A city sidewalk environment evaluation research method based on multi-source data

Abstract

This invention discloses a research method for urban sidewalk environment assessment based on multi-source data, comprising the following research steps: Based on research data, current environmental assessment research indicators are summarized; based on NLP-text similarity analysis, the indicators are classified and analyzed to propose primary indicators; urban sidewalk environment assessment data is selected; based on sentiment analysis-binary clustering, the indicators are initially clustered and then merged and screened as secondary indicators; the selected secondary indicators are then classified into primary indicators. This invention obtains more suitable sidewalk environment assessment indicators by merging, screening, and clustering sidewalk environment indicators in advance; then, road network data of the sidewalk environment is selected, and environmental analysis and evaluation are performed based on the indicators; finally, the sidewalk environment is divided into different orientations, and suggestions are made for the environment of different orientations, making the sidewalk environment analysis more accurate and helping to improve the urban sidewalk environment.

Description

Technical Field

[0001] This invention relates to the field of sidewalk environment research technology, specifically a method for urban sidewalk environment assessment based on multi-source data. Background Technology

[0002] With the acceleration of urbanization and the continuous increase in car ownership, urban traffic has become primarily motor vehicle-oriented. In this process, problems such as traffic congestion, competition between pedestrians and vehicles for road space, air pollution, and noise pollution have emerged one after another, becoming unique "urban diseases" that seriously threaten public health and reduce the willingness to walk. Therefore, improving the urban built environment, finding a balance between urban development and public interests, and seeking sustainable urban development have become key areas of focus for modern urban development.

[0003] However, the current evaluation of urban sidewalk environment has few indicators and the classification is not detailed enough. It is only a general evaluation and the evaluation of urban sidewalk environment is not accurate enough, making it difficult to make reasonable suggestions for sidewalk environment. Summary of the Invention

[0004] This invention provides a research method for urban sidewalk environment assessment based on multi-source data, which can effectively solve the problems mentioned in the background art: the current assessment of urban sidewalk environment has few indicators and the classification is not detailed enough. It is only a general assessment, which is not accurate enough for the assessment of urban sidewalk environment and makes it difficult to make reasonable suggestions for sidewalk environment.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a research method for urban sidewalk environmental assessment based on multi-source data, comprising the following research steps:

[0006] S1. Based on research data, summarize current environmental assessment research indicators, classify and analyze the indicators based on NLP-text similarity analysis, and propose primary indicators;

[0007] S2. Select urban sidewalk environment assessment data, and based on sentiment analysis-binary clustering method, initially cluster the indicators and then merge and screen secondary indicators, and classify the selected secondary indicators into primary indicators;

[0008] S3. Evaluate the primary indicators based on the calculated secondary indicators, and then evaluate the urban sidewalk environment;

[0009] S4. Use a GIS platform to overlay the evaluation results to form a visual graphic;

[0010] S5. Select a research sample and analyze its basic data;

[0011] S6. Conduct a sidewalk environment assessment of the research sample based on the primary and secondary indicators;

[0012] S7. Based on the evaluation results, propose suggestions for optimizing the sidewalk environment;

[0013] In step S1, a sidewalk environment feature word library is constructed based on the influencing factors of the sidewalk environment listed by relevant domestic and international research. The feature word library is decomposed into several word groups, and then the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm is used to convert the influencing factor word groups into word vectors and calculate the feature word values. The calculation method is as follows:

[0014]

[0015]

[0016] TF-IDF (Seed Feature Word Weight) = TF * IDF;

[0017] Based on the calculated word vectors, the text data of sidewalk environmental impact factors were categorized using Euclidean distance text similarity analysis. Indicators of 0.5 < D(X, Y) < 1 were considered similar and grouped into one category. Euclidean distance similarity:

[0018] D(X,Y)=||XY||2 (3);

[0019] Then, a similarity analysis matrix is ​​obtained using Gensim. Based on similarity, influencing factors (seeds) are initially screened to construct a concise, effective, and semantically relatively independent feature vocabulary. Finally, weight analysis is performed on the feature vocabulary using the formula:

[0020]

[0021] Calculate the weights of seed feature words and select seed feature words with weights greater than 0.7 as seed features affecting the sidewalk environment;

[0022] Through screening and weighted visualization, four primary indicators were identified: comfort, safety, accessibility, and convenience.

[0023] According to the above technical solution, in step S2, basic data of six first-tier and new first-tier cities in China are collected, a basic database is established, and feature words, i.e., secondary indicators, are analyzed and extracted.

[0024] The collected data was then processed using the mean-filling method. The fill formula is as follows:

[0025]

[0026] Next, dimensionless processing is performed, using standardization as the dimensionless processing method:

[0027]

[0028] The correlation coefficient model is used to represent the correlation between features and the sidewalk environment. The correlation coefficient matrix is ​​calculated as follows:

[0029]

[0030] In equation (7): ρ is the correlation coefficient, and COV(X,Y) is the covariance of X and Y. Let X be the standard deviation of X, and X be the mean.

[0031] Emotion judgment criteria were established, and emotion analysis and emotion naming were performed on the characteristics of the sidewalk environment. Emotion categories with no impact were removed, and key factors were extracted. Thirty-three independent features were selected, and k-means++ clustering was performed using a binary approach. Based on binary clustering, two initial cluster centers were randomly selected, and the distances from the remaining samples to these two centers were calculated. Clustering was then performed based on the closest distance. The distance calculation formula is as follows:

[0032]

[0033] The formula for calculating the category center points of two categories is as follows:

[0034]

[0035] Repeat the above steps until the category center no longer changes;

[0036] The selection of secondary indicators corresponds to the primary indicators. Preliminary clustering divides the secondary indicators into categories A, B, C, and D.

[0037] Finally, the secondary indicators are reclassified into the primary indicators.

[0038] According to the above technical solution, in S3, the pedestrian environmental comfort indicators include green view rate, sky visibility, interface enclosure degree and environmental diversity.

