Rural residential area improvement area identification method, system, equipment and medium

Abstract

The invention discloses a rural residential area improvement area identification method, system, device and medium, and the method comprises the steps: firstly obtaining the environment development data and land utilization data of a target improvement area, and obtaining the land utilization data of the target improvement area based on the environment development data; obtaining a residential area improvement scale constraint of a target year in the target improvement area through a first agent behavior rule model, and obtaining an improvement willingness degree of a second agent of the target year in the target improvement area through a second agent behavior rule model based on the land utilization data, and finally, according to the residence renovation scale constraint and the renovation willingness of the second agent, performing optimization selection on the target renovation area through a seed area growth algorithm to obtain the residence renovation scale and spatial layout of the target year in the target renovation area. According to the invention, the flexibility of the residential area renovation area identification result can be enhanced, and the rationality and stability of the residential area renovation area identification result can be improved.

Description

Technical Field

[0001] The present invention relates to the field of land consolidation technology, and in particular to a method, system, equipment and medium for identifying a rural residential consolidation area. Background Art

[0002] In order to ensure that regional land use development meets the requirements of social development and ecological environment, human settlements need to be rectified. Therefore, it is necessary to efficiently and accurately identify settlement remediation areas in complex nonlinear land use patterns. However, the settlement remediation process is affected by the joint decision-making of different relevant remediation subjects. Traditional methods mostly use per capita construction land indicators, per capita construction land indicators and land idle rate as single indicators for theoretical identification, resulting in a large gap between the identification results and the actual results; or through the empowerment of indicators for multi-factor comprehensive correction, but this is too dependent on expert prior knowledge, subjective interference is strong, resulting in the neglect of the role of the main subject behavior in the remediation process. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method, system, device and medium for identifying rural residential area renovation areas, which can reduce subjective factors in the process of residential area renovation area identification and improve the authenticity and reliability of residential area renovation results.

[0004] In order to solve the above technical problems, an embodiment of the present invention provides a method for identifying a rural residential area renovation area, comprising:

[0005] Acquire decision-making environment data of the target remediation area; wherein the decision-making environment data includes environmental development data and land use data;

[0006] Based on the environmental development data, a first agent behavior rule model is used to obtain a target constraint for residential renovation in the target renovation area in the target year;

[0007] Based on the land use data, obtaining a second agent that agrees to conduct remediation in the target remediation area in the target year through a second agent behavior rule model;

[0008] According to the target constraint of settlement renovation, the renovation willingness of the second intelligent body and the preset scale constraint of settlement renovation, the target renovation area is optimized and selected through the seed region growing algorithm to obtain the scale and spatial layout of settlement renovation in the target renovation area in the target year.

[0009] Optionally, the first agent behavior rule model is trained by the following steps:

[0010] Obtain historical environmental development data of the target remediation area as training samples;

[0011] The training samples are used to train the improved ensemble learning algorithm to obtain a trained first agent behavior rule model.

[0012] Optionally, the second agent behavior rule model is obtained by training through the following steps:

[0013] Acquire historical land use data of the target remediation area; wherein the historical land use data of the target remediation area includes historical land use data of the remediation-implemented area and historical land use data of the remediation-unimplemented area;

[0014] The historical land use data of the areas where remediation has been implemented is used as positive training samples, and the historical land use data of the areas where remediation has not been implemented is used as negative training samples, and the long short-term memory network is trained to obtain a trained second agent behavior rule model.

[0015] Optionally, the second intelligent entity that agrees to the renovation includes the area of ​​the residential area that agrees to be renovated and the renovation willingness of the second intelligent entity that agrees to be renovated.

[0016] Furthermore, according to the settlement renovation target constraint, the renovation willingness of the second intelligent agent and the preset settlement renovation contiguous scale constraint, the target renovation area is optimized and selected through a seed region growing algorithm to obtain the settlement renovation scale and spatial layout in the target renovation area in the target year, including:

[0017] According to the preset scale constraints of the contiguous area of ​​residential area renovation and the renovation willingness of the second intelligent agent who agrees to the renovation, the second intelligent agent that meets the requirements and agrees to the renovation is selected;

[0018] If the settlement area owned by the second agent that agrees to conduct the renovation in the target renovation area is larger than the contiguous scale constraint, a seed is selected according to the constraint threshold; wherein the seed is a second agent among the second agents that agree to conduct the renovation whose renovation willingness is greater than the constraint threshold;

[0019] All seeds are sorted according to the degree of willingness to be regulated, and the sorted seeds are expanded to the second intelligent agent in the eight neighborhoods in turn until the expansion area threshold is met; wherein, the edge level is given to the second intelligent agent in the eight neighborhoods during the expansion process;

[0020] The expanded second agent is optimized and selected according to the edge level of the second agent until the renovation target constraint is met, and the renovation scale and spatial layout of the settlements in the target renovation area in the target year are obtained.

