Urban update region and time sequence delimiting method and system based on network evaluation data

Abstract

The invention discloses a city update area and time sequence delimiting method and system based on network evaluation data. The method comprises the steps that the network evaluation data and a city update text are collected; negative feedback evaluation is screened out, and city problems are identified; analyzing damage or aging of the building surface; dividing the urban problem point set into a point type facility updating area, a linear street style updating area and a surface type environment improvement area, calculating an updating time sequence order of all the areas, and dividing updating time sequence stages; through a Word2Vec feature vectorization algorithm, respectively outputting the core features of the city update region and the city update text as vector sets; calculating a cosine similarity calculation formula to obtain a similarity value of each pair of vector sets, and taking a maximum value to obtain a matching result with the highest similarity; negative feedback evaluation quantities before and after updating are collected and compared, if the negative feedback evaluation quantities are increased, the updating of the area is judged to be invalid, and the urban problem identification step is automatically returned; according to the invention, accurate planning and efficient and orderly implementation of city updating are effectively realized.

Description

Technical Field

[0001] The present invention relates to a method and system for delineating urban renewal areas and time series, and in particular to a method and system for delineating urban renewal areas and time series based on network evaluation data. Background Art

[0002] In recent years, with the development of cities and the changes in residents' needs, the traditional urban renewal model faces many challenges. Especially in the process of urban renewal area and time sequence delineation, the traditional method is based on data collection and experience judgment, relying on field surveys and on-site mapping, which increases a lot of time and labor costs. At the same time, the public participation survey method of traditional questionnaire surveys has a small sample size and is difficult to collect, which reduces the accuracy and speed of urban problems and renewal area identification, making it difficult to effectively solve practical problems, and may even cause resource waste or new contradictions. With the widespread application of online evaluation data, especially on social media and comment platforms, residents' feedback has gradually become an important reference information for urban renewal. However, the current use of these data is still relatively preliminary, mainly relying on simple statistics and analysis, and lacks systematic and intelligent processing methods. Traditional renewal area delineation is usually based on administrative divisions or historical issues, and fails to effectively reflect the real-time changes in residents' needs. With the help of artificial intelligence and big data technologies, especially the application of natural language processing and deep learning models, it has become more feasible to delineate urban renewal areas and time sequences based on online evaluation data. Summary of the invention

[0003] Purpose of the invention: The purpose of the present invention is to provide a method and system for delineating urban renewal areas and timing based on network evaluation data, which can improve the accuracy of urban problem identification and the accuracy of urban renewal plan matching.

[0004] Technical solution: The method for delineating urban renewal areas and time series based on network evaluation data described in the present invention comprises:

[0005] (1) Collecting online evaluation data and urban renewal text; the online evaluation data includes evaluation text, evaluation photos, number of likes and time information, and the online evaluation data carries location information and time information;

[0006] (2) Based on the collected online evaluation data, city-related semantics extraction and sentiment semantics analysis are performed to screen out negative feedback evaluations and identify urban problems;

[0007] (3) Use image preprocessing technology to preprocess the collected evaluation photos, identify abnormal buildings through the residual network model, and analyze the damage or aging of the building surface;

[0008] (4) The urban problems and the buildings with abnormalities are combined to generate an urban problem point set, which is then divided into point-type facility renewal areas, linear streetscape renewal areas, and surface-type environmental improvement areas. The renewal time series of all areas is calculated and the renewal time series stages are divided.

[0009] (5) Using the Word2Vec feature vectorization algorithm, the core features of the urban renewal area and the urban renewal text are output as vector sets respectively; the core features include function, location, core issues and facility type; the cosine similarity calculation formula is calculated to obtain the similarity value of each pair of vector sets, and the maximum value is taken to obtain the matching result with the highest similarity;

[0010] (6) Collect and compare the number of negative feedback evaluations before and after the update. If the number of negative feedback evaluations increases, it is determined that the update of the area is invalid and return to the step of identifying urban problems.

[0011] Furthermore, the step (1) comprises:

[0012] Through social media, we can obtain online evaluation data with location information and time information, including evaluation text, evaluation photos, and the number of likes. We can process the obtained data in real time through edge computing devices and store the processed data.

[0013] The urban renewal texts obtained from local planning institutes are converted into text format and cleaned, the urban renewal content is identified from the cleaned text, and the text of the content is converted into structured data and saved.

[0014] Furthermore, the step (2) comprises:

[0015] The collected evaluation texts are cleaned and segmented to obtain vectors of fixed dimensions and position codes. The processed vectors and position codes are input into the Transformer model, and the probability distribution representing the sentiment polarity is output. The texts with negative sentiment probability greater than the preset threshold are regarded as negative feedback evaluations. The semantic relationship between words is captured through the self-attention mechanism of the Transformer model, and the text corresponding to the negative feedback evaluation is associated with the context to identify the key semantic information related to the urban issues targeted by the negative feedback evaluation.

[0016] The key semantic information, time information and location information corresponding to the screened negative feedback evaluations are input into the Latent Dirichlet Allocation generative topic model LDA to identify the topics of the texts at the same time node and at the same location. Topics with a frequency higher than the preset number are regarded as persistent urban problems in the area, and topics with a frequency lower than the preset number are regarded as occasional urban problems.