[0039] The green space ratio index of built-up areas in Category A is converted into the green view rate index of pedestrian environment. The formula for calculating the green view rate is:

[0040] Rate (green view rate) = S (plants) / 1 (10);

[0041] In equation (10), S represents the sum of pixels occupied by all plants in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0042] The urban heat island intensity, air quality index, and noise index in Category A are converted into sky visibility indicators. The formula for calculating sky visibility is as follows:

[0043] Rate (sky visibility area) = S (sky) / 1 (11);

[0044] In equation (11), S represents the sum of the pixels occupied by the sky in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0045] The walking interface indicators in Category A are transformed into two quantifiable indicators: interface enclosure degree and environmental diversity. The formula for calculating interface enclosure degree is as follows:

[0046] Rate (interface enclosure degree) = [S (wall) + S (building)] / 1 (12);

[0047] In Equation (12), [S(wall)+S(building)] represents the sum of pixels occupied by walls and buildings in the street view photo, and 1 represents the total pixel area of ​​the street view photo;

[0048] The formula for calculating environmental diversity is:

[0049] Rate (environmental diversity) = N (types of elements) / 1 (13);

[0050] In equation (13), N represents the sum of the types of elements in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0051] Pedestrian safety indicators include natural surveillance, street lighting availability, motor vehicle interference, and the availability of traffic-related facilities.

[0052] The crime rate index in Category B is converted into a natural surveillance index. The formula for calculating the natural surveillance index is as follows:

[0053] Rate (Natural Surveillance) = S (Pedestrians) / 1 (14);

[0054] In equation (14), S represents the number of pixels occupied by pedestrians in the street view photo, and 1 represents the total pixel area.

[0055] The road lighting facility indicators in Category B are converted into street light completeness indicators. The formula for calculating street light completeness is as follows:

[0056] Rate (streetlight completeness) = S (streetlights) / 1 (15);

[0057] In equation (15), S represents the number of pixels occupied by streetlights in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0058] The indicators of car ownership, environmental noise, SO2, and smoke in Category B are converted into vehicle interference indicators. The formula for calculating vehicle interference is as follows:

[0059] Rate (motor vehicle interference degree) = S (motor vehicle) / 1 (16);

[0060] In equation (16), S represents the number of pixels occupied by motor vehicles in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0061] The traffic fatality rate index in Category B is converted into a traffic infrastructure completeness index. The formula for calculating the completeness of traffic infrastructure is as follows:

[0062] Rate (completeness of transportation ancillary facilities) = S (ancillary facilities) / 1 (17);

[0063] In Equation (17), S represents the number of pixels occupied by traffic ancillary facilities such as traffic lights, road guardrails, and zebra crossings in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0064] Pedestrian accessibility indicators include road network density, intersection density, public transport station coverage, and relative pedestrian width.

[0065] The road network length, per capita urban road area, and built-up area road network density in Category C are converted into road network density indices. The formula for calculating road network density is as follows:

[0066] Density (road network density) = L (roads) / S (area of ​​region) (18);

[0067] In Equation (18), L (road) represents the sum of the lengths of the road network in the vector map data, and S (area) represents the area of ​​the study area;

[0068] The formula for calculating intersection density is:

[0069] Density (intersection density) = N (intersection) / S (area of ​​area) (19);

[0070] In equation (19), N represents the sum of the number of intersections in the vector map data, and S represents the area of ​​the study area;

[0071] The public transport station quantity index in Category C is converted into a public transport station coverage index. The formula for calculating the public transport station coverage rate is as follows:

[0072] Rate (bus stop coverage rate) = L (bus stop coverage length) / L (road length) (20);

[0073] Rate (subway station coverage rate) = L (subway station coverage length) / L (road length) (21);

[0074] In equations (20) and (21), L (site coverage length) represents the length of the road network covered by the reasonable service radius of all sites in the study area, and L (road length) represents the total length of the road network in the study area.

[0075] The sidewalk width index in Category C is converted into the sidewalk relative width index. The formula for calculating the sidewalk relative width is as follows:

[0076] Rate (relative width of sidewalk) = S (sidewalk) / S (roadway) (22);

[0077] In Equation (22), S(sidewalk) represents the number of pixels occupied by the sidewalk in the street view photo, and S(roadway) represents the pixel area occupied by the roadway in the street view photo.

[0078] Pedestrian accessibility indicators include POI density, POI mix, and building density;

[0079] The public service facility density, park area, number of medical institutions, number of schools, public utility land use, and population indicators in Category D are converted into POI density indicators. The formula for calculating POI density is as follows:

[0080] Density (POI density) = N (number of POIs) / S (area of ​​region) (23);

[0081] In equation (23), N represents the sum of the number of POIs superimposed on the vector map data, and S represents the area of ​​the study area;

[0082] The functional blending index in category D is converted into the POI blending index. The formula for calculating POI blending is as follows:

[0083] Rate (POI mixing degree) = N (number of POI types) / S (area of ​​region) (24);

[0084] In equation (24), N represents the sum of the number of POI types superimposed on the vector map data, and S represents the area of ​​the study area;

[0085] The land development intensity and land area requisitioned this year in Category D are converted into building density indicators. The formula for calculating building density is as follows:

[0086] Density (building density) = S (buildings) / S (area of ​​region) (25);

[0087] In Equation (25), S(building) represents the sum of the building base areas in the vector map data, and S(area) represents the area of ​​the study area.

[0088] According to the above technical solution, in step S4, a series of tasks such as data collection, data storage, data management, and data analysis are completed using a geographic information system (GIS). Spatial analysis, as the main function of GIS, is used to quantify the indicator data and then perform spatial positioning. Overlay analysis, as a subordinate function, is used to overlay the evaluation results of individual indicators to obtain a visual graphic.

[0089] The evaluation results data are mapped one-to-one with geospatial coordinates and assigned color levels. The evaluation results are then converted into visual graphics and overlaid to comprehensively reflect the characteristics of the evaluation indicators.

[0090] According to the above technical solution, in step S5, the pedestrian environment in the Gusu District of Suzhou is selected as the research sample, and the research area is divided with the streets in the Gusu District as the basic unit.

[0091] Road network data was obtained through OSM maps, roads outside the research scope were removed, POI data within the Gusu District were obtained, business types were simplified and classified, and road coordinate points with equal spacing were obtained through the Gusu District urban vector map. A panoramic static image of the Gusu District urban space was collected. After the static image was captured, semantic segmentation was performed on the panoramic static image to identify the pixel composition of the panoramic static image. 24 identification elements were selected, and the number and proportion of pixels of the elements were further calculated.

