A method and system for evaluating urban visual landscape based on structural equation model

Abstract

The invention discloses an urban visual landscape evaluation method and system based on a structural equation model, belonging to the field of urban planning; the evaluation method comprises: crawling urban visual landscape pictures and accompanying information, cleaning the data, and establishing an urban visual landscape picture data set; quantifying urban visual landscape elements based on a deep convolutional neural network, semantically segmenting the urban visual landscape pictures in the data set, and obtaining urban visual landscape element indicators; discriminating the relationship between the quantified urban visual landscape element indicators and landscape characteristics, summarizing and classifying them through factor analysis of SPSS software, and finding the corresponding relationship between the urban visual landscape element indicators and the landscape characteristics; constructing a structural equation model of the influence of urban visual landscape characteristics on public preferences based on the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics; and adjusting the indicators of the structural equation model to obtain an urban visual landscape evaluation model.

Description

Technical Field

[0001] The invention belongs to the field of urban planning, and in particular relates to an urban visual landscape evaluation method and system based on a structural equation model. Background Art

[0002] Urban landscape is a comprehensive reflection of the artificial and natural environments. Visual landscape covers more than 80% of human perception and is a key factor affecting people's quality of life and happiness. Therefore, it is crucial to evaluate the quality of urban visual landscape for the creation and maintenance of urban landscape.

[0003] In order to create high-quality urban visual landscape and improve the public's satisfaction with landscape perception, it is necessary to conduct a quantitative and standardized evaluation of urban visual landscape. Nowadays, big data technology and social media images are widely used in the field of urban planning, and deep learning technology is becoming more and more mature. However, the existing urban visual landscape evaluation methods are limited to general statistical methods, such as regression analysis and principal component analysis, which are limited to the qualitative description of urban visual landscape and mainly focus on the single impact of landscape indicators on landscape quality, ignoring the relationship between indicators and landscape characteristics and the comprehensive impact of landscape characteristics on landscape quality. Summary of the invention

[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide an urban visual landscape evaluation method and system based on a structural equation model, clarify the important indicators that affect the urban visual landscape and the landscape characteristics to which they belong, and construct a structural equation model of the impact of urban visual landscape characteristics on public preferences, so as to conduct a quantitative and standardized urban visual landscape evaluation.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] An urban visual landscape evaluation method based on a structural equation model comprises the following steps:

[0007] Define the distribution range of urban visual landscapes, crawl urban visual landscape images and accompanying information according to the range, clean the data, and establish an urban visual landscape image dataset;

[0008] Based on the deep convolutional neural network, the urban visual landscape elements are quantified. The deeplabV3+ model is used to perform semantic segmentation on the urban visual landscape images in the dataset, and the pixel ratio of each element in the urban visual landscape images is counted to obtain the urban visual landscape element index.

[0009] The relationship between the quantified urban visual landscape element indicators and landscape characteristics was determined, and the corresponding relationship between the urban visual landscape element indicators and landscape characteristics was found through factor analysis using SPSS software.

[0010] Based on the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics, a structural equation model of the impact of urban visual landscape characteristics on public preferences is constructed;

[0011] The indicators of the structural equation model were adjusted to obtain the urban visual landscape evaluation model.

[0012] Furthermore, the data cleaning process is as follows: first, ensure that the number of images for each user ID is only 1; second, remove images of non-urban visual landscapes; and finally remove images with the number of likes, reposts, and comments below 1.

[0013] Furthermore, the urban visual landscape image dataset, as a training set for the deeplabV3+ model, can identify 21 urban visual landscape elements. The urban visual landscape elements are measured by measuring their pixel ratios in the picture to obtain urban visual landscape element indicators.

[0014] Furthermore, when summarizing and classifying them through factor analysis of the SPSS software, the reliability of the collected data was determined based on KMO and Bartlett's sphericity test; then the maximum variance method was used to extract the correlation matrix, obtain the rotated component matrix, and eliminate useless urban visual landscape element indicators.

[0015] Furthermore, when the correspondence between the urban visual landscape element indicators and the landscape characteristics is obtained, the correspondence between the urban visual landscape element indicators and the landscape characteristics is determined according to the total variance explanation and the related parameters, and the following is obtained: the landscape naturalness corresponds to the vegetation visibility, the water surface visibility, the mountain visibility and the land visibility; the landscape richness corresponds to the landscape sketch visibility and the element richness; the landscape coherence corresponds to the crowd visibility, the building visibility and the sky visibility; the landscape interference corresponds to the traffic sign visibility, the vehicle visibility and the road visibility.