[0017] Combine the name, date, spatial location and number of negative feedback evaluations of the sporadic urban problems to build a case library of historical urban sporadic problems and network evaluations, and use the data in the case library to train the Prophet trend prediction model to predict the probability and area of ​​occurrence of sporadic urban problems in the next year, and inspect areas with probabilities higher than a preset ratio to determine the spatial location of sporadic urban problems;

[0018] The case library includes the names, dates, spatial locations and numbers of negative feedback evaluations of occasional urban problems.

[0019] Furthermore, the step (3) comprises:

[0020] The collected evaluation photos are preprocessed using image preprocessing technology, wherein the preprocessing includes: using a median filter algorithm to remove random noise in the image, applying a histogram equalization technique to improve the contrast and brightness of the image and enhance the building details, and performing geometric and color correction;

[0021] A preset proportion is divided from the preprocessed evaluation photos, and the abnormalities of the buildings therein are marked to obtain an abnormal image data set, where the abnormalities are damage or aging; a pre-trained ResNet residual network model is called through a deep learning framework, and supervised learning is performed using the abnormal image data set of the buildings to fine-tune the model parameters; the fine-tuned model is used to detect the abnormal parts of the buildings in all the preprocessed evaluation photos, identify the damaged or aged areas, and locate them in combination with the spatial coordinates;

[0022] Generate 3D point cloud data of buildings with anomalies through field measurements.

[0023] Furthermore, the step (4) comprises:

[0024] The Kruskal algorithm is used to generate a minimum spanning tree from the urban problem point set. By setting the minimum length threshold and the maximum density threshold, the linear structures that meet the conditions are screened out and marked as linear streetscape renewal areas.

[0025] The DBSCAN density clustering algorithm is used to cluster the remaining point sets in the urban problem point set, and the densely populated area with the minimum number of points in the neighborhood is marked as the area of ​​surface environmental improvement, and the remaining isolated point sets are marked as point-type facility renewal areas.

[0026] The urgency weighted index of each area is calculated by combining the number of negative feedback evaluations, the number of visitors and the population density data, and the update time series ranking of all areas is obtained, which is then divided into several update time series stages using the natural breakpoint method.

[0027] Furthermore, generating a minimum spanning tree from the urban problem point set by using the Kruskal algorithm includes:

[0028] The identified abnormal points are used as nodes v i And the coordinates are expressed as (x i ,y i ), let V = {v1,v2,…,v n} is a node set, n is the total number of nodes, i = 1, 2, ..., n;

[0029] Calculate the weights of all edges, let E = {e1, e2, …, e m} is the set of edges, m is the total number of edges, and each edge e k Connect two nodes and represent them as e k =(v i ,v j ), j = 1, 2, ..., n; k = 1, 2, ..., m; the Manhattan distance between nodes is used as the weight of each edge w(e k ):

[0030] w(e k )=|x i -x j |+|y i -y j |;

[0031] Sort the edge set E, and the sorted edge set is sorted in ascending order of weight:

[0032] E ′ ={e′1,e′2,…,e′ m}

[0033] w(e′1)≤w(e′2)≤…≤w(e′ m )

[0034] Initialize the spanning tree and set the edge set to be empty Add each edge that meets the conditions to the set T. The final set T contains n-1 edges connecting all nodes, and the total weight is the smallest;

[0035] Gradually add edges, perform loop checks starting from the edge with the smallest weight, and use the Union-Find data structure to check the sorted edge e′ k The two nodes v connected i and v j Whether they belong to different sets; if v i and v j If they are not in the same set, add this edge to set T and merge v i and v j The set in which it is located; if vi and v j If they are already in the same set, skip this edge to ensure that no loop is formed;

[0036] By setting the minimum length threshold and the maximum density threshold, the linear structures meeting the conditions are screened out and marked as linear streetscape renewal areas, including:

[0037] Let the minimum length threshold be L min , retain all length l ≥ L in the minimum spanning tree MST min , and filter out the linear structure formed by the edges;

[0038] Let the maximum density threshold be D max Used to control the density of linear structures and calculate the total length L and node density D of linear structures nodes , edge density D edges and the comprehensive density D com , assuming that the linear structure consists of m edges e1,e2,…,e m The length of each side is l(e1),l(e2),…,l(e m ), then the total length Node Density Edge Density Comprehensive density D com =α×D nodes +β×D edges , where α represents the node density D nodes In the comprehensive density D com The weight in , β represents the edge density D edges In the comprehensive density D com The weights in D com <D max The linear structure is marked as a linear streetscape renewal area.

[0039] Furthermore, the DBSCAN density clustering algorithm is used to perform cluster analysis on the remaining point sets, and the densely populated area with the minimum number of points in the neighborhood is marked as a surface environment improvement area, and the remaining isolated point sets are marked as point-type facility update areas, including:

[0040] Let the set of all remaining points p be D, that is, D = {p1, p2, ..., p n}; Let the neighborhood radius be ∈, and all points q whose distance p does not exceed the neighborhood radius ∈ constitute the neighborhood N ∈ (p), that is:

[0041] N ∈ (p)={q∈D|distance(p,q)≤∈}

[0042] Let the threshold of the number of points in the neighborhood be MinPts. If the number of q points m meets the predetermined threshold, that is, m ≥ MinPts, then p is considered to be a core point;

[0043] p and other points in its neighborhood expand to form a dense region C, |C| is the number of points in C, S min is the threshold of C; if |C|≥S min It is marked as a surface environmental remediation area; if |C| min It is marked as a point-type facility update area.