[0092] According to the above technical solution, in step S6, the comfort of the sidewalk environment is evaluated by assessing the greening rate, the visible sky area, the interface enclosure degree, and the environmental diversity.

[0093] The safety of sidewalks is evaluated based on factors such as natural surveillance, street lighting availability, motor vehicle interference, and the availability of traffic safety facilities.

[0094] The accessibility of the pedestrian environment is evaluated by assessing road network density, intersection density, public transportation station coverage, and relative width of the pedestrian walkway.

[0095] The accessibility of the pedestrian environment is evaluated by assessing POI density, POI mix, and building density.

[0096] According to the above technical solution, in S7, the 15 evaluation indicators of the pedestrian environment in Gusu District mainly form three different orientations of pedestrian environment: traffic-oriented, living-oriented, and leisure-oriented. Then, suggestions are made for the three different orientations of pedestrian environment based on the evaluation results.

[0097] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0098] 1. By extracting feature words from existing studies, we propose primary indicators. We then collect sidewalk data from first-tier and new first-tier cities in China, calculate the relevance, merge and screen the indicators to initially cluster them, and then transform the initially clustered indicators into the required secondary indicators. Finally, we classify the required indicators into primary indicators. Through screening, we make the sidewalk environment indicators more in line with the evaluation needs of the urban sidewalk environment, which is convenient for subsequent evaluation.

[0099] 2. By collecting map road network data of urban sidewalks, we can obtain the distribution of green view rate, sky visibility, interface enclosure degree, and environmental diversity to evaluate the comfort of the sidewalk environment. We can also obtain the distribution of natural surveillance degree, street lighting completeness, motor vehicle interference degree, and safety facility completeness to evaluate the safety of the sidewalk environment. Furthermore, we can obtain the distribution of average road network density, average road intersection density, sidewalk coverage of bus stops, sidewalk coverage of subway stations, and relative width of sidewalks to evaluate the accessibility of the sidewalk environment. Finally, we can obtain the distribution of POI density, POI mixing degree, and building density to evaluate the convenience of the sidewalk environment, making the analysis of the sidewalk environment more accurate.

[0100] 3. By using evaluation indicators to form different orientations of pedestrian environment, and making suggestions for different orientations of pedestrian environment, we can help improve the pedestrian environment according to actual needs, which can promote the continuous improvement of the urban pedestrian environment and play a positive role in alleviating urban traffic congestion, improving the quality of pedestrian environment and pedestrian landscape.

[0101] In summary, by merging, screening, and clustering pedestrian environment indicators in advance, we can obtain pedestrian environment evaluation indicators that better meet our needs. Then, we can select road network data for the pedestrian environment, conduct environmental analysis and evaluation based on the indicators, and further divide the pedestrian environment into different orientations, providing suggestions for the environment of different orientations. This makes the pedestrian environment analysis more accurate and helps improve the urban pedestrian environment. Attached Figure Description

[0102] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0103] Figure 1 This is a schematic diagram of the indicator evaluation process of the present invention;

[0104] Figure 2 This is the OSM map road network data of Gusu District processed by this invention;

[0105] Figure 3 This invention provides a spatial distribution map of the green visibility rate in Gusu District.

[0106] Figure 4 This invention provides a spatial distribution map of the visible sky over the Gusu District.

[0107] Figure 5 This is the kernel density map of the interface enclosure degree in the Suzhou area of ​​this invention;

[0108] Figure 6 This invention provides a spatial distribution map of environmental diversity in the Gusu District.

[0109] Figure 7 This is the kernel density map of the natural monitoring degree in Gusu District of this invention;

[0110] Figure 8 This is a spatial distribution map of the street light perfection level in Gusu District according to the present invention;

[0111] Figure 9 This invention provides a spatial distribution map of manually identified streetlights in Gusu District.

[0112] Figure 10 This is the kernel density map of the interference level of motor vehicles in Gusu District according to the present invention;

[0113] Figure 11 This is a kernel density diagram of the completeness of safety facilities in Gusu District according to the present invention;

[0114] Figure 12 This invention provides an average road network density map of Gusu District.

[0115] Figure 13 This invention provides a map showing the average length of streets in Gusu District.

[0116] Figure 14 This invention is a map showing the average density of road intersections in Gusu District.

[0117] Figure 15 This invention provides a diagram showing the coverage of sidewalks at bus stops in Gusu District.

[0118] Figure 16 This is a diagram showing the pedestrian walkway coverage of subway stations in Gusu District according to the present invention;

[0119] Figure 17 This invention provides a relative width diagram of sidewalks in Gusu District.

[0120] Figure 18 This invention is a distribution map of Points of Interest (POIs) in the Gusu District.

[0121] Figure 19 This is the POI kernel density map of the Suzhou area in this invention;

[0122] Figure 20 This is a POI mixing degree map of Suzhou District in this invention;

[0123] Figure 21 This invention is a building density map of Gusu District. Detailed Implementation

[0124] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0125] Example: Figure 1 As shown, this invention provides a technical solution, a research method for urban sidewalk environmental assessment based on multi-source data, comprising the following research steps:

[0126] S1. Based on research data, summarize current environmental assessment research indicators, classify and analyze the indicators based on NLP-text similarity analysis, and propose primary indicators;

[0127] S2. Select urban sidewalk environment assessment data, and based on sentiment analysis-binary clustering method, initially cluster the indicators and then merge and screen secondary indicators, and classify the selected secondary indicators into primary indicators;

[0128] S3. Evaluate the primary indicators based on the calculated secondary indicators, and then evaluate the urban sidewalk environment;

[0129] S4. Use a GIS platform to overlay the evaluation results to form a visual graphic;

[0130] S5. Select a research sample and analyze its basic data;

[0131] S6. Conduct a sidewalk environment assessment of the research sample based on the primary and secondary indicators;

[0132] S7. Based on the evaluation results, propose suggestions for optimizing the sidewalk environment.

