-term memory. It also uses an improved whale optimization search algorithm to provide more performance optimization for the parameters of complex hybrid models, thereby improving the efficiency and accuracy of urban and rural planning management as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A schematic diagram of a flow chart of an urban and rural planning management method based on artificial intelligence provided by the present invention;

[0074] Figure 2 A schematic diagram of an urban and rural planning management system based on artificial intelligence provided by the present invention;

[0075] Figure 3 is a schematic flow chart of step S3;

[0076] Figure 4 It is a schematic diagram of the process of step S4.

[0077] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0078] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0079] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0080] Example 1, see Figure 1 The present invention provides an urban and rural planning management method based on artificial intelligence, which comprises the following steps:

[0081] Step S1: data collection;

[0082] Step S2: data preprocessing;

[0083] Step S3: urban expansion prediction;

[0084] Step S4: air quality prediction;

[0085] Step S5: Urban and rural planning management;

[0086] By executing the above operations, it is found that in the existing intelligent urban and rural planning management methods, there is a problem that urban and rural management planning needs to consider both the urban expansion at the macro level and the specific urban and rural data at the micro level. The existing methods mostly characterize urban and rural planning from a macro level through changes in urban area, but lack micro-level considerations, as well as the combined analysis, prediction and management of the two, which makes it difficult to meet the actual needs of urban and rural planning management. This solution creatively adopts the combination of urban expansion prediction and air quality prediction at the border between urban and rural areas to carry out urban and rural planning management. Through the coordinated analysis of macro expansion and micro air indicators, it effectively deepens the dimension of planning management, and also provides smarter, more comprehensive and more comprehensive technical support for urban and rural planning management.

[0087] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the data collection includes the following steps:

[0088] Step S11: collecting original urban expansion data, specifically, collecting urban remote sensing image data, social economic data and historical urban expansion data to obtain original urban expansion data;

[0089] The socio-economic data, including population growth, GDP, economic development level and land prices;

[0090] Step S12: collecting original air quality data, specifically, collecting air quality monitoring data and meteorological data in the urban-rural border area to obtain the original air quality data;

[0091] The air quality monitoring data includes PM2.5, SO 2 and CO 2 Pollutant concentration;

[0092] The meteorological data include temperature, humidity, wind speed, wind direction and precipitation.

[0093] Example 3, see Figure 1 This embodiment is based on the above embodiment. In step S2, the data preprocessing includes the following steps:

[0094] Step S21: pre-processing the original urban expansion data, specifically, performing data cleaning and data standardization on the original urban expansion data, and converting the socio-economic data and historical urban expansion data in the original urban expansion data into time series data, so as to obtain high-quality urban expansion data;

[0095] Step S22: pre-processing the original air quality data, specifically, obtaining high-quality air quality data by performing data cleaning and data standardization on the original air quality data.

[0096] Example 4, see Figure 1 and Figure 3 , this embodiment is based on the above embodiment. In step S3, the urban expansion prediction is used to predict the urban expansion area. Specifically, based on the high-quality urban expansion data, a convolutional long-term and short-term parallel bidirectional gated recurrent network is used to predict the urban expansion to obtain the predicted urban expansion area;

[0097] The convolutional long short-term parallel bidirectional gated recurrent network includes a parallel convolutional long short-term memory subnet, a bidirectional gated recurrent subnet and a prediction output layer;

[0098] The parallel convolutional long short-term memory subnetwork includes a parallel convolutional subnetwork and a long short-term memory subnetwork;

[0099] The convolutional subnet is used to extract spatial features from urban remote sensing image data;

[0100] The long short-term memory subnet is used to capture complex temporal dependencies;

[0101] The bidirectional gated recurrent subnetwork is used to capture bidirectional time series information, thereby improving the model prediction performance;

[0102] The urban expansion prediction comprises the following steps:

[0103] Step S31: constructing a parallel convolutional long short-term memory subnet, specifically setting a parallel convolutional subnet and a long short-term memory subnet in the parallel convolutional long short-term memory subnet, processing the high-quality data of urban expansion through the parallel convolutional long short-term memory subnet, and obtaining the spatial characteristics and temporal characteristics of urban expansion, including the following steps:

[0104] Step S311: constructing a convolutional subnet, specifically receiving urban remote sensing image data in the urban expansion high-quality data as convolutional subnet input data, stacking N convolutional pooling layers in the convolutional subnet, and setting a flattening layer after the last convolutional pooling layer, and processing the convolutional subnet input data through the convolutional subnet to obtain the urban expansion spatial features;

[0105] The convolutional pooling layer includes a convolutional layer and a pooling layer, and the calculation formula is:

[0106] ;

[0107] In the formula, h cnn is the convolutional feature map, specifically the output feature map of the convolutional layer, Conv2D(·) is a two-dimensional convolution operation, X is the convolutional layer input, and w cnn is the convolution kernel, b cnn is the convolution bias term, h cp It is the convolutional pooling feature map, specifically the output feature map of the pooling layer, and MaxPooling(·) is the maximum pooling operation;

[0108] Step S312: constructing a long short-term memory subnet, specifically receiving the socioeconomic data and historical urban expansion data in the urban expansion high-quality data as the long short-term memory subnet input data, stacking M long short-term memory layers in the long short-term memory subnet, and processing the long short-term memory subnet input data through the long short-term memory subnet to obtain the urban expansion time series characteristics;

[0109] Step S32: constructing a bidirectional gated recurrent subnetwork, specifically receiving the urban expansion spatial features and the urban expansion temporal features as the bidirectional gated recurrent subnetwork input data, setting a self-attention mechanism and K bidirectional gated recurrent layers in the bidirectional gated recurrent subnetwork, and processing the bidirectional gated recurrent subnetwork input data through the bidirectional gated recurrent subnetwork to obtain the urban expansion bidirectional features;

[0110] Step S33: constructing a prediction output layer, specifically, processing the bidirectional features of urban expansion through a fully connected layer and a linear activation function to obtain a prediction output layer result;

[0111] Step S34: model training, specifically, performing model training by constructing a parallel convolutional long short-term memory subnet, constructing a bidirectional gated recurrent subnet, and constructing a prediction output layer to obtain an urban expansion prediction model;

[0112] Step S35: Calculate the urban expansion prediction result, specifically, perform urban expansion prediction through the urban expansion prediction model to obtain the urban expansion prediction area;

[0113] By performing the above operations, in view of the technical problem that in the existing urban expansion prediction methods, urban expansion prediction involves complex information of two different dimensions, image and time series data, and the existing methods have low parallel processing capabilities for image features and time series features, and the existing technology is difficult to meet the dual requirements of urban and rural planning and construction for computing efficiency and computing accuracy, this solution creatively adopts a convolutional long-short-term parallel bidirectional gated recurrent network to predict urban expansion. Through the parallel structure of convolution and long-short-term memory design, the urban expansion characteristics at the spatial and temporal levels are effectively extracted, and by using the two features as the input of the bidirectional gated recurrent network respectively, the bidirectional structure effectively improves the analysis capability of complex data and improves the overall accuracy of the method.

[0114] Example 5, see Figure 1 and Figure 4 This embodiment is based on the above embodiment. In step S4, the air quality prediction is used to predict the air quality of the urban-rural border area. Specifically, the air quality prediction is performed based on the high-quality air quality data and using a temporal convolutional bidirectional long short-term memory network combined with improved parameter optimization, including the following steps:

[0115] Step S41: constructing an improved parameter optimization algorithm, specifically constructing an improved whale optimization algorithm, including the following steps:

[0116] Step S411: Initialize the whale population by chaotic mapping, specifically by initializing each whale position in the search space through Logistic mapping, and the calculation formula is:

[0117] ;

[0118] In the formula, x k+1 is the chaotic value generated by the k+1th mapping, k is the mapping index, c is the Logistic mapping control parameter, x k is the chaotic value generated by the kth mapping, the initial value of which is a random number in the range [0,1], pos i,jis the position of the i-th whale in the j-th dimension, i is the whale index, j is the dimension index, up j is the upper bound of the search space of the jth dimension, low j is the lower bound of the search space of the jth dimension;

[0119] Step S412: Evaluate the fitness of the whale population, specifically, select the model accuracy as the fitness function, calculate the fitness value of each whale position through the fitness function, and select the whale position with the best fitness value as the optimal whale position;