[0044] Furthermore, the weighted urgency index of each area is calculated by combining the number of negative feedback evaluations, the number of visitors and the population density data to obtain the update time sequence ranking of all areas, including:

[0045] Define the weights of negative feedback quantity, visitor number and population density data as w1, w2, w3 respectively;

[0046] Normalize the number of negative feedback, number of visitors, and population density data;

[0047] Use the standardized data to calculate the urgency weighted index I of each update area:

[0048] I = w1×Norm(number of negative feedback)+w2×Norm(number of visitors)+w2×Norm(population density)

[0049] Norm(·) represents the standardized value;

[0050] The calculated urgency weighted index I of the update area is sorted by size to form an update time sequence table. The higher the index, the more urgent the update demand of the area.

[0051] Furthermore, the core features of the urban renewal area and the urban renewal text are output as vector sets respectively through the Word2Vec feature vectorization algorithm, including:

[0052] Set the characteristics of the urban renewal area to T region , including function, location, core issues and facility types, urban renewal text is T text :

[0053] T region ={Function, Location, Core Issue, Facility Type}

[0054] T text ={Urban renewal text content}

[0055] Suppose each word x in the core feature a The vector is represented as v(x a ), T​region Convert to V region , T text Convert to V text ,Right now:

[0056] v(x a )=Word2Vec(x a )

[0057]

[0058] Among them, λ a is word x a The weight, μ b is word x b The weight of

[0059] The cosine similarity calculation formula is used to obtain the similarity value of each pair of vector sets, and the maximum value is taken to obtain the matching result with the highest similarity, including: using the feature vector V region and the text vector V text Calculate similarity via cosine similarity:

[0060]

[0061] The calculated similarity value indicates the degree of match between the characteristics of the area and a certain urban renewal text strategy; the higher the similarity, the more suitable the urban renewal text strategy is for the renewal of the area.

[0062] Based on the same inventive concept, the present invention also provides a system for delineating urban renewal areas and time series based on network evaluation data, comprising:

[0063] A collection module, used to collect online evaluation data and urban renewal text; the online evaluation data includes evaluation text, evaluation photos and the number of likes, and the online evaluation data carries location information and time information;

[0064] The urban problem identification module is used to extract urban related semantics and perform sentiment semantic analysis based on the collected online evaluation data, filter out negative feedback evaluations, and identify urban problems;

[0065] The anomaly recognition module is used to pre-process the collected evaluation photos using image pre-processing technology, identify abnormal buildings through the residual network model, and analyze the damage or aging of the building surface;

[0066] The division module is used to merge urban problems and abnormal buildings to generate an urban problem point set, divide the urban problem point set into point-type facility renewal areas, linear street style renewal areas and surface environment improvement areas, and calculate the renewal time sequence of all areas and divide the renewal time sequence stages;

[0067] A calculation module is used to output the core features of the urban renewal area and the urban renewal text as vector sets respectively through the Word2Vec feature vectorization algorithm; the core features include function, location, core issues and facility type; calculate the cosine similarity calculation formula to obtain the similarity value of each pair of vector sets, and take the maximum value to obtain the matching result with the highest similarity;

[0068] The verification module is used to collect and compare the number of negative feedback evaluations before and after the update. If the number of negative feedback evaluations increases, it is determined that the update of the area is invalid and returns to the urban problem identification module.

[0069] Beneficial effects: Compared with the prior art, the present invention reduces the time cost of regional analysis and delineation by about 30% by integrating network evaluation data, problem point set analysis and intelligent algorithms, and reduces resource consumption by about 25% through automated data processing and real-time calculation. With the help of automated analysis technology, the present invention shortens the urban renewal regional analysis work that traditionally takes 4-6 weeks to 1 week, which significantly shortens the planning and implementation cycle, optimizes the allocation and utilization efficiency of resources, and reduces the time cost and resource consumption of urban renewal regional and time series delineation; compared with traditional methods, the present invention further improves the accuracy of urban problem identification by about 50% through the application of Transformer-based natural language processing algorithms and LDA generative probability models, and improves the accuracy of strategy matching by about 45% through cosine similarity calculation. Based on the deep semantic analysis of negative feedback data and the DBSCAN density clustering algorithm, the present invention can accurately identify potential persistent and sporadic problem areas in the city, and through intelligent strategy matching, ensure that each renewal area can get the most suitable solution, which improves the overall effect of urban renewal and the scientific nature of planning, and improves the accuracy of urban problem identification and the accuracy of urban renewal scheme matching; compared with traditional methods, the present invention uses image preprocessing technology and the Word2Vec feature vectorization algorithm to scientifically divide the area-type environmental improvement area and the point-type facility renewal area, which expands the scope of application of urban renewal schemes in different regional types by about 35% and improves the design depth by about 25%. This improves the flexibility of urban renewal planning to meet the needs of diversified urban development and expands the scope of application and design depth of urban renewal schemes in different regional types. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 is a flow chart of the method of the present invention;

[0071] Figure 2 A schematic diagram of building anomaly recognition according to an embodiment of the present invention;

[0072] Figure 3A schematic diagram of urban renewal area and time sequence delineation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0073] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in combination with the accompanying drawings and embodiments. 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.