[0133] In S1, a feature word library for the sidewalk environment is constructed based on the influencing factors of the sidewalk environment listed by relevant research at home and abroad. The feature word library is decomposed into several word groups, and then the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm is used to convert the influencing factor word groups into word vectors and calculate the feature word values. The calculation method is as follows:

[0134]

[0135]

[0136] TF-IDF (Seed Feature Word Weight) = TF * IDF;

[0137] Based on the calculated word vectors, the text data of sidewalk environmental impact factors are classified by Euclidean distance text similarity analysis. The index of 0.5 < D(X,Y) < 1 is judged to be similar and classified into one category.

[0138] Euclidean distance similarity:

[0139] D(X,Y)=||XY||2 (3);

[0140] Then, a similarity analysis matrix is ​​obtained using Gensim. Based on similarity, influencing factors (seeds) are initially screened to construct a concise, effective, and semantically relatively independent feature vocabulary. Finally, weight analysis is performed on the feature vocabulary using the formula:

[0141]

[0142] Calculate the weights of seed feature words and select seed feature words with weights greater than 0.7 as seed features affecting the sidewalk environment;

[0143] Through screening and visualizing the weights, four primary indicators were determined: comfort, safety, accessibility, and convenience. The weights of these indicators are shown in the table below:

[0144] category Comfort Security Accessibility Convenience Weight 13.02％ 10.50％ 9.66％ 7.14％

[0145] According to the above technical solution, in S2, secondary indicators are extracted based on the evaluation of primary indicators;

[0146] We collected basic OSM map data, Baidu street view images, and business POI data from six first-tier and new first-tier cities in China, obtained relevant statistical data, and established a basic database. We then analyzed and mined the sidewalk environmental indicator database to extract feature words, i.e., secondary indicators.

[0147] The collected data was then processed using the mean-filling method. The fill formula is as follows:

[0148]

[0149] Next, dimensionless processing is performed, using standardization as the dimensionless processing method:

[0150]

[0151] The correlation coefficient analysis method was used to determine the correlation between indicators, and the sentiment judgment of the indicators was carried out. The "sidewalk length" indicator in the map data, which is strongly correlated, was selected to transfer the quantitative indicators of the sidewalk environment and analyze the relationship between the relevant indicators and the sidewalk environment.

[0152] Specifically, a correlation coefficient model is used to represent the correlation between features and the sidewalk environment. The correlation coefficient matrix is ​​calculated as follows:

[0153]

[0154] In equation (7): ρ is the correlation coefficient, and COV(X,Y) is the covariance of X and Y. Let X be the standard deviation of X, and X be the mean.

[0155] Establish emotional judgment criteria, conduct emotional analysis and emotional naming of pedestrian environment characteristics, remove emotional categories that have no impact, and then conduct the next step of dimensional classification research.

[0156] Based on the sentiment classification results, we extracted key factors, removed collinearity-related features, and filtered the original indicators into 33 independent features. Using this as a basis, we performed k-means++ clustering with a binary approach to address the sample imbalance between categories. First, based on binary clustering, we randomly selected two initial cluster centers and calculated the distances of the remaining samples to these two centers. Then, we clustered the samples based on the closest distance. The distance calculation formula is as follows:

[0157]

[0158] The formula for calculating the category center points of two categories is as follows:

[0159]

[0160] Repeat the above operation until the class centers no longer change. The first step of kmeans++ clustering is completed. Then select the class with a larger sample size and perform kmeans++ clustering again. Iterate in this way until the sample reaches the number of classes to be clustered. Then the clustering is complete.

[0161] The selection of secondary indicators corresponds to the primary indicators, and the preliminary clustering classifies the secondary indicators into 4 categories;

[0162] Based on the clustering results, indicators are screened. The initial clustered indicators are merged and screened to form selected secondary indicators. The weights of the merged clusters are then categorized into four primary indicators. The indicator merging and screening comparison table is as follows:

[0163] In S3, evaluation data is selected from OSM urban vector road network data, Baidu POI data, and panoramic static map data to conduct a sidewalk environment evaluation;

[0164] Pedestrian walkway environmental comfort indicators include green view rate, sky visibility, interface enclosure degree, and environmental diversity;

[0165] The green space ratio of built-up areas is converted into the green view rate of sidewalk environment. The formula for calculating the green view rate is:

[0166] Rate (green view rate) = S (plants) / 1 (10);

[0167] In equation (10), S (plant) represents the sum of the pixels occupied by all plants in the street view photo;

[0168] 1 represents the total pixel area of ​​the street view photo;

[0169] The higher the proportion of S (plants), the higher the greening level of the sidewalk environment, and the closer the Rate (green view rate) is to 1.

[0170] The urban heat island intensity, air quality index, and noise index are converted into sky visibility indicators. The formula for calculating sky visibility is as follows:

[0171] Rate (sky visibility area) = S (sky) / 1 (11);

[0172] In equation (11), S(sky) represents the sum of the pixels occupied by the sky in the street view photo;

[0173] 1 represents the total pixel area of ​​the street view photo;

[0174] The larger the proportion of S (sky), the larger the visible sky area in the sidewalk environment, and the closer the Rate (visible sky area) is to 1.

[0175] The walking interface index is transformed into two quantifiable indicators: interface enclosure degree and environmental diversity. The formula for calculating interface enclosure degree is as follows:

[0176] Rate (interface enclosure degree) = [S (wall) + S (building)] / 1 (12);

[0177] In Equation (12), [S(wall)+S(building)] represents the sum of pixels occupied by walls and buildings in the street view photo, and 1 represents the total pixel area of ​​the street view photo;

[0178] When the proportion of [S(wall)+S(building)] is larger, the interface enclosure of pedestrian walking is stronger, the possibility of interaction between pedestrians and the sidewalk environment is higher, and the Rate (interface enclosure) is closer to 1.

[0179] The formula for calculating environmental diversity is:

[0180] Rate (environmental diversity) = N (types of elements) / 1 (13);

[0181] In Equation (13), N (feature type) represents the sum of all feature types in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0182] The greater the proportion of N (type of elements), the stronger the diversity of the sidewalk environment, and the closer the Rate (environmental diversity) is to 1. The smaller the proportion of N (type of elements), the weaker the diversity of the sidewalk environment, and the closer the Rate (environmental diversity) is to 0.

[0183] Pedestrian safety indicators include natural surveillance, street lighting availability, motor vehicle interference, and the availability of traffic-related facilities.