[0021] Optionally, the constraint threshold is calculated by the following formula:

[0022]

[0023] Wherein, T1 is the constraint threshold, Persential(a,b) is the percentile function used to obtain the bth percentile of a, Prop2 is the remediation willingness of the second agent who agrees to the remediation, and R is the target remediation area. is the settlement area of ​​the second agent that agrees to be renovated, Constrained by the remediation objectives.

[0024] Optionally, the expansion area threshold is calculated by the following formula:

[0025]

[0026] Among them, T E is the expansion area threshold.

[0027] In order to solve the above technical problems, an embodiment of the present invention further provides a residential area improvement area identification system, comprising:

[0028] A decision environment module, used to obtain decision environment data of the target remediation area; wherein the decision environment data includes environmental development data and land use data;

[0029] A first agent module is used to obtain, based on the environmental development data, a settlement renovation scale constraint in the target renovation area in the target year through a first agent behavior rule model;

[0030] A second agent module is used to obtain the remediation willingness of the second agent in the target remediation area in the target year through a second agent behavior rule model based on the land use data;

[0031] The settlement remediation area identification module is used to optimize the selection of the target remediation area through a seed region growing algorithm according to the settlement remediation scale constraint and the remediation willingness of the second intelligent agent, so as to obtain the settlement remediation scale and spatial layout in the target remediation area in the target year.

[0032] To solve the above technical problems, an embodiment of the present invention further provides a terminal device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the rural settlement remediation area identification method as described in any one of the above items.

[0033] In order to solve the above technical problems, an embodiment of the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the rural residential area improvement area identification methods described above.

[0034] Compared with the prior art, the method, system, device and medium for identifying rural residential remediation areas provided by the embodiments of the present invention first obtain environmental development data and land use data of the target remediation area, and based on the environmental development data, obtain the scale constraint of residential remediation in the target remediation area in the target year through the first agent behavior rule model, and based on the land use data, obtain the remediation willingness of the second agent in the target remediation area in the target year through the second agent behavior rule model, and finally optimize the target remediation area according to the scale constraint of residential remediation and the remediation willingness of the second agent through the seed region growth algorithm to obtain the scale and spatial layout of residential remediation in the target year in the target remediation area. The present invention combines actual decision-making environment factors, establishes a model by coupling a deep learning algorithm, and thus constructs the behavioral rules of relevant subjects in the process of resident remediation, which can reduce subjective empowerment intervention, thereby enhancing the flexibility of the identification results of residential remediation areas, and improving the rationality and stability of the identification results of residential remediation areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solution of the present invention, the drawings used in the implementation mode will be briefly introduced below. Obviously, the drawings described below are only some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 is a flow chart of a method for identifying a rural residential area renovation area provided by an embodiment of the present invention;

[0037] Figure 2 It is a schematic diagram of the architecture of a method for identifying a rural residential area renovation area provided by an embodiment of the present invention;

[0038] Figure 3 is a schematic diagram of a training process of a first agent behavior rule model provided by an embodiment of the present invention;

[0039] Figure 4 is a cross-validation accuracy result of a first agent behavior rule model provided by an embodiment of the present invention;

[0040] Figure 5is a schematic diagram of the training principle of a second agent behavior rule model provided by an embodiment of the present invention;

[0041] Figure 6 is a verification accuracy result of a second agent behavior rule model provided by an embodiment of the present invention;

[0042] Figure 7 This is a flow chart of optimizing the selection of a target remediation area through a seed region growing algorithm provided by an embodiment of the present invention;

[0043] Figure 8 It is a structural block diagram of a rural residential area improvement area identification system provided by an embodiment of the present invention;

[0044] Fig. 9 It is a structural block diagram of a terminal device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] See also Figure 1 , Figure 1 1 is a flow chart of a method for identifying a rural residential area renovation area provided by an embodiment of the present invention, wherein the method for identifying a rural residential area renovation area comprises steps S1 to S4:

[0047] S1. Obtaining decision-making environment data of the target remediation area; wherein the decision-making environment data includes environmental development data and land use data;

[0048] It should be noted that the decision-making environment includes the natural environment and social environment in which the relevant subjects are located during the settlement renovation process. Based on the analysis of the role connotations of the relevant subjects in the settlement renovation process, the decision-making factors that affect the behavior of the relevant subjects are selected for the intelligent agent model in the future, and specific indicators are selected to reflect the impact of various decision-making factors.