[0074] Embodiment 1: As attached Figure 1 As shown, the method for delineating urban renewal areas and time series based on network evaluation data in this embodiment includes:

[0075] (1) Collecting online evaluation data and urban renewal text; the online evaluation data includes evaluation text, evaluation photos and the number of likes, and the online evaluation data carries location information and time information;

[0076] (2) Based on the collected online evaluation data, city-related semantics extraction and sentiment semantics analysis are performed to screen out negative feedback evaluations and identify urban problems;

[0077] (3) Use image preprocessing technology to preprocess the collected evaluation photos, identify abnormal buildings through the residual network model, and analyze the damage or aging of the building surface;

[0078] (4) The urban problems and the buildings with abnormalities are combined to generate an urban problem point set, which is then divided into point-type facility renewal areas, linear streetscape renewal areas, and surface-type environmental improvement areas. The renewal time series of all areas is calculated and the renewal time series stages are divided.

[0079] (5) Using the Word2Vec feature vectorization algorithm, the core features of the urban renewal area and the urban renewal text are output as vector sets respectively; the core features include function, location, core issues and facility type; the cosine similarity calculation formula is calculated to obtain the similarity value of each pair of vector sets, and the maximum value is taken to obtain the matching result with the highest similarity;

[0080] (6) Collect and compare the number of negative feedback evaluations before and after the update. If the number of negative feedback evaluations increases, it is determined that the update of the area is invalid and return to the step of identifying urban problems.

[0081] Specifically, step (1) collection of online evaluation data and urban renewal text

[0082] The data collection module is connected to the social media API interface to obtain online evaluation data with location information, including evaluation text, evaluation photos, number of likes, and time information. The data is processed and transmitted in real time through the NVIDIA Jetson Nano edge computing device. 500 sets of urban renewal texts are collected by using the ScanSnap iX1600 scanner and stored in a big data management platform that supports structured retrieval.

[0083] Step (2) Identification of urban problems based on negative feedback evaluation

[0084] Filter negative feedback evaluations based on evaluation texts to identify persistent and sporadic urban problems. Through the Transformer model based on the self-attention mechanism, city-related semantics extraction and sentiment semantics analysis are performed on the evaluation text data to filter out negative feedback evaluations. Further identification of persistent urban problems is carried out, and the filtered key semantics, time information and location information are used as input parameters. The LDA generative topic model is used to identify frequently occurring urban problems in the region. For sporadic urban problems, a case library of historical urban sporadic problems-network evaluation associations is constructed, and the Prophet trend prediction model is trained. The negative feedback evaluation is used as the case library input to predict the probability of occurrence and risk areas of sporadic urban problems. A drone equipped with a LiDAR laser radar sensor is used to inspect the risk area to determine the problem points.

[0085] Step (3) Architectural style diagnosis based on residual network model

[0086] The evaluation images are denoised, enhanced and corrected through image preprocessing technology. The pre-trained ResNet residual network model is called through deep learning frameworks such as PyTorch, and the abnormal parts of the building are detected by fine-tuning the training model. The inspection vehicle equipped with a light projector and a high-resolution camera is used to obtain 3D point cloud data of buildings in abnormal locations through structured light technology to analyze the deformation and damage of the building surface.

[0087] Step (4) Delineation of urban renewal areas and timelines

[0088] The urban problem point set is divided into point-type facility renewal areas, linear street appearance renewal areas, and surface environment improvement areas. The minimum spanning tree is generated from the urban problem point set by the Kruskal algorithm. By setting the minimum length threshold and the maximum density threshold, the linear structure that meets the conditions is screened out and marked as the linear street appearance renewal area. The DBSCAN density clustering algorithm is used to perform cluster analysis on the remaining point sets, and the areas with the minimum number of points in the neighborhood are marked as surface environment improvement areas, and the remaining isolated point sets are marked as point-type facility renewal areas. The urgency weighted index of each renewal area is calculated by combining the number of negative feedbacks, the number of visitors, and the population density data to obtain the renewal time sequence ranking of all areas, and then divided into three renewal time sequence stages (update within 5 years, update within 10 years, and update within 15 years) by the natural breakpoint method. Figure 3 shown.

[0089] Step (5) Urban renewal strategy matching

[0090] Through the Word2Vec feature vectorization algorithm, the core features of the urban renewal area (function, location, core issues, facility type) and the urban renewal text are output as vector sets respectively. The cosine similarity calculation formula is calculated to obtain the similarity value of each pair of vector sets, and the maximum value is taken to obtain the matching result with the highest similarity.

[0091] Step (6) Urban renewal effect verification and iteration

[0092] Collect and compare the number of negative feedback evaluations before and after the update. If the number of negative feedback evaluations increases, it is considered that the update of the area is ineffective, and automatically return to step (2) to re-identify and inspect the problem.

[0093] This embodiment may also include step (7): digital sand table display of urban renewal plan

[0094] Build a sandbox for the strategic development of urban renewal. Use augmented reality (AR) and virtual reality (VR) devices to display the overall layout and strategic planning of urban renewal. Use the Tableau data visualization platform to display urban problems based on network evaluation data analysis in the form of heat maps and trend maps, and use the MATSim time series simulation tool to perform an animated demonstration of the urban renewal time series.

[0095] Furthermore, in step (1), the DGX computer workstation with built-in data acquisition module is connected to the social media API interface to obtain the network evaluation data of the target city with location information and time information, including evaluation text, URL of evaluation photo, and number of likes. The NVIDIA Jetson Nano edge computing device is used to process the acquired data in real time and store the processed data in the workstation.