[0120] Step S413: Perform mutation operation on the whale positions with the bottom 5% of fitness value ranking, and the calculation formula is:

[0121] ;

[0122] In the formula, is the position of the ith whale in the jth dimension after mutation, G j This is the best position for whales at present. is the position of the first whale randomly selected in the current iteration in the jth dimension, is the position of the second whale randomly selected in the current iteration in the jth dimension, is a binary random number whose value is 0 or 1, and is used to ensure the randomness of the mutation operation;

[0123] Step S414: updating the whale position, specifically updating the whale position by simulating the whale's prey-seeking behavior and prey-catching behavior, including the following steps:

[0124] Step S4141: Prey-seeking behavior is used to explore the search space over a large area. Specifically, the whale position is updated by simulating the prey-seeking behavior of the whale. The calculation formula is:

[0125] ;

[0126] In the formula, is the position of the ith whale in the jth dimension at the t+1th iteration, t is the iteration index, is the position of the ith whale in the jth dimension at the tth iteration, U i is the step size coefficient of the ith whale, which is used to control the amplitude of the whale position update. is the mean position of all whales in the jth dimension, u is the fitness coefficient, r is a random number in the range [0,1], t max is the maximum number of iterations, is the decay factor, which is used to control the decay rate of the fitness coefficient;

[0127] Step S4142: Prey hunting behavior is used to refine the development of the search space. Specifically, the prey hunting behavior of whales is simulated and the spiral update strategy is adopted to update the whale position. The calculation formula is:

[0128] ;

[0129] In the formula, is the adaptive step size coefficient, is the opposite position of the ith whale in the jth dimension at the tth iteration, cos(·) is the cosine function, is the circumference of a circle, q is the spiral control parameter, is the initial adaptive step coefficient, a is the spiral amplitude constant, which is used to control the spiral update amplitude;

[0130] Step S4143: Selecting a whale position update method, specifically randomly generating a whale behavior factor in the range of [0,1]. When the whale behavior factor is greater than or equal to 0.5, the whale position is updated through the prey-finding behavior described in step S4141. Otherwise, the whale position is updated through the prey-hunting behavior described in step S4142.

[0131] Step S415: Iteration of the whale population, specifically, iteratively updating the whale position by repeatedly executing steps S412 to S414 until the maximum number of iterations is reached;

[0132] Step 42: Construct a temporal convolutional bidirectional long short-term memory network, including the following steps:

[0133] Step S421: construct a time series convolutional subnet to extract local time series features, specifically, stack L convolutional layers in the time series convolutional subnet, set a causal dilated convolutional layer, a ReLU activation function and a random dropout layer in each convolutional layer, and set a residual connection between each convolutional layer. The air quality high-quality data is processed through the time series convolutional subnet to obtain the air quality time series convolutional features;

[0134] Step S422: constructing a bidirectional long short-term memory subnet to capture global dependencies, specifically setting a forward long short-term memory layer and a backward long short-term memory layer in the bidirectional long short-term memory subnet, and processing the air quality temporal convolution features through the bidirectional long short-term memory subnet to obtain the air quality long-term and short-term features;

[0135] Step S423: constructing an output layer, specifically, processing the bidirectional features of urban expansion through a fully connected layer and a Softmax activation function to obtain an output layer result;

[0136] Step S43: constructing an air quality prediction model, specifically by training a temporal convolutional bidirectional long short-term memory network and using an improved parameter optimization algorithm to tune model parameters to obtain an air quality prediction model;

[0137] Step S44: Calculate the air quality prediction result, specifically, perform air quality prediction through an air quality prediction model to obtain an air quality prediction level, wherein the air quality prediction level includes excellent, good, light pollution, moderate pollution, heavy pollution and severe pollution;

[0138] By performing the above operations, in view of the fact that in the existing air quality prediction methods, the changes in air quality at the border of urban and rural areas involve both short-term high-frequency changes and long-term trend changes, and the traditional methods focus on trend change analysis, which easily ignores short-term air quality fluctuations, resulting in insufficient depth of air quality analysis and difficulty in providing support for urban and rural planning analysis at the micro level, this solution creatively adopts a time series convolution bidirectional long short-term memory network combined with improved parameter optimization for air quality prediction, extracts short-term air change characteristics through time series convolution, and processes long-term trend air quality changes through bidirectional long short-term memory, and adopts an improved whale optimization search algorithm to provide more performance optimization for the parameters of complex hybrid models, which improves the efficiency and accuracy of urban and rural planning management as a whole.