[0049] For example, environmental development data may include various environmental development indicators of the target remediation area, such as the development needs, economic level, distribution pattern and terrain conditions of the target remediation area, which can reflect the current situation and development needs of the target remediation area. Land use data involves the use of land in the target remediation area, including land use conditions and pattern conditions for different purposes, which helps to understand the rationality and optimization space of land use in the target remediation area.

[0050] S2. Based on the environmental development data, obtaining the target constraint of residential renovation in the target renovation area in the target year through the first agent behavior rule model;

[0051] S3, based on the land use data, obtaining a second agent that agrees to conduct remediation in the target remediation area in the target year through a second agent behavior rule model;

[0052] S4. Based on the target constraint of settlement renovation, the renovation willingness of the second intelligent agent and the preset scale constraint of settlement renovation, the target renovation area is optimized and selected through the seed region growing algorithm to obtain the scale and spatial layout of settlement renovation in the target renovation area in the target year.

[0053] It should be noted that, in the specific implementation, the first agent is mainly responsible for the decision-making of the remediation of the settlement. Such agents are usually government agencies, planning departments or relevant decision-making bodies. By establishing a specific decision-making model, the first agent can evaluate the scale constraints of the remediation based on environmental development data to ensure the compliance and rationality of the remediation plan. The second agent is mainly the stakeholders who may be affected by the remediation. The second agent reflects the needs and concerns of stakeholders by evaluating the support and expectations of the stakeholders for the remediation plan.

[0054] It is worth noting that the interaction between the first agent and the second agent is very important. On the one hand, the first agent needs to adjust the remediation target based on the feedback from the second agent; on the other hand, the implementation of the remediation must also obtain the consent of the second agent. By integrating the remediation target constraints of the first agent and the willingness of the second agent, a feedback loop can be established to ensure that the planning of the remediation area not only meets policy requirements but also effectively meets the actual needs of residents.

[0055] Specifically, the second intelligent entity that agrees to the renovation includes the area of ​​the residential area that agrees to be renovated and the renovation willingness of the second intelligent entity that agrees to be renovated.

[0056] Furthermore, according to the settlement renovation target constraint, the renovation willingness of the second intelligent agent and the preset settlement renovation contiguous scale constraint, the target renovation area is optimized and selected through a seed region growing algorithm to obtain the settlement renovation scale and spatial layout in the target renovation area in the target year, including:

[0057] According to the preset scale constraints of the contiguous area of ​​residential area renovation and the renovation willingness of the second intelligent agent who agrees to the renovation, the second intelligent agent that meets the requirements and agrees to the renovation is selected;

[0058] If the settlement area owned by the second agent that agrees to conduct the renovation in the target renovation area is larger than the contiguous scale constraint, a seed is selected according to the constraint threshold; wherein the seed is a second agent among the second agents that agree to conduct the renovation whose renovation willingness is greater than the constraint threshold;

[0059] All seeds are sorted according to the degree of willingness to be regulated, and the sorted seeds are expanded to the second intelligent agent in the eight neighborhoods in turn until the expansion area threshold is met; wherein, the edge level is given to the second intelligent agent in the eight neighborhoods during the expansion process;

[0060] The expanded second agent is optimized and selected according to the edge level of the second agent until the renovation target constraint is met, and the renovation scale and spatial layout of the settlements in the target renovation area in the target year are obtained.

[0061] In one embodiment, the constraint threshold is calculated by the following formula:

[0062]

[0063] Wherein, T1 is the constraint threshold, Persential(a,b) is the percentile function used to obtain the bth percentile of a, Prop2 is the remediation willingness of the second agent who agrees to the remediation, and R is the target remediation area. is the settlement area of ​​the second agent that agrees to be renovated, Constrained by the remediation objectives.

[0064] Furthermore, the expansion area threshold is calculated by the following formula:

[0065]

[0066] Among them, T E is the expansion area threshold.

[0067] It is worth noting that the seed region growing algorithm is an algorithm that starts from the initial "seed" point and gradually expands the region. It grows by evaluating the suitability of the surrounding area and can handle the clustering and regional optimization problems of spatial data. The embodiment of the present invention combines the target constraints of settlement renovation, renovation willingness and contiguous scale constraints, and uses the seed region growing algorithm to perform optimization selection, which can effectively realize the scientific identification and layout of renovation areas.