[0096] 500 sets of urban renewal texts obtained from local planning institutes were scanned at high resolution using ScanSnap iX1600 scanners, and the scanned results were converted into searchable text format using OCR technology. Text data was cleaned and formatted, and natural language processing (NLP) technology and keyword matching were used to identify content related to urban renewal from the cleaned text. Specific information was extracted, and the text was converted into structured data and uploaded and stored in Google BigQuery, a big data management platform that supports structured retrieval.

[0097] In step (2), the evaluation text obtained in step 1 is cleaned and segmented using a DGX computer workstation, and the words are converted into vectors of fixed dimensions using a pre-trained GloVe word embedding model. The processed vector sequence and position encoding are input into the Transformer model, and the probability distribution representing the sentiment polarity is output. Texts with a negative sentiment probability greater than 0.7 are selected as negative feedback evaluations. The semantic relationship between words is captured through the self-attention mechanism of the Transformer model, and the text related to the negative feedback evaluation is associated with its context, identifying the key semantic information related to the specific urban issues targeted by the negative feedback evaluation.

[0098] The screened key semantic information, time information, and location information are further input into the LDA generative topic model to identify text topics at the same time node and at the same location. Topics that appear more than 10 times are considered persistent urban problems in the area, and topics that appear less than 5 times are considered occasional urban problems.

[0099] The name, date, spatial location, and number of negative feedback evaluations of sporadic urban problems are combined to build a case library of historical urban sporadic problems-network evaluation associations. The data is fitted into the model to train the Prophet trend prediction model to predict the probability of occurrence and risk areas of sporadic urban problems in the next year. Drones equipped with LiDAR laser radar sensors are used to conduct field inspections of risk areas with a probability higher than 50% to determine the spatial location of sporadic urban problems.

[0100] In step (3), the evaluation photos obtained in step 1 are preprocessed using a DGX computer workstation, including using a median filter algorithm to remove random noise in the image; applying a histogram equalization technique to improve the contrast and brightness of the image and enhance the details of the building; and performing geometric and color correction to ensure the authenticity and accuracy of the image.

[0101] We selected 10% of the processed photos and manually annotated the abnormal buildings. We used the PyTorch deep learning framework to call the pre-trained ResNet residual network model, and used the labeled abnormal building image dataset for supervised learning to fine-tune the model parameters. We used the fine-tuned model to detect the abnormal parts of the buildings in all the evaluated photos, identify the areas with damage, aging or other problems, and locate them in combination with spatial coordinates, such as Figure 2 shown.

[0102] Inspection vehicles equipped with light projectors and high-resolution cameras are used to project light of known patterns in combination with structured light technology to calculate the deformation of abnormal building surfaces. High-precision three-dimensional point cloud data of abnormal buildings is generated through field measurements to analyze the deformation and damage of the building surface.

[0103] In step (4), the minimum spanning tree is generated from the urban problem point set by the Kruskal algorithm, which means:

[0104] (411) The identified urban problem point set (buildings, facilities and other areas in the city that need to be updated) is used as node v i And the coordinates are expressed as (x i ,y i ). Let V = {v1,v2,…,v n} is a node collection.

[0105] (412) Calculate the weights of all edges. Let E = {e1, e2, …, e m} as a set of edges, each edge e k Connect two nodes and represent them as e k =(v i ,v j ), the Manhattan distance between nodes is used as the weight of each edge w(e k ):

[0106] w(e k )=|x i -x j |+|y i -y j |.

[0107] (413) Sort the edge set E. The sorted edge set is sorted in ascending order of weight:

[0108] E ′ ={e′1,e′2,…,e′ m}

[0109] w(e′1)≤w(e′2)≤…≤w(e′ m )

[0110] (414) Initialize the spanning tree. Set the edge set to be empty Add each edge that meets the conditions to the set T. The final set T contains n-1 edges connecting all nodes, and the total weight is the smallest.

[0111] (415) Gradually add edges and perform loop check from the edge with the smallest weight, that is, use the Union-Find data structure to check the sorted edge e′ k The two nodes v connected i and v j Whether they belong to different sets. i and v j If they are not in the same set, add this edge to set T and merge v i and v j The set in which it is located; if v i and v j If they are already in the same set, the edge is skipped to ensure that no loop is formed.

[0112] By setting the minimum length threshold and the maximum density threshold, the linear structures that meet the conditions are screened out and marked as linear streetscape renewal areas, which are:

[0113] (421) Let the minimum length threshold be L min , retain all length l ≥ L in the minimum spanning tree (MST) min 's edges and filter out the linear structure formed by these edges.

[0114] (422) Let the maximum density threshold be D max Used to control the density of linear structures and calculate the total length L and node density D of linear structures nodes , edge density D edges and the comprehensive density D com , assuming that the linear structure consists of m edges e1,e2,…,e m The length of each side is l(e1),l(e2),…,l(e m ), then the total length Node Density Edge Density Comprehensive density D com =α×D nodes +β×D edges (α+β=1), where α represents the node density D nodes In the comprehensive density D com The weight in , β represents the edge density D edges In the comprehensive density D com The weights in D com <D max The linear structure is marked as a linear streetscape renewal area.