[0184] The crime rate index is converted into a natural surveillance index. The formula for calculating natural surveillance is as follows:

[0185] Rate (Natural Surveillance) = S (Pedestrians) / 1 (14);

[0186] In Equation (14), S (pedestrian) represents the number of pixels occupied by pedestrians in the street view photo, and 1 represents the total pixel area;

[0187] The larger the proportion of pedestrians (S), the stronger the natural surveillance of the sidewalk environment, and the closer the Rate (natural surveillance) is to 1.

[0188] The road lighting facility index is converted into a street light completeness index. The formula for calculating street light completeness is as follows:

[0189] Rate (streetlight completeness) = S (streetlights) / 1 (15);

[0190] In Equation (15), S (street lamp) represents the number of pixels occupied by the street lamp in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0191] When S (streetlight) has a pixel, it means that there is streetlight coverage in the pedestrian environment, and Rate (streetlight completeness) has a value. When S (streetlight) does not have a pixel, it means that there is no streetlight coverage in the pedestrian environment, and Rate (streetlight completeness) is 0.

[0192] The indicators of car ownership, environmental noise, SO2, and smoke are converted into a vehicle interference index. The formula for calculating vehicle interference is as follows:

[0193] Rate (motor vehicle interference degree) = S (motor vehicle) / 1 (16);

[0194] In Equation (16), S (motor vehicle) represents the number of pixels occupied by motor vehicles (cars, motorcycles, buses, trucks) in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0195] The larger the number of pixels occupied by S (motor vehicles), the stronger the interference of motor vehicles on the sidewalk environment, and the closer the Rate (motor vehicle interference degree) is to 1.

[0196] The traffic fatality rate indicator is converted into a traffic infrastructure completeness indicator. The formula for calculating the completeness of traffic infrastructure is as follows:

[0197] Rate (completeness of transportation ancillary facilities) = S (ancillary facilities) / 1 (17);

[0198] In Equation (17), S (ancillary facilities) represents the number of pixels occupied by traffic ancillary facilities such as traffic lights, road guardrails, and zebra crossings in the street view photo, and 1 represents the total pixel area of ​​the street view photo.

[0199] The larger the number of pixels occupied by S (ancillary facilities), the more complete the traffic ancillary facilities in the pedestrian environment, the better the pedestrian safety is protected, and the closer the Rate (completeness of traffic ancillary facilities) is to 1.

[0200] Pedestrian accessibility indicators include road network density, intersection density, public transport station coverage, and relative pedestrian width.

[0201] The road network length, per capita urban road area, and built-up area road network density indicators are converted into road network density indicators. The formula for calculating road network density is as follows:

[0202] Density (road network density) = L (roads) / S (area of ​​region) (18);

[0203] In Equation (18), L (road) represents the sum of the lengths of the road network in the vector map data, and S (area) represents the area of ​​the study area;

[0204] With a fixed area of ​​region S, the longer the length of L (road), the higher the road network density, the higher the density of pedestrian walkways, the stronger the pedestrian capacity, and the higher the Density (road network density) value.

[0205] The formula for calculating intersection density is:

[0206] Density (intersection density) = N (intersection) / S (area of ​​area) (19);

[0207] In equation (19), N (intersection) represents the sum of the number of intersections in the vector map data, and S area represents the area of ​​the study area;

[0208] With a fixed area of ​​region S, the more intersections there are, the higher the intersection density becomes, the more travel routes are available, and the higher the Density value.

[0209] The indicator of the number of public transportation stops is converted into the indicator of public transportation stop coverage. The formula for calculating public transportation stop coverage is as follows:

[0210] Rate (bus stop coverage rate) = L (bus stop coverage length) / L (road length) (20);

[0211] Rate (subway station coverage rate) = L (subway station coverage length) / L (road length) (21);

[0212] In equations (20) and (21), L (site coverage length) represents the length of the road network covered by the reasonable service radius of all sites in the study area, and L (road length) represents the total length of the road network in the study area.

[0213] The higher the coverage of public transportation stops, the easier it is for pedestrians to reach public transportation stops, the higher the accessibility of their travel, and the closer the Rate (stop coverage) ratio is to 1.

[0214] The sidewalk width index is converted into a relative sidewalk width index. The formula for calculating the relative sidewalk width is as follows:

[0215] Rate (relative width of sidewalk) = S (sidewalk) / S (roadway) (22);

[0216] In Equation (22), S(sidewalk) represents the number of pixels occupied by the sidewalk in the street view photo, and S(roadway) represents the pixel area occupied by the roadway in the street view photo.

[0217] When the proportion of S (sidewalk) is larger, the relative width of the sidewalk is larger, the pedestrian capacity is stronger, and the Rate (relative width of sidewalk) is closer to 1.

[0218] Pedestrian accessibility indicators include POI density, POI mix, and building density;

[0219] The density of public service facilities, park area, number of medical institutions, number of schools, public utility land, and population indicators are converted into POI density indicators. The formula for calculating POI density is as follows:

[0220] Density (POI density) = N (number of POIs) / S (area of ​​region) (23);

[0221] In equation (23), N (number of POIs) represents the sum of the number of POIs superimposed on the vector map data, and S (area of ​​the region) represents the area of ​​the study area;

[0222] With a fixed area S, the higher the number of points of interest (POIs) N, the higher the POI density, the stronger the service capacity within the area, the more convenient the services enjoyed by pedestrians, and the higher the Density (POI density) value.

[0223] The functional blending index is converted into a POI blending index. The POI blending index is calculated using the following formula:

[0224] Rate (POI mixing degree) = N (number of POI types) / S (area of ​​region) (24);

[0225] In Equation (24), N (number of POI types) represents the sum of the number of POI types superimposed on the vector map data, and S (area of ​​region) represents the area of ​​the study area;

[0226] With a fixed area S, the higher the number of POI types N, the higher the POI mixing degree. Pedestrians can obtain more types of services in a more concentrated area, the more convenient the services enjoyed by pedestrians are, and the higher the Rate (POI mixing degree) value.