[0068] In one embodiment, the first agent behavior rule model is trained by the following steps:

[0069] Obtain historical environmental development data of the target remediation area as training samples;

[0070] The training samples are used to train the improved ensemble learning algorithm to obtain a trained first agent behavior rule model.

[0071] Exemplarily, the ensemble learning improvement algorithm can be LGBM (Light Gradient Boost Machine), which is a tree-based ensemble learning method that uses gradient boosting technology. By combining multiple weak learners (usually decision trees) into a powerful model, it has better generalization ability, stronger robustness and better feature category segmentation effect than general machine learning algorithms. Compared with other ensemble learning algorithms such as GBRT (GradientBoosting Regression Tree) and XGBoost (eXtreme Gradient Boosting, an optimized distributed gradient boosting library), LGBM can achieve higher computational efficiency while maintaining the same level of accuracy.

[0072] In one embodiment, the second agent behavior rule model is trained by the following steps:

[0073] Acquire historical land use data of the target remediation area; wherein the historical land use data of the target remediation area includes historical land use data of the remediation area and land use data of the non-remediation area;

[0074] The historical land use data of the areas where remediation has been implemented is used as positive training samples, and the historical land use data of the areas where remediation has not been implemented is used as negative training samples, and the long short-term memory network is trained to obtain a trained second agent behavior rule model.

[0075] For example, the long short-term memory network may be LSTM (Long Short-term Memory), which is a recurrent neural network variant designed to solve the "vanishing gradient" problem. For land use problems with spatiotemporal dependencies, LSTM is applicable to spatiotemporal data mining.

[0076] In order to enable those skilled in the art to more clearly understand the rural residential area improvement area identification method provided by the present invention, a detailed description is given below using the local government as the first intelligent agent and farmers as the second intelligent agent to identify rural residential area improvement areas as an example.

[0077] See also Figure 2 , Figure 2 Schematic diagram of a rural residential area improvement area identification method provided by an embodiment of the present invention. Figure 2As shown, the overall framework includes a decision environment module, an agent preference module, and an agent interaction module.

[0078] First, based on the analysis of the roles of local governments and farmers in the process of rural settlement renovation, we can select decision factors that affect the behaviors of local governments and farmers, and determine specific indicators to reflect the impact of various factors. Specific decision factors can be adjusted appropriately according to actual conditions.

[0079] For example, according to the decision-making environment of rural settlement renovation, the decision factors and corresponding indicators selected for the government agent and the farmer agent can be referred to Table 1.

[0080] Table 1 Decision factors for rural settlement renovation

[0081]

[0082]

[0083] Secondly, the behavioral rules of the government agent constructed in advance through the LGBM algorithm are used to clarify the rectification goals, that is, to obtain the scale constraints of the rectification goals; at the same time, the behavioral rules of the farmer agent constructed in advance through LSTM are used to determine the rectification preferences, that is, to obtain the degree of rectification willingness.

[0084] It should be noted that the decision-making behavior of the government agent constructed using LGBM can be represented by formula (1) to describe the rectification goal of the government agent under the feedback of the decision-making environment:

[0085]

[0086] In formula (1), gov i Represents the government agent i, Area gov Represents the remediation goals determined by the government agent, is the j-type decision factor that affects the government agent (see Table 1).

[0087] Specifically, the land use change data of historical years are superimposed as training samples, and the change in the area of ​​rural settlements in each unit is obtained as the label value for constructing the behavior rules of the government agent. The samples are divided into training sets and test sets for model accuracy verification, and the coefficient of determination (R 2 ), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) are used to evaluate the accuracy of the trained LGBM model in learning the rule-based behaviors of government agents.

[0088] See also Figure 3 , is a schematic diagram of a training process of a first agent behavior rule model provided by an embodiment of the present invention, such as Figure 3 As shown, firstly, we select urban and rural development needs, government fiscal pressure, agricultural development level, regional food security, rural distribution pattern, construction site distribution pattern and location and terrain conditions. The decision factors of seven government agents are used to match the time and space of the data set, which is divided into a training data set and a test data set. LGBM is used to train the model on the training data set, initialize a weak learner, and for each sample, calculate the gradient of its loss function with respect to the model output. These gradients are used as new target variables to train a new weak learner, and the newly trained weak learner is added to the model after being weighted, and the model prediction value is updated. At the same time, it is necessary to specify the parameters of the model, such as the learning rate, the number of trees, the depth of the tree, the feature splitting strategy, etc. The parameters can be tuned by cross-validation and other methods to obtain better model performance. The trained model is evaluated on the test data set, and common evaluation indicators such as accuracy, precision, recall, F1-score, etc. can be used. Finally, the multiple decision factors selected above are used as real sample labels, and the trained model is used to predict the new data to obtain the output of the model.