[0115] The DBSCAN density clustering algorithm is used to cluster the remaining point sets. The densely populated areas with the minimum number of points in the neighborhood are marked as area-type environmental improvement areas, and the remaining isolated point sets are marked as point-type facility update areas, which are:

[0116] (431) Let D be the set of all remaining points p, that is, D = {p1, p2, …, p n}. Let the neighborhood radius be ∈, and all points q whose distance p does not exceed the neighborhood radius ∈ constitute the neighborhood N ∈ (p), that is:

[0117] N ∈ (p)={q∈D|distance(p,q)≤∈}

[0118] (432) Let the threshold of the number of points in the neighborhood be MinPts. If the number of points m in q satisfies the predetermined threshold, that is, m ≥ MinPts, then p is considered to be a core point.

[0119] (433)p and other points in its neighborhood expand to form a dense region C, where |C| is the number of points in C, S min is the threshold of C. If |C|≥S min It is marked as a surface environmental remediation area; if |C| min It is marked as a point-type facility update area.

[0120] The number of points represents the number of points in a certain area, which is used to indicate the degree of density. The method for determining the dense area with the minimum number of points is:

[0121] Define the neighborhood and draw a circle with radius ∈ for each point in DBSCAN;

[0122] Define the minimum number of points MinPts, which is the minimum number of points that must be included in the specified area; when the neighborhood contains at least MinPts points, it is considered to be a core point and constitutes a dense area;

[0123] Mark the points whose number of points in the neighborhood is greater than or equal to MinPts as the core points of the dense area;

[0124] Classify the points within the range ∈ adjacent to the core point into the same cluster to form a cluster;

[0125] Identify and cluster dense areas to form a minimum point-count dense area.

[0126] By combining the number of negative feedbacks, the number of visitors, and the population density data in the network evaluation, the urgency weighted index of each update area is calculated, and the update time sequence ranking of all areas is obtained, which is:

[0127] ​(441) Define the weights of the number of negative feedbacks, number of visitors, and population density data as w1, w2, and w3 respectively.

[0128] (442) The data on the number of negative feedback, the number of visitors, and the population density are standardized (Norm(·) represents the standardized value) so that the above data can be compared on the same scale:

[0129]

[0130] (443) Use the standardized data to calculate the urgency weighted index I of each update area:

[0131] I = w1×Norm(number of negative feedback)+w2×Norm(number of visitors)+w2×Norm(population density)

[0132] (444) The calculated urgency weighted index I of the update area is sorted by size to form an update schedule. The higher the index, the more urgent the update demand of the area.

[0133] In step (5), the core features (function, location, core issues, facility type) of the urban renewal area and the urban renewal text are output as vector sets respectively through the Word2Vec feature vectorization algorithm, which means:

[0134] (51) Assume the characteristics of the urban renewal area as T region , including functions, locations, core issues and facility types, urban renewal text is T text :

[0135] T region ={Function, Location, Core Issue, Facility Type}

[0136] T text ={Urban renewal text content}

[0137] (52) Let each word x in the core feature a The vector is represented as v(x a ), T region Convert to V region , T text Convert to V text (λ a is word x a The weight, μ b is word x b The weight of

[0138] v(x a )=Word2Vec(x a )

[0139]

[0140]

[0141] Among them, λ a is word x a The weight, μ b is word x b The weight of

[0142] By calculating the cosine similarity formula, we can get the similarity value of each pair of vector sets, and take the maximum value to get the matching result with the highest similarity. region and the text vector V text Calculate similarity via cosine similarity:

[0143]

[0144] The calculated similarity value indicates the degree of match between the characteristics of the area and a certain urban renewal text strategy. The higher the similarity, the more suitable the urban renewal text strategy is for the renewal of the area.

[0145] Based on the same inventive concept, this embodiment also provides a system for delineating urban renewal areas and time series based on network evaluation data, including:

[0146] A collection module, used to collect online evaluation data and urban renewal text; the online evaluation data includes evaluation text, evaluation photos and the number of likes, and the online evaluation data carries location information and time information;

[0147] The urban problem identification module is used to extract urban related semantics and perform sentiment semantic analysis based on the collected online evaluation data, filter out negative feedback evaluations, and identify urban problems;

[0148] The anomaly recognition module is used to pre-process the collected evaluation photos using image pre-processing technology, identify abnormal buildings through the residual network model, and analyze the damage or aging of the building surface;

[0149] The division module is used to merge urban problems and abnormal buildings to generate an urban problem point set, divide the urban problem point set into point-type facility renewal areas, linear street style renewal areas and surface environment improvement areas, and calculate the renewal time sequence of all areas and divide the renewal time sequence stages;

[0150] A calculation module is used to output the core features of the urban renewal area and the urban renewal text as vector sets respectively through the Word2Vec feature vectorization algorithm; the core features include function, location, core issues and facility type; calculate the cosine similarity calculation formula to obtain the similarity value of each pair of vector sets, and take the maximum value to obtain the matching result with the highest similarity;

[0151] The verification module is used to collect and compare the number of negative feedback evaluations before and after the update. If the number of negative feedback evaluations increases, it is determined that the update of the area is invalid and returns to the urban problem identification module.