[0227] The land development intensity and the land area requisitioned this year are converted into building density indicators. The formula for calculating building density is as follows:

[0228] Density (building density) = S (buildings) / S (area of ​​region) (25);

[0229] In Equation (25), S(building) represents the sum of the building base areas in the vector map data, and S(area) represents the area of ​​the study area;

[0230] With a fixed area S (area), the larger the sum of the building areas S (areas), the more convenient the services enjoyed by pedestrians, and the higher the Density (building density) value.

[0231] Based on the evaluation methods for the indicators, the evaluation methods for the indicators will be summarized.

[0232] In S4, a series of tasks such as data collection, data storage, data management, and data analysis are completed using Geographic Information System (GIS). Spatial analysis, as the main function of GIS, is used to quantify the indicator data and then perform spatial positioning. Overlay analysis, as a subordinate function, is used to overlay the evaluation results of individual indicators to obtain visualization graphics.

[0233] The evaluation results data are mapped one-to-one with geospatial coordinates and assigned color levels. The evaluation results are then converted into visual graphics and overlaid to comprehensively reflect the characteristics of the evaluation indicators.

[0234] like Figure 2As shown in Figure S5, the pedestrian environment within the Gusu District of Suzhou was selected as the research sample, and the research area was divided into basic units, namely Baiyangwan Street, Sujin Street, Huqiu Street, Pingjiang Street, Jinchang Street, Canglang Street, Shuangta Street, and Wumenqiao Street within the Gusu District.

[0235] Road network data was obtained using OSM maps. Roads outside the study area and road segments smaller than 5 meters were removed. The final road network data contained 1144 roads.

[0236] We used Python to connect with the location search API provided by Baidu Developer Platform to acquire POI data within the Gusu District, simplified and classified the business types, and finally divided the business types into four major categories: social services, productive services, consumer services, and manufacturing.

[0237] Using Python and Baidu's panoramic static image API interface, road coordinate points with equal spacing are obtained through the urban vector map of Gusu District. Based on the coordinate points, a panoramic static image of Gusu District is collected from a 360-degree perspective. After the static image is captured, semantic segmentation is performed on the panoramic static image to identify the pixel composition of the panoramic static image. 24 identification elements are selected and the number and proportion of pixels of each element are further calculated.

[0238] The filtered data was finally imported into the ArcGIS platform for processing, and the coordinates of the panoramic static map were matched one-to-one with the urban space.

[0239] like Figure 3 As shown, in S6, the comfort index evaluation is as follows: Green view rate evaluation: The green view rate of the sidewalk environment in the ancient city area is among the better greening areas in Gusu District;

[0240] like Figure 4 As shown, the sky visibility evaluation shows a spatial distribution pattern of lower inner areas and higher outer areas. The sky visibility in the outer areas of the city is generally higher than that in the city center, exhibiting a contiguous and clustered spatial layout.

[0241] like Figure 5 As shown, the interface enclosure degree evaluation: the high value of the interface enclosure degree forms three distinct areas in the north, central and south within the Gusu District, with the north and south moats as the dividing lines.

[0242] like Figure 6 As shown, the environmental diversity assessment shows that the overall environmental diversity exhibits a distribution characteristic of generally high values.

[0243] like Figure 7As shown, the safety indicators are evaluated as follows: Natural surveillance level evaluation: The overall natural surveillance level is low. The pixel ratio of people in the pedestrian environment clearly forms a two-level concentric structure. The pedestrian distribution on the pedestrian walkways in the entire Gusu District presents two strip-shaped distributions: east-west and north-south.

[0244] like Figure 8-9 As shown, the street light completeness evaluation: the street light coverage rate of the selected roads reached 96%, and the street light completeness level of the entire Gusu District study area was very high.

[0245] like Figure 10 As shown, the evaluation of vehicle interference shows that the proportion of vehicle pixels is negatively correlated with the street level. The proportion of vehicle pixels on main roads and expressways is generally low, while the proportion of vehicle pixels on secondary roads and branch roads is relatively high.

[0246] From a spatial distribution perspective, the proportion of motor vehicles shows a staggered pattern across the entire area. In terms of the number of motor vehicles in the pedestrian environment, there are two core clusters in space: the central area of ​​the ancient city and the area around Suzhou Station.

[0247] like Figure 11 As shown, the evaluation of the completeness of traffic safety facilities shows that the pixel ratio of traffic safety facilities at the scale of the old city area is weaker at the inside and stronger at the outside.

[0248] like Figure 12-13 As shown, the accessibility indicators are evaluated as follows: Road network density evaluation: The overall average road network density shows a step-by-step ranking: Sujin Street > Pingjiang Street > Wumenqiao Street > Jinchang Street > Canglang Street > Shuangta Street > Baiyangwan Street > Huqiu Street;

[0249] like Figure 14 As shown, the intersection density evaluation is as follows: the intersection density of the 8 streets is ranked as follows: Pingjiang Street > Sujin Street > Shuangta Street > Wumenqiao Street > Canglang Street > Jinchang Street > Baiyangwan Street > Huqiu Street;

[0250] like Figure 15-16 As shown, the coverage rate of public transportation stations is evaluated as follows: within the service radius of buses, the overall coverage rate of sidewalks in Gusu District reaches 75.88%, and within the service radius of subways, the overall coverage rate of sidewalks in Gusu District reaches 41.30%.

[0251] like Figure 17 As shown, the relative width of sidewalks in Gusu District is 34.84%.

[0252] The primary indicator of convenience includes three secondary indicators: POI density, POI mix, and building density;

[0253] like Figure 18-19As shown, the POI density evaluation shows that the distribution of POIs in Gusu District is characterized by cohesion and dispersion, and cluster distribution. The spatial aggregation is obvious. Within the ancient city area, there are many POIs that are clustered, while in the outer areas they are more dispersed. The structural distribution of POIs in Gusu District is positively correlated with the urban center system structure. The main distribution characteristics can be summarized as follows: (1) Core area clustered, with multiple points distributed around it; (2) Multiple corridors coexist, and linear dispersion along the roads; (3) Mountain and water elements divide the business layout.

[0254] like Figure 20 As shown, the POI mixing degree evaluation shows that the POI mixing degree in the ancient city center of Gusu District is significantly higher than that in the surrounding areas. The POI richness in the ancient city center area is significantly higher than that in other areas in terms of difference, and the mixing degree in the southern streets is significantly higher than that in the northern streets.