[0089] For example, using Figure 3 The training process shown in the figure shows that the main hyper parameters of the LGBM model obtained by training based on the historical data of Province A in the past ten years include the number of weak learners, subsampling rate, number of leaves, tree depth and learning rate; the number of weak learners in the model reaches 1200, the subsampling rate is 0.6, the number of leaves is 15, the tree depth is 4, and the learning rate is 0.05. The results show that the trained LGBM model has achieved a high accuracy.

[0090] See also Figure 4 , Figure 4 is a cross-validation accuracy result of a first agent behavior rule model provided by an embodiment of the present invention. Figure 4 As shown, the R2 value of the model is 0.9875, showing excellent explanatory power for the characteristic variables; the MAE is 22.16hm 2 Compared with the calculation results of the traditional method for the size constraints of rural settlements, the error is reduced by three orders of magnitude; the MAPE is 6.95%, which meets the availability condition of less than 20% and achieves an overall good accuracy level of less than 10%.

[0091] It should be noted that by using LSTM to construct the decision-making behavior rules of the farmer agent, the farmer agent's preference for regulation under the observation and perception of the decision-making environment can be represented by formula (2):

[0092]

[0093] In formula (2), farmer i represents farmer agent i, and Prop framer characterizes the willingness of the farmer agent to participate in the renovation. Its value range is 0 to 1. The closer it is to 1, the more willing the farmer agent is to participate in the renovation. is the j-type decision factor affecting the farmer agent (refer to Table 1).

[0094] Specifically, there is a congruent spatial mapping relationship between the farmer agent and the rural residential area. Training samples can be obtained by overlaying the land use data of historical years. Among them, the area that was a rural residential area in year A (year A < year B) but not a rural residential area in year B can be considered as the renovated area that has been implemented, and it is used as a positive training sample with a label value of 1; the area that was not a rural residential area in year A but is a rural residential area in year B, or the area that was a rural residential area in year A and is still a rural residential area in year B can be considered as the area that is not suitable for renovation for the time being, and it is used as a negative training sample with a label value of 0. Then, the training samples are divided into a training set and a test set for the accuracy verification of the module. The probability threshold of the renovation willingness of the farmer agent is obtained through the Receiver Operating Characteristic curve (ROC curve). This threshold can achieve the optimal classification of the training samples. If the renovation willingness probability of a certain farmer agent is greater than this threshold, it is considered an individual who agrees to participate in the renovation. In addition, the precision (formula 3), recall rate (formula 4), and accuracy rate (formula 5) are used to evaluate the accuracy of the test results.

[0095]

[0096] In the formula, TP, TN, FP, and FN represent true positive (actually positive, predicted positive), true negative (actually negative, predicted negative), false positive (actually negative, predicted positive), and false negative (actually positive, predicted negative), respectively.

[0097] See Figure 5 , Figure 5 is a schematic diagram of the training principle of a second agent behavior rule model provided by an embodiment of the present invention. As Figure 5 shown, decision factors of five farmer agents such as social environmental conditions, natural environmental conditions, land use conditions, landscape pattern conditions, and image texture features are input into the LSTM network, and finally the renovation willingness degree and decision preference of the corresponding farmer agent are output through the fully connected layer.

[0098] For example, the main hyperparameters of the LSTM model trained based on the historical data of Province A in the past ten years are 2 hidden layers, (100, 80) hidden layer nodes, 512 batch sizes, 30 iterations, 0.2 dropout rate, 1 early termination tolerance, Adam optimization function, and 0.001 learning rate. Among them, the precision rate of negative samples is 91.17%, the recall rate is 88.13%, and the accuracy rate is 89.80%. The precision rate of positive samples is 88.52%, the recall rate is 91.47%, and the accuracy rate is 89.80%, all of which show that the trained LSTM model has achieved high accuracy.