[0139] Example 6, see Figure 1 , this embodiment is based on the above embodiment. In step S5, the urban and rural planning management is specifically to generate an urban and rural planning management report by integrating the predicted urban expansion area and the predicted air quality level, and to focus on supervising the urban and rural bordering areas with poor predicted air quality levels;

[0140] The predicted air quality level is poor, specifically referring to the predicted air quality level reaching light pollution or above;

[0141] Embodiment 7, see Figure 2 , this embodiment is based on the above embodiment, and the present invention provides an urban and rural planning management system based on artificial intelligence, including: a data collection module, a data preprocessing module, an urban expansion prediction module, an air quality prediction module and an urban and rural planning management module;

[0142] The data collection module is used for data collection, and obtains urban expansion raw data and air quality raw data through data collection, and sends the urban expansion raw data and the air quality raw data to the data preprocessing module;

[0143] The data preprocessing module is used for data preprocessing, and obtains urban expansion high-quality data and air quality high-quality data through data preprocessing, and sends the urban expansion high-quality data to the urban expansion prediction module, and sends the air quality high-quality data to the air quality prediction module;

[0144] The urban expansion prediction module is used for urban expansion prediction, obtains the urban expansion prediction area through urban expansion prediction, and sends the urban expansion prediction area to the urban and rural planning management module;

[0145] The air quality prediction module is used for air quality prediction, obtains the air quality prediction level through air quality prediction, and sends the air quality prediction level to the urban and rural planning management module;

[0146] The urban and rural planning management module is used for urban and rural planning management, and generates an urban and rural planning management report by integrating the predicted urban expansion area and the predicted air quality level.

[0147] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0148] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0149] The present invention and its embodiments are described above, which is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it and design structural methods and embodiments similar to the technical solution without creativity without departing from the purpose of the invention, they should all fall within the protection scope of the present invention.

Claims

1. An urban and rural planning management method based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1: data collection; Step S2: data preprocessing; Step S3: urban expansion prediction, which is used to predict the urban expansion area. Specifically, based on the high-quality urban expansion data, a convolutional long-term and short-term parallel bidirectional gated recurrent network is used to predict urban expansion and obtain the predicted urban expansion area. The convolutional long short-term parallel bidirectional gated recurrent network includes a parallel convolutional long short-term memory subnet, a bidirectional gated recurrent subnet and a prediction output layer; The parallel convolutional long short-term memory subnetwork includes a parallel convolutional subnetwork and a long short-term memory subnetwork; The convolutional subnet is used to extract spatial features from urban remote sensing image data; The long short-term memory subnet is used to capture complex temporal dependencies; The bidirectional gated recurrent subnet is used to capture bidirectional timing information; Step S4: air quality prediction; Step S5: Urban and rural planning management.