[0255] like Figure 21 As shown, the building density evaluation shows that the building density distribution in Gusu District exhibits a very obvious concentric radial structure. The building density within the moat of the ancient city forms a first-level step, gradually decreasing towards the outer edge. A clear boundary is formed along the moat, and the area within the moat also shows a pattern of higher density in the north and lower density in the south.

[0256] In S7, the 15 evaluation indicators of the pedestrian environment in Gusu District mainly formed three different orientations: (1) traffic-oriented pedestrian environment, (2) living-oriented pedestrian environment, and (3) leisure-oriented pedestrian environment;

[0257] The following recommendations are made for pedestrian environments with traffic flow: Ensure smooth pedestrian networks: add pedestrian safety facilities, increase intersection density, widen sidewalks, and provide safe and convenient transfer options;

[0258] Balanced service industry layout: Improve the configuration of service facilities and pay attention to the mixed functions of business formats;

[0259] The following recommendations are made for pedestrian environments in residential areas: Improve the quality of the spatial environment: provide a pleasant spatial scale, create rich environmental interfaces, and create a vibrant pedestrian landscape;

[0260] Reduce interference from motor vehicles: Add traffic facilities to reduce conflicts between pedestrians and vehicles, and set up signs to guide green travel;

[0261] The following suggestions are made for the environment of leisure pedestrian walkways: Create a good landscape environment: Create distinctive landscape spaces and pay attention to the richness and layering of the landscape;

[0262] Improve pedestrian accessibility: optimize the configuration of service nodes and create a quiet pedestrian environment.

[0263] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A research method for urban sidewalk environmental assessment based on multi-source data, characterized in that: The research steps include the following: S1. Based on research data, summarize current environmental assessment research indicators, classify and analyze the indicators based on NLP-text similarity analysis, and propose primary indicators; S2. Select urban sidewalk environment assessment data, and based on sentiment analysis-binary clustering method, initially cluster the indicators and then merge and screen secondary indicators, and classify the selected secondary indicators into primary indicators; S3. Evaluate the primary indicators based on the calculated secondary indicators, and then evaluate the urban sidewalk environment; S4. Use a GIS platform to overlay the evaluation results to form a visual graphic; S5. Select a research sample and analyze its basic data; S6. Conduct a sidewalk environment assessment of the research sample based on the primary and secondary indicators; S7. Based on the evaluation results, propose suggestions for optimizing the sidewalk environment; In step S1, a sidewalk environment feature word library is constructed based on the influencing factors of the sidewalk environment listed by relevant domestic and international research. The feature word library is decomposed into several word groups, and then the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm is used to convert the influencing factor word groups into word vectors and calculate the feature word values. The calculation method is as follows: TF-IDF (Seed Feature Word Weight) = TF * IDF; Based on the calculated word vectors, the text data of sidewalk environmental impact factors were categorized using Euclidean distance text similarity analysis. Indicators of 0.5 < D(X, Y) < 1 were considered similar and grouped into one category. Euclidean distance similarity: D(X,Y)=||XY||2 (3); Then, a similarity analysis matrix is ​​obtained using Gensim. Based on similarity, influencing factors (seeds) are initially screened to construct a concise, effective, and semantically relatively independent feature vocabulary. Finally, weight analysis is performed on the feature vocabulary using the formula: Calculate the weights of seed feature words and select seed feature words with weights greater than 0.7 as seed features affecting the sidewalk environment; Through screening and visualizing the weights, four primary indicators were identified: comfort, safety, accessibility, and convenience. In S2, basic data of six first-tier and new first-tier cities in China are collected, and a basic database is established to analyze and extract feature words, i.e., secondary indicators. The collected data was then processed using the mean-filling method. The fill formula is as follows: Next, dimensionless processing is performed, using standardization as the dimensionless processing method: The correlation coefficient model is used to represent the correlation between features and the sidewalk environment. The correlation coefficient matrix is ​​calculated as follows: In equation (7): ρ is the correlation coefficient, and COV(X,Y) is the covariance of X and Y. Let X be the standard deviation of X, and X be the mean. Emotion judgment criteria were established, and emotion analysis and emotion naming were performed on the characteristics of the sidewalk environment. Emotion categories with no impact were removed, and key factors were extracted. Thirty-three independent features were selected, and k-means++ clustering was performed using a binary approach. Based on binary clustering, two initial cluster centers were randomly selected, and the distances from the remaining samples to these two centers were calculated. Clustering was then performed based on the closest distance. The distance calculation formula is as follows: The formula for calculating the category center points of two categories is as follows: Repeat the above steps until the category center no longer changes; The selection of secondary indicators corresponds to the primary indicators. Preliminary clustering divides the secondary indicators into categories A, B, C, and D. Finally, the secondary indicators are reclassified into the primary indicators; In S3, the pedestrian environmental comfort indicators include green view rate, sky visibility, interface enclosure degree, and environmental diversity. The green space ratio index of built-up areas in Category A is converted into the green view rate index of pedestrian environment. The formula for calculating the green view rate is: Rate (green view rate) = S (plants) / 1 (10); In equation (10), S represents the sum of pixels occupied by all plants in the street view photo, and 1 represents the total pixel area of ​​the street view photo. The urban heat island intensity, air quality index, and noise index in Category A are converted into sky visibility indicators. The formula for calculating sky visibility is as follows: Rate (sky visibility area) = S (sky) / 1 (11); In equation (11), S represents the sum of the pixels occupied by the sky in the street view photo, and 1 represents the total pixel area of ​​the street view photo. The walking interface indicators in Category A are transformed into two quantifiable indicators: interface enclosure degree and environmental diversity. The formula for calculating interface enclosure degree is as follows: Rate (interface enclosure degree) = [S (wall) + S (building)] / 1 (12); In Equation (12), [S(wall)+S(building)] represents the sum of pixels occupied by walls and buildings in the street view photo, and 1 represents the total pixel area of ​​the street view photo; The formula for calculating environmental diversity is: Rate (environmental diversity) = N (types of elements) / 1 (13); In equation (13), N represents the sum of the types of elements in the street view photo, and 1 represents the total pixel area of ​​the street view photo. Pedestrian safety indicators include natural surveillance, street lighting availability, motor vehicle interference, and the availability of traffic-related facilities. The crime rate index in Category B is converted into a natural surveillance index. The formula for calculating the natural surveillance index is as follows: Rate (Natural Surveillance) = S (Pedestrians) / 1 (14); In equation (14), S represents the number of pixels occupied by pedestrians in the street view photo, and 1 represents the total pixel area. The road lighting facility indicators in Category B are converted into street light completeness indicators. The formula for calculating street light completeness is as follows: Rate (streetlight completeness) = S (streetlights) / 1 (15); In equation (15), S represents the number of pixels occupied by streetlights in the street view photo, and 1 represents the total pixel area of ​​the street view photo. The indicators of car ownership, environmental noise, SO2, and smoke in Category B are converted into vehicle interference indicators. The formula for calculating vehicle interference is as follows: Rate (motor vehicle interference degree) = S (motor vehicle) / 1 (16); In equation (16), S represents the number of pixels occupied by motor vehicles in the street view photo, and 1 represents the total pixel area of ​​the street view photo. The traffic fatality rate index in Category B is converted into a traffic infrastructure completeness index. The formula for calculating the completeness of traffic infrastructure is as follows: Rate (completeness of transportation ancillary facilities) = S (ancillary facilities) / 1 (17); In Equation (17), S represents the number of pixels occupied by traffic ancillary facilities such as traffic lights, road guardrails, and zebra crossings in the street view photo, and 1 represents the total pixel area of ​​the street view photo. Pedestrian accessibility indicators include road network density, intersection density, public transport station coverage, and relative pedestrian width. The road network length, per capita urban road area, and built-up area road network density in Category C are converted into road network density indices. The formula for calculating road network density is as follows: Density (road network density) = L (roads) / S (area of ​​region) (18); In Equation (18), L (road) represents the sum of the lengths of the road network in the vector map data, and S represents the area of ​​the study area; The formula for calculating intersection density is: Density (intersection density) = N (intersection) / S (area of ​​area) (19); In equation (19), N represents the sum of the number of intersections in the vector map data, and S represents the area of ​​the study area; The public transport station quantity index in Category C is converted into a public transport station coverage index. The formula for calculating the public transport station coverage rate is as follows: Rate (bus stop coverage rate) = L (bus stop coverage length) / L (road length) (20); Rate (subway station coverage rate) = L (subway station coverage length) / L (road length) (21); In equations (20) and (21), L (site coverage length) represents the length of the road network covered by the reasonable service radius of all sites in the study area, and L (road length) represents the total length of the road network in the study area. The sidewalk width index in Category C is converted into the sidewalk relative width index. The formula for calculating the sidewalk relative width is as follows: Rate (relative width of sidewalk) = S (sidewalk) / S (roadway) (22); In Equation (22), S(sidewalk) represents the number of pixels occupied by the sidewalk in the street view photo, and S(roadway) represents the pixel area occupied by the roadway in the street view photo. Pedestrian accessibility indicators include POI density, POI mix, and building density; The public service facility density, park area, number of medical institutions, number of schools, public utility land use, and population indicators in Category D are converted into POI density indicators. The formula for calculating POI density is as follows: Density (POI density) = N (number of POIs) / S (area of ​​region) (23); In equation (23), N represents the sum of the number of POIs superimposed on the vector map data, and S represents the area of ​​the study area; The functional blending index in category D is converted into the POI blending index. The formula for calculating POI blending is as follows: Rate (POI mixing degree) = N (number of POI types) / S (area of ​​region) (24); In equation (24), N represents the sum of the number of POI types superimposed on the vector map data, and S represents the area of ​​the study area; The land development intensity and land area requisitioned this year in Category D are converted into building density indicators. The formula for calculating building density is as follows: Density (building density) = S (buildings) / S (area of ​​region) (25); In Equation (25), S(building) represents the sum of the building base areas in the vector map data, and S(area) represents the area of ​​the study area.