[0099] For example, see Figure 6 , Figure 6 is a verification accuracy result of a second agent behavior rule model provided by an embodiment of the present invention. Figure 6 As shown, the horizontal axis is FPR (False Positive Rate), the vertical axis is TPR (True Positive Rate, that is, recall rate), and the AUC value of the model is 0.9574. The AUC value is the area under the ROC curve. The closer it is to 1.0, the better the model performance. Therefore, it shows that the second agent behavior rule model provided in this embodiment is a better classifier. The probability threshold of farmers' willingness to rectify the situation obtained in the figure is 0.6330.

[0100] Furthermore, during the specific implementation, the base year data of the target remediation area can be collected, and the quantity and scale constraints of the rural settlement remediation potential in the target year (government agent preference) can be obtained through the government agent behavior rule model, and the farmer agents who agree to remediation in the target year (farmer agent preference) can be obtained through the farmer agent behavior rule model.

[0101] Finally, through the potential scale constraints of remediation formulated by the government agent and the remediation willingness of the farmer agent, the interaction between the government and the farmer agent is realized, and the farmer subjects who agree to remediation are optimized to obtain the final potential scale and spatial layout of rural settlement remediation.

[0102] It should be noted that the implementation of the final decision on rural settlement renovation often relies on the cooperation of formal administrative agencies (government agents) and major interest groups (farmer agents). On the one hand, due to the limitation of capital cost, control requirements and the need for contiguous renovation, the government cannot meet the needs of all farmers who express positive intentions to implement settlement renovation; on the other hand, when conditions such as funds and land use indicators permit, the implementation of renovation projects must also obtain the consent of farmers. Therefore, it is necessary to consider the wishes of farmers under the constraints of government agents and obtain the final result through the interaction between agents.

[0103] Specifically, see Figure 7 , Figure 7 It is a flowchart for optimizing the selection of a target remediation area through a seed region growing algorithm provided by an embodiment of the present invention.

[0104] like Figure 7 As shown in the figure, firstly, the precondition of agent interaction is clarified, that is, the remediation target scale constraint of the target remediation area R is obtained by the government agent behavior rule model. And the scale of the contiguous area is restricted conti , and the farmer agents that agree to the regulation obtained by the farmer agent behavior rules And constrain Area according to the contiguous scale conti and the second agent's willingness to regulate, remove those that do not meet the requirements (i.e., the farmer agent who refuses to be regulated ), thereby selecting the farmer agents that agree to the regulation in, include Area of ​​rural settlements and Prob farmer ;

[0105] Then select the seeds If the target remediation area R Area of ​​rural settlements Less than If all of them are included in the rectification, they will not participate in the subsequent steps. Otherwise, the rectification willingness Prob is selected first according to the constraint threshold (Formula 6). farmer Higher For seeds (such as Figure 7 , as shown in a);

[0106]

[0107] Expand into potential areas: Based on Prob farmer right Sort in descending order and sort by To the potential farmer agents in the eight neighborhoods Expansion (such as Figure 7 , as shown in b), until the expansion area threshold is met (Formula 7). In this process, Edge level L (such as Figure 7 , as shown in c), where the farther away from the screening seed, the larger L is.

[0108]

[0109] Finally, the redundant potential is discarded: to ensure that the scale of the optimized selected potential is still First, after the potential area expands, adjacent and continuous potential patches will appear. The average value of the remediation willingness of the set is the average remediation willingness of the patch. according to In ascending order, select the one with the largest L Set, discard Prob farmer Smallest Until satisfied Constraints, finally get the optimized selection That is, the final scale and spatial layout of rural settlement renovation.

[0110] In summary, the embodiments of the present invention combine the actual decision-making environment factors, establish a model by coupling the deep learning algorithm, and thus construct the behavioral rules of relevant subjects in the process of resident remediation, which can reduce subjective empowerment intervention, thereby enhancing the flexibility of the results of residential remediation area identification, and improving the rationality and stability of the results of residential remediation area identification.

[0111] On the basis of the above-mentioned method items, the present invention provides corresponding embodiments of the system items.

[0112] See also Figure 8 , Figure 8 The following is a structural block diagram of a system for identifying residential area renovation provided by an embodiment of the present invention. The system for identifying rural residential area renovation includes:

[0113] A decision environment module 21 is used to obtain decision environment data of the target remediation area; wherein the decision environment data includes environmental development data and land use data;

[0114] The first agent module 22 is used to obtain the target constraint of settlement renovation in the target renovation area in the target year through the first agent behavior rule model based on the environmental development data;

[0115] The second agent module 23 is used to obtain the second agent that agrees to conduct remediation in the target remediation area in the target year through the second agent behavior rule model based on the land use data;

[0116] The rural settlement renovation area identification module 24 optimizes the selection of the target renovation area through a seed area growing algorithm according to the settlement renovation target constraints, the renovation willingness of the second intelligent body and the preset settlement renovation contiguous scale constraints, and obtains the settlement renovation scale and spatial layout in the target renovation area in the target year.