Claims

1. A method for delineating urban renewal areas and time series based on network evaluation data, characterized in that: include: (1) Collect online evaluation data and urban renewal texts; The online evaluation data includes evaluation text, evaluation photos and the number of likes, and the online evaluation data carries location information and time information; (2) Based on the collected online evaluation data, city-related semantics extraction and sentiment semantics analysis are performed to screen out negative feedback evaluations and identify urban problems; (3) Use image preprocessing technology to preprocess the collected evaluation photos, identify abnormal buildings through the residual network model, and analyze the damage or aging of the building surface; (4) The urban problems and the buildings with abnormalities are combined to generate an urban problem point set, which is then divided into point-type facility renewal areas, linear streetscape renewal areas, and surface-type environmental improvement areas. The renewal time series of all areas is calculated and the renewal time series stages are divided. (5) Using the Word2Vec feature vectorization algorithm, the core features of the urban renewal area and the urban renewal text are output as vector sets respectively; the core features include function, location, core issues and facility type; Calculate the cosine similarity calculation formula to obtain the similarity value of each pair of vector sets, and take the maximum value to obtain the matching result with the highest similarity; (6) Collect and compare the number of negative feedback evaluations before and after the update. If the number of negative feedback evaluations increases, it is determined that the update of the area is invalid and return to the step of identifying urban problems.

2. The method for delineating urban renewal areas and time series based on network evaluation data according to claim 1 is characterized in that: The step (1) comprises: Through social media, we can obtain online evaluation data with location information and time information, including evaluation text, evaluation photos, and the number of likes. We can process the obtained data in real time through edge computing devices and store the processed data. The urban renewal texts obtained from local planning institutes are converted into text format and cleaned, the urban renewal content is identified from the cleaned text, and the text of the content is converted into structured data and saved.

3. The method for delineating urban renewal areas and time series based on network evaluation data according to claim 1 is characterized in that: The step (2) comprises: The collected evaluation texts are cleaned and segmented to obtain vectors of fixed dimensions and position codes. The processed vectors and position codes are input into the Transformer model, and the probability distribution representing the sentiment polarity is output. The texts with negative sentiment probability greater than the preset threshold are regarded as negative feedback evaluations. The semantic relationship between words is captured through the self-attention mechanism of the Transformer model, and the text corresponding to the negative feedback evaluation is associated with the context to identify the key semantic information related to the urban issues targeted by the negative feedback evaluation. The key semantic information, time information and location information corresponding to the screened negative feedback evaluations are input into the Latent Dirichlet Allocation generative topic model LDA to identify the topics of the texts at the same time node and at the same location. Topics with a frequency higher than the preset number are regarded as persistent urban problems in the area, and topics with a frequency lower than the preset number are regarded as occasional urban problems. Combine the name, date, spatial location and number of negative feedback evaluations of the sporadic urban problems to build a case library of historical urban sporadic problems and network evaluations, and use the data in the case library to train the Prophet trend prediction model to predict the probability and area of ​​occurrence of sporadic urban problems in the next year, and inspect areas with probabilities higher than a preset ratio to determine the spatial location of sporadic urban problems; The case library includes the names, dates, spatial locations and numbers of negative feedback evaluations of occasional urban problems.

4. The method for delineating urban renewal areas and time series based on network evaluation data according to claim 1 is characterized in that: The step (3) comprises: The collected evaluation photos are preprocessed using image preprocessing technology, wherein the preprocessing includes: using a median filter algorithm to remove random noise in the image, applying a histogram equalization technique to improve the contrast and brightness of the image and enhance the building details, and performing geometric and color correction; A preset proportion is divided from the preprocessed evaluation photos, and the abnormalities of the buildings therein are marked to obtain an abnormal image data set, where the abnormalities are damage or aging; a pre-trained ResNet residual network model is called through a deep learning framework, and supervised learning is performed using the abnormal image data set of the buildings to fine-tune the model parameters; the fine-tuned model is used to detect the abnormal parts of the buildings in all the preprocessed evaluation photos, identify the damaged or aged areas, and locate them in combination with the spatial coordinates; Generate 3D point cloud data of buildings with anomalies through field measurements.

5. The method for delineating urban renewal areas and time series based on network evaluation data according to claim 1 is characterized in that: The step (4) comprises: The Kruskal algorithm is used to generate a minimum spanning tree from the urban problem point set. By setting the minimum length threshold and the maximum density threshold, the linear structures that meet the conditions are screened out and marked as linear streetscape renewal areas. The DBSCAN density clustering algorithm is used to cluster the remaining point sets in the urban problem point set, and the densely populated area with the minimum number of points in the neighborhood is marked as the area of ​​surface environmental improvement, and the remaining isolated point sets are marked as point-type facility renewal areas. The urgency weighted index of each area is calculated by combining the number of negative feedback evaluations, the number of visitors and the population density data, and the update time sequence ranking of all areas is obtained, which is then divided into several update time sequence stages through the natural breakpoint method.