2. The urban and rural planning management method based on artificial intelligence according to claim 1 is characterized in that: In step S3, the urban expansion prediction includes the following steps: Step S31: constructing a parallel convolutional long short-term memory subnet, specifically setting a parallel convolutional subnet and a long short-term memory subnet in the parallel convolutional long short-term memory subnet, processing the high-quality data of urban expansion through the parallel convolutional long short-term memory subnet, and obtaining the spatial characteristics and temporal characteristics of urban expansion, including the following steps: Step S311: constructing a convolutional subnet, specifically receiving urban remote sensing image data in the urban expansion high-quality data as convolutional subnet input data, stacking N convolutional pooling layers in the convolutional subnet, and setting a flattening layer after the last convolutional pooling layer, and processing the convolutional subnet input data through the convolutional subnet to obtain the urban expansion spatial features; The convolutional pooling layer includes a convolutional layer and a pooling layer, and the calculation formula is: ； In the formula, h cnn is the convolutional feature map, specifically the output feature map of the convolutional layer, Conv2D(·) is a two-dimensional convolution operation, X is the convolutional layer input, and w cnn is the convolution kernel, b cnn is the convolution bias term, h cp It is the convolutional pooling feature map, specifically the output feature map of the pooling layer, and MaxPooling(·) is the maximum pooling operation; Step S312: constructing a long short-term memory subnet, specifically receiving the socioeconomic data and historical urban expansion data in the urban expansion high-quality data as the long short-term memory subnet input data, stacking M long short-term memory layers in the long short-term memory subnet, and processing the long short-term memory subnet input data through the long short-term memory subnet to obtain the urban expansion time series characteristics; Step S32: constructing a bidirectional gated recurrent subnetwork, specifically receiving the urban expansion spatial features and the urban expansion temporal features as the bidirectional gated recurrent subnetwork input data, setting a self-attention mechanism and K bidirectional gated recurrent layers in the bidirectional gated recurrent subnetwork, and processing the bidirectional gated recurrent subnetwork input data through the bidirectional gated recurrent subnetwork to obtain the urban expansion bidirectional features; Step S33: constructing a prediction output layer, specifically, processing the bidirectional features of urban expansion through a fully connected layer and a linear activation function to obtain a prediction output layer result; Step S34: model training, specifically, performing model training by constructing a parallel convolutional long short-term memory subnet, constructing a bidirectional gated recurrent subnet, and constructing a prediction output layer to obtain an urban expansion prediction model; Step S35: Calculate the urban expansion prediction result, specifically, perform urban expansion prediction through the urban expansion prediction model to obtain the urban expansion prediction area.

3. The urban and rural planning management method based on artificial intelligence according to claim 2 is characterized in that: In step S4, the air quality prediction is used to predict the air quality of the urban-rural border area. Specifically, based on the high-quality air quality data, the air quality prediction is performed using a temporal convolutional bidirectional long short-term memory network combined with improved parameter optimization, including the following steps: Step S41: constructing an improved parameter optimization algorithm, specifically constructing an improved whale optimization algorithm, including the following steps: Step S411: Initialize the whale population by chaotic mapping, specifically by initializing each whale position in the search space through Logistic mapping, and the calculation formula is: ； In the formula, x k+1 is the chaotic value generated by the k+1th mapping, k is the mapping index, c is the Logistic mapping control parameter, x k is the chaotic value generated by the kth mapping, the initial value of which is a random number in the range [0,1], pos i,j is the position of the i-th whale in the j-th dimension, i is the whale index, j is the dimension index, up j is the upper bound of the search space of the jth dimension, low j is the lower bound of the search space of the jth dimension; Step S412: Evaluate the fitness of the whale population, specifically, select the model accuracy as the fitness function, calculate the fitness value of each whale position through the fitness function, and select the whale position with the best fitness value as the optimal whale position; Step S413: Perform mutation operation on the whale positions with the bottom 5% of fitness value ranking, and the calculation formula is: ； In the formula, is the position of the ith whale in the jth dimension after mutation, G j This is the best position for whales at present. is the position of the first whale randomly selected in the current iteration in the jth dimension, is the position of the second whale randomly selected in the current iteration in the jth dimension, is a binary random number whose value is 0 or 1, and is used to ensure the randomness of the mutation operation; Step S414: updating the whale position, specifically updating the whale position by simulating the whale's prey-seeking behavior and prey-catching behavior, including the following steps: Step S4141: Prey-seeking behavior is used to explore the search space over a large area. Specifically, the whale position is updated by simulating the prey-seeking behavior of the whale. The calculation formula is: ； In the formula, is the position of the ith whale in the jth dimension at the t+1th iteration, t is the iteration index, is the position of the ith whale in the jth dimension at the tth iteration, U i is the step size coefficient of the ith whale, which is used to control the amplitude of the whale position update. is the mean position of all whales in the jth dimension, u is the fitness coefficient, r is a random number in the range [0,1], t max is the maximum number of iterations, is the decay factor, which is used to control the decay rate of the fitness coefficient; Step S4142: Prey hunting behavior is used to refine the development of the search space. Specifically, the prey hunting behavior of whales is simulated and the spiral update strategy is adopted to update the whale position. The calculation formula is: ； In the formula, is the adaptive step size coefficient, is the opposite position of the ith whale in the jth dimension at the tth iteration, cos(·) is the cosine function, is the circumference of a circle, q is the spiral control parameter, is the initial adaptive step coefficient, a is the spiral amplitude constant, which is used to control the spiral update amplitude; Step S4143: Selecting a whale position update method, specifically randomly generating a whale behavior factor in the range of [0,1]. When the whale behavior factor is greater than or equal to 0.5, the whale position is updated through the prey-finding behavior described in step S4141. Otherwise, the whale position is updated through the prey-hunting behavior described in step S4142. Step S415: Iteration of the whale population, specifically, iteratively updating the whale position by repeatedly executing steps S412 to S414 until the maximum number of iterations is reached; Step 42: Construct a temporal convolutional bidirectional long short-term memory network, including the following steps: Step S421: construct a time series convolutional subnet to extract local time series features, specifically, stack L convolutional layers in the time series convolutional subnet, set a causal dilated convolutional layer, a ReLU activation function and a random dropout layer in each convolutional layer, and set a residual connection between each convolutional layer. The air quality high-quality data is processed through the time series convolutional subnet to obtain the air quality time series convolutional features; Step S422: constructing a bidirectional long short-term memory subnet to capture global dependencies, specifically setting a forward long short-term memory layer and a backward long short-term memory layer in the bidirectional long short-term memory subnet, and processing the air quality temporal convolution features through the bidirectional long short-term memory subnet to obtain the air quality long-term and short-term features; Step S423: constructing an output layer, specifically, processing the bidirectional features of urban expansion through a fully connected layer and a Softmax activation function to obtain an output layer result; Step S43: constructing an air quality prediction model, specifically by training a temporal convolutional bidirectional long short-term memory network and using an improved parameter optimization algorithm to tune model parameters to obtain an air quality prediction model; Step S44: Calculate the air quality prediction result, specifically, perform air quality prediction through an air quality prediction model to obtain an air quality prediction level.