2. The method for urban sidewalk environmental assessment based on multi-source data according to claim 1, characterized in that, In S4, a series of tasks, including data collection, data storage, data management, and data analysis, are completed using a Geographic Information System (GIS). Spatial analysis, as the main function of GIS, is used to quantify the indicator data and then perform spatial positioning. Overlay analysis, as a subordinate function, is used to overlay the evaluation results of individual indicators to obtain a visual graphic. The evaluation results data are mapped one-to-one with geospatial coordinates and assigned color levels. The evaluation results are then converted into visual graphics and overlaid to comprehensively reflect the characteristics of the evaluation indicators.

3. The method for urban sidewalk environmental assessment based on multi-source data according to claim 1, characterized in that, In S5, the pedestrian environment within the Gusu District of Suzhou is selected as the research sample, and the research area is divided using streets within the Gusu District as the basic unit. Road network data was obtained through OSM maps, roads outside the research scope were removed, POI data within the Gusu District were obtained, business types were simplified and classified, and road coordinate points with equal spacing were obtained through the Gusu District urban vector map. A panoramic static image of the Gusu District urban space was collected. After the static image was captured, semantic segmentation was performed on the panoramic static image to identify the pixel composition of the panoramic static image. 24 identification elements were selected, and the number and proportion of pixels of the elements were further calculated.

4. The method for urban sidewalk environmental assessment based on multi-source data according to claim 1, characterized in that, In S6, the comfort of the sidewalk environment is evaluated by assessing the greening rate, the visible sky area, the interface enclosure degree, and the environmental diversity. The safety of sidewalks is evaluated based on factors such as natural surveillance, street lighting availability, motor vehicle interference, and the availability of traffic safety facilities. The accessibility of the pedestrian environment is evaluated by assessing road network density, intersection density, public transportation station coverage, and relative width of the pedestrian walkway. The accessibility of the pedestrian environment is evaluated by assessing POI density, POI mix, and building density.

5. The method for urban sidewalk environmental assessment based on multi-source data according to claim 1, characterized in that, In S7, the 15 evaluation indicators of the pedestrian environment in Gusu District mainly form three different orientations of pedestrian environment: traffic-oriented, living-oriented, and leisure-oriented. Based on the evaluation results, suggestions are made for the three different orientations of pedestrian environment.

Images