6. The method for delineating urban renewal areas and time series based on network evaluation data according to claim 5 is characterized in that: The method of generating a minimum spanning tree from a set of urban problem points by using the Kruskal algorithm includes: The identified abnormal points are used as nodes v i And the coordinates are expressed as (x i ,y i ), let V = {v1,v2,…,v n } is a node set, n is the total number of nodes, i = 1, 2, ..., n; Calculate the weights of all edges, let E = {e1, e2, …, e m } is the set of edges, m is the total number of edges, and each edge e k Connect two nodes and represent them as e k =(v i ,v j ), j = 1, 2, ..., n; k = 1, 2, ..., m; the Manhattan distance between nodes is used as the weight of each edge w(e k ): w(e k )=|x i -x j +|y i -y j |; Sort the edge set E, and the sorted edge set is sorted in ascending order of weight: E′={e'1,e'2,…,e' m } w(e′1)≤w(e′2)≤…≤w(e' m ) Initialize the spanning tree and set the edge set to be empty Add each edge that meets the conditions to the set T. The final set T contains n-1 edges connecting all nodes, and the total weight is the smallest; Gradually add edges, perform loop checks starting from the edge with the smallest weight, and use the Union-Find data structure to check the sorted edge e′ k The two nodes v connected i and v j Whether they belong to different sets; if v i and v j If they are not in the same set, add this edge to set T and merge v i and v j The set in which it is located; if v i and v j If they are already in the same set, skip this edge to ensure that no loop is formed; By setting the minimum length threshold and the maximum density threshold, the linear structures meeting the conditions are screened out and marked as linear streetscape renewal areas, including: Let the minimum length threshold be L min , retain all length l ≥ L in the minimum spanning tree MST min , and filter out the linear structure formed by the edges; Let the maximum density threshold be D max Used to control the density of linear structures and calculate the total length L and node density D of linear structures nodes , edge density D edges and the comprehensive density D com , assuming that the linear structure consists of m edges e1,e2,…,e m The length of each side is l(e1),l(e2),…,l(e m ), then the total length Node Density Edge Density Comprehensive density D com =α×D nodes +β×D edges , where α represents the node density D nodes In the comprehensive density D com The weight in , β represents the edge density D edges In the comprehensive density D com The weights in D com <D max The linear structure is marked as a linear streetscape renewal area.

7. The method for delineating urban renewal areas and time series based on network evaluation data according to claim 5 is characterized in that: The DBSCAN density clustering algorithm is used to perform cluster analysis on the remaining point sets, and the densely populated area with the minimum number of points in the neighborhood is marked as the surface environment improvement area, and the remaining isolated point sets are marked as point facility update areas, including: Let the set of all remaining points p be D, that is, D = {p1, p2, ..., p n }; Let the neighborhood radius be ∈, and all points q whose distance p does not exceed the neighborhood radius ∈ constitute the neighborhood N ∈ (p), that is: N ∈ (p)={q∈D|distance(p,q)≤∈} Let the threshold of the number of points in the neighborhood be MinPts. If the number of q points m meets the predetermined threshold, that is, m ≥ MinPts, then p is considered to be a core point; p and other points in its neighborhood expand to form a dense region C, |C| is the number of points in C, S min is the threshold of C; if |C|≥S min It is marked as a surface environmental remediation area; if |C| min It is marked as a point-type facility update area.​ 8. The method for defining urban renewal areas and time series based on network evaluation data according to claim 5 is characterized in that: The weighted urgency index of each area is calculated by combining the number of negative feedback evaluations, the number of visitors and the population density data to obtain the update time sequence ranking of all areas, including: Define the weights of negative feedback quantity, visitor number and population density data as w1, w2, w3 respectively; Normalize the number of negative feedback, number of visitors, and population density data; Use the standardized data to calculate the urgency weighted index I of each update area: I = w1×Norm(number of negative feedback)+w2×Norm(number of visitors)+w2×Norm(population density) Norm(·) represents the standardized value; The calculated urgency weighted index I of the update area is sorted by size to form an update time sequence table. The higher the index, the more urgent the update demand of the area.

9. The method for delineating urban renewal areas and time series based on network evaluation data according to claim 1 is characterized in that: The core features of the urban renewal area and the urban renewal text are output as vector sets respectively through the Word2Vec feature vectorization algorithm, including: Set the characteristics of the urban renewal area to T region , including function, location, core issues and facility types, urban renewal text is T text : T region ={Function, Location, Core Issue, Facility Type} T text ={Urban renewal text content} Suppose each word x in the core feature a The vector is represented as v(x a ), T rion Convert to V region , T text Convert to V text ,Right now: v(x a )=Word2Vec(x a ) Among them, λ a is word x a The weight, μ b is word x b The weight of The cosine similarity calculation formula is used to obtain the similarity value of each pair of vector sets, and the maximum value is taken to obtain the matching result with the highest similarity, including: using the feature vector V region and the text vector V text Calculate similarity via cosine similarity: The calculated similarity value indicates the degree of match between the characteristics of the area and a certain urban renewal text strategy; the higher the similarity, the more suitable the urban renewal text strategy is for the renewal of the area.

10. A system for defining urban renewal areas and time series based on network evaluation data, characterized in that: include: The collection module is used to collect online evaluation data and urban renewal texts; The online evaluation data includes evaluation text, evaluation photos and the number of likes, and the online evaluation data carries location information and time information; The urban problem identification module is used to extract urban related semantics and perform sentiment semantic analysis based on the collected online evaluation data, filter out negative feedback evaluations, and identify urban problems; The anomaly recognition module is used to pre-process the collected evaluation photos using image pre-processing technology, identify abnormal buildings through the residual network model, and analyze the damage or aging of the building surface; The division module is used to merge urban problems and abnormal buildings to generate an urban problem point set, divide the urban problem point set into point-type facility renewal areas, linear street style renewal areas and surface environment improvement areas, and calculate the renewal time sequence of all areas and divide the renewal time sequence stages; A calculation module, for outputting the core features of the urban renewal area and the urban renewal text as vector sets respectively through a Word2Vec feature vectorization algorithm; the core features include functions, locations, core issues and facility types; Calculate the cosine similarity calculation formula to obtain the similarity value of each pair of vector sets, and take the maximum value to obtain the matching result with the highest similarity; The verification module is used to collect and compare the number of negative feedback evaluations before and after the update. If the number of negative feedback evaluations increases, it is determined that the update of the area is invalid and returns to the urban problem identification module.

Images