4. The urban and rural planning management method based on artificial intelligence according to claim 3 is characterized in that: In step S2, the data preprocessing includes the following steps: Step S21: pre-processing the original urban expansion data, specifically, performing data cleaning and data standardization on the original urban expansion data, and converting the socio-economic data and historical urban expansion data in the original urban expansion data into time series data, so as to obtain high-quality urban expansion data; Step S22: pre-processing the original air quality data, specifically, obtaining high-quality air quality data by performing data cleaning and data standardization on the original air quality data.

5. The urban and rural planning management method based on artificial intelligence according to claim 4 is characterized in that: In step S1, the data collection includes the following steps: Step S11: collecting original urban expansion data, specifically, collecting urban remote sensing image data, social economic data and historical urban expansion data to obtain original urban expansion data; Step S12: Collecting original air quality data, specifically, collecting air quality monitoring data and meteorological data in the urban-rural border area to obtain original air quality data.

6. The urban and rural planning management method based on artificial intelligence according to claim 5 is characterized by: In step S5, the urban and rural planning management specifically generates an urban and rural planning management report by comprehensively considering the predicted urban expansion area and the predicted air quality level, with a focus on supervising the urban-rural border areas with poor predicted air quality levels.

7. An urban and rural planning management system based on artificial intelligence, used to implement an urban and rural planning management method based on artificial intelligence as described in any one of claims 1 to 6, characterized in that: It includes data collection module, data preprocessing module, urban expansion prediction module, air quality prediction module and urban and rural planning management module.

8. The urban and rural planning management system based on artificial intelligence according to claim 8 is characterized by: The data collection module is used for data collection, and obtains urban expansion raw data and air quality raw data through data collection, and sends the urban expansion raw data and the air quality raw data to the data preprocessing module; The data preprocessing module is used for data preprocessing, and obtains urban expansion high-quality data and air quality high-quality data through data preprocessing, and sends the urban expansion high-quality data to the urban expansion prediction module, and sends the air quality high-quality data to the air quality prediction module; The urban expansion prediction module is used for urban expansion prediction, obtains the urban expansion prediction area through urban expansion prediction, and sends the urban expansion prediction area to the urban and rural planning management module; The air quality prediction module is used for air quality prediction, obtains the air quality prediction level through air quality prediction, and sends the air quality prediction level to the urban and rural planning management module; The urban and rural planning management module is used for urban and rural planning management, and generates an urban and rural planning management report by integrating the predicted urban expansion area and the predicted air quality level.

Images