[0117] In an optional embodiment, the first agent module 22 further includes a first agent behavior rule model training unit, which is used to:

[0118] Obtain historical environmental development data of the target remediation area as training samples;

[0119] The training samples are used to train the improved ensemble learning algorithm to obtain a trained first agent behavior rule model.

[0120] In an optional embodiment, the second agent module 23 further includes a second agent behavior rule model training unit, which is used to:

[0121] Acquire historical land use data of the target remediation area; wherein the historical land use data of the target remediation area includes historical land use data of the remediation-implemented area and historical land use data of the remediation-unimplemented area;

[0122] The historical land use data of the areas where remediation has been implemented is used as positive training samples, and the historical land use data of the areas where remediation has not been implemented is used as negative training samples, and the long short-term memory network is trained to obtain a trained second agent behavior rule model.

[0123] Specifically, the second intelligent entity that agrees to the renovation includes the area of ​​the residential area that agrees to be renovated and the renovation willingness of the second intelligent entity that agrees to be renovated.

[0124] Furthermore, the rural residential area improvement area identification module 24 includes:

[0125] The second agent screening unit is used to screen out the second agents that meet the requirements and agree to the renovation according to the preset contiguous scale constraint of the residential area renovation and the renovation willingness of the second agents that agree to the renovation;

[0126] A screening seed determination unit is used to select screening seeds according to a constraint threshold if the residential area owned by the second agent that agrees to remediate in the target remediation area is larger than the contiguous scale constraint; wherein the screening seed is a second agent among the second agents that agree to remediate whose remediation willingness is greater than the constraint threshold;

[0127] The regional expansion unit is used to sort all the screening seeds according to the degree of remediation willingness, and sequentially expand the sorted screening seeds to the second intelligent agents in the eight neighborhoods until the expansion area threshold is met; wherein, the edge level is given to the second intelligent agents in the eight neighborhoods during the expansion process;

[0128] The regional optimization selection unit is used to optimize the selection of the expanded second intelligent agent according to the edge level of the second intelligent agent until the renovation target constraint is met, and the renovation scale and spatial layout of the settlements in the target renovation area in the target year are obtained.

[0129] It should be noted that the rural residential area remediation area identification system provided in the embodiment of the present invention is used to execute all the process steps of the rural residential area remediation area identification method of the above embodiment. The working principles and beneficial effects of the two correspond one to one, and thus will not be repeated here.

[0130] The embodiment of the present invention also provides a terminal device, such as Fig. 9 As shown, Fig. 9 3 is a block diagram of a terminal device provided in an embodiment of the present invention. The terminal device includes a processor 31, a memory 32, and a computer program stored in the memory 32 and configured to be executed by the processor 31. When the processor 31 executes the computer program, the rural residential area improvement area identification method described in any of the above embodiments is implemented.

[0131] In addition, an embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the rural residential area improvement area identification method as described in any of the above embodiments.

[0132] When the processor 31 executes the computer program, the steps in the above-mentioned rural residential area improvement area identification method embodiment are implemented, for example: Figure 1 Alternatively, when the processor 31 executes the computer program, the functions of each module in the above-mentioned rural residential area improvement area identification system embodiment are realized, for example Figure 8 The functions of each module of the rural settlement remediation area identification system are shown.

[0133] Preferably, the computer program can be divided into one or more modules / units, which are stored in the memory 32 and executed by the processor 31 to implement the present invention. The one or more modules / units can be a series of computer program instruction segments that can implement specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.

[0134] The processor 31 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor, or the processor 31 can be any conventional processor. The processor 31 is the control center of the terminal device, and various parts of the terminal device are connected using various interfaces and lines.

[0135] The memory 32 mainly includes a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function, etc., and the data storage area can store related data, etc. In addition, the memory 32 can be a high-speed random access memory, or a non-volatile memory, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, and a flash card (Flash Card), etc., or the memory 32 can also be other volatile solid-state storage devices.

[0136] It should be noted that the above terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that Fig. 9 The structural block diagram shown is only a structural example of the above-mentioned terminal device and does not constitute a structural limitation of the above-mentioned terminal device. The above-mentioned terminal device may include more or fewer components than shown in the figure, or combine certain components, or different components.

[0137] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for identifying rural residential area renovation areas, characterized in that: include: Acquire decision-making environment data of the target remediation area; wherein the decision-making environment data includes environmental development data and land use data; Based on the environmental development data, a first agent behavior rule model is used to obtain a target constraint for residential renovation in the target renovation area in the target year; Based on the land use data, obtaining a second agent that agrees to conduct remediation in the target remediation area in the target year through a second agent behavior rule model; According to the target constraint of settlement renovation, the renovation willingness of the second intelligent body and the preset scale constraint of settlement renovation, the target renovation area is optimized and selected through the seed region growing algorithm to obtain the scale and spatial layout of settlement renovation in the target renovation area in the target year.

2. The rural residential area improvement area identification method according to claim 1, characterized in that: The first agent behavior rule model is obtained by training through the following steps: Obtain historical environmental development data of the target remediation area as training samples; The training samples are used to train the improved ensemble learning algorithm to obtain a trained first agent behavior rule model.

3. The rural residential area improvement area identification method according to claim 1, characterized in that: The second agent behavior rule model is trained by the following steps: Acquire historical land use data of the target remediation area; wherein the historical land use data of the target remediation area includes historical land use data of the remediation-implemented area and historical land use data of the remediation-unimplemented area; The historical land use data of the areas where remediation has been implemented is used as positive training samples, and the historical land use data of the areas where remediation has not been implemented is used as negative training samples, and the long short-term memory network is trained to obtain a trained second agent behavior rule model.

4. The rural residential area improvement area identification method according to claim 1, characterized in that: The second intelligent entity that agrees to renovate includes the area of ​​the residential area that agrees to be renovated and the renovation willingness of the second intelligent entity that agrees to renovate.

5. The rural residential area improvement area identification method according to claim 4, characterized in that: According to the settlement renovation target constraint, the renovation willingness of the second intelligent agent and the preset settlement renovation contiguous scale constraint, the target renovation area is optimized and selected through the seed region growing algorithm to obtain the settlement renovation scale and spatial layout in the target renovation area in the target year, including: According to the preset scale constraints of the contiguous area of ​​residential area renovation and the renovation willingness of the second intelligent agent who agrees to the renovation, the second intelligent agent that meets the requirements and agrees to the renovation is selected; If the settlement area owned by the second agent that agrees to conduct the renovation in the target renovation area is larger than the contiguous scale constraint, a seed is selected according to the constraint threshold; wherein the seed is a second agent among the second agents that agree to conduct the renovation whose renovation willingness is greater than the constraint threshold; All seeds are sorted according to the degree of willingness to be regulated, and the sorted seeds are expanded to the second intelligent agent in the eight neighborhoods in turn until the expansion area threshold is met; wherein, the edge level is given to the second intelligent agent in the eight neighborhoods during the expansion process; The expanded second agent is optimized and selected according to the edge level of the second agent until the renovation target constraint is met, and the renovation scale and spatial layout of the settlements in the target renovation area in the target year are obtained.

6. The rural residential area improvement area identification method according to claim 5, characterized in that: The constraint threshold is calculated by the following formula: Wherein, T1 is the constraint threshold, Persential(a,b) is the percentile function used to obtain the bth percentile of a, Prop2 is the remediation willingness of the second agent who agrees to the remediation, and R is the target remediation area. is the settlement area of ​​the second agent that agrees to be renovated, Constrained by the remediation objectives.

7. The rural residential area improvement area identification method according to claim 6, characterized in that: The expansion area threshold is calculated by the following formula: Among them, T E is the expansion area threshold.

8. A rural residential area improvement area identification system, characterized in that: include: A decision environment module, used to obtain decision environment data of the target remediation area; wherein the decision environment data includes environmental development data and land use data; A first agent module is used to obtain, based on the environmental development data, a target constraint for residential renovation in the target renovation area in the target year through a first agent behavior rule model; A second agent module is used to obtain, based on the land use data, a second agent that agrees to conduct remediation in the target remediation area in the target year through a second agent behavior rule model; The settlement remediation area identification module is used to optimize the selection of the target remediation area through a seed region growing algorithm according to the settlement remediation target constraints, the remediation willingness of the second intelligent agent and the preset settlement remediation contiguous scale constraints, so as to obtain the settlement remediation scale and spatial layout in the target remediation area in the target year.

9. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for identifying rural residential area renovation areas as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the rural residential area improvement area identification method as described in any one of claims 1 to 7.

Images