[0016] Furthermore, the urban visual landscape index calculation formula is as follows:

[0017] Vegetation visibility GVI:

[0018] Mountain visibility MVI:

[0019] Water surface visibility WVI:

[0020] Land visibility TVI:

[0021] Sky visibility SVI:

[0022] Crowd visibility PVI:

[0023] Building visibility UCL: Traffic sign visibility TSVI:

[0024] Vehicle visibility VVI:

[0025] Road visibility rate RVI:

[0026] Element richness (LEC) Landscape Visibility (OVI)

[0027] In the formula, A t_i is the sum of the pixels in image i, G t_i is the sum of vegetation pixels in image i, M t_i is the sum of the mountain pixels in image i, W t_i is the total number of water pixels in image i, T t_i is the total number of land pixels in image i, S t_i is the sum of sky pixels in image i, P t_i is the sum of the crowd pixels in image i, B t_i is the sum of building pixels in image i, TS t_i is the sum of the traffic sign pixels in image i, TL t_i is the total number of traffic light pixels in image i, CA t_i is the sum of the car pixels in image i, TR t_i is the sum of train pixels in image i, BU t_i is the total number of bus pixels in image i, MO t_i is the sum of the motorcycle pixels in image i, R t_i is the sum of road pixels in image i, E i is the total number of landscape elements in image i, WA t_i is the sum of wall pixels in image i, PO t_i is the total number of column pixels in image i, RO t_i is the total number of rock pixels in image i, FE t_i is the sum of fence pixels in image i.

[0028] Furthermore, the steps of constructing the structural equation model include:

[0029] S41, using AMOS software, the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics was imported into AMOS software to achieve data matching. Based on the standardized path coefficients between various landscape characteristics in AMOS software, the influence relationship between landscape naturalness, landscape richness, landscape coherence and landscape disturbance on public preference was obtained;

[0030] S42, the influence of landscape naturalness, landscape richness, landscape coherence and landscape disturbance on public preference was tested respectively, and the relationship between different landscape characteristics was tested to obtain the structural equation model.

[0031] Furthermore, the indicators of the structural equation model are adjusted as follows: the DF value must be less than 5, the GFI value must be greater than 0.9, the AGFI value must be greater than 0.9, and the RMSEA value must be less than 0.8; and the operating logic of the model is adjusted according to the above indicators, and the paths that do not show significant correlation are removed until they are fully adapted to obtain the final urban visual landscape evaluation model.

[0032] An urban visual landscape evaluation system based on structural equation model, including:

[0033] Dataset establishment module: define the distribution range of urban visual landscape, crawl urban visual landscape pictures and accompanying information according to the range, clean the data, and establish an urban visual landscape picture dataset;

[0034] Element index acquisition module: quantify urban visual landscape elements based on deep convolutional neural networks, use the deeplabV3+ model to perform semantic segmentation on urban visual landscape images in the data set, and count the pixel ratios of various elements in urban visual landscape images to obtain urban visual landscape element indicators;

[0035] Correspondence determination module: Determine the relationship between the quantified urban visual landscape element indicators and landscape characteristics, summarize and classify them through factor analysis of SPSS software, and find out the correspondence between urban visual landscape element indicators and landscape characteristics;

[0036] Model building module: Based on the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics, a structural equation model of the impact of urban visual landscape characteristics on public preferences is constructed;

[0037] And the model adjustment module: adjust the indicators of the structural equation model to obtain the urban visual landscape evaluation model.

[0038] A computer storage medium stores a readable program, characterized in that when the program is run, the above method can be executed.

[0039] Beneficial effects of the present invention:

[0040] With the help of big data and deep convolutional neural network technology, the representativeness of public perception data obtained in urban visual landscape evaluation and the accuracy of quantitative urban visual landscape element indicators are improved; with the help of factor analysis, the correspondence between landscape indicators and characteristics in urban visual landscape evaluation is obtained, and a hierarchical evaluation indicator system is constructed; based on the structural equation model, an urban visual landscape evaluation model is constructed, which takes into account the relationship between various landscape indicators and their comprehensive impact on the evaluation results, accurately explains the internal transmission mechanism of urban visual landscape evaluation indicators, improves the scientificity and persuasiveness of the evaluation, and forms a complete set of standardized and quantitative urban visual landscape evaluation processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 is a flow chart of the urban visual landscape evaluation method of the present invention;

[0043] Figure 2 It is a schematic diagram of the urban visual landscape evaluation model of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] like Figure 1 As shown, a method for evaluating urban visual landscape based on a structural equation model includes the following steps:

[0046] S1, define the distribution range of urban visual landscape, crawl urban visual landscape pictures and accompanying information according to the relevant range, clean the data according to certain rules, and establish an urban visual landscape picture dataset;

[0047] Among them, when defining the distribution range of urban visual landscape, the central urban area of ​​the analyzed city is generally selected as the specific range for collecting urban visual landscape pictures;

[0048] When crawling urban visual landscape pictures and accompanying information, Python software is used to crawl urban visual landscape pictures, the number of likes, reposts, comments, labels, descriptions, IDs and other related information of the pictures; the collected pictures are generally sourced from well-known domestic social media platforms to ensure the richness of data volume and data types (such as Weibo, Xiaohongshu, Dianping.com, Mafengwo, etc.);

[0049] The steps of cleaning the data according to certain rules include: first, ensuring that each user ID has only one image, in order to avoid a single user uploading too many pictures and interfering with the widely collected public attitudes; second, eliminating pictures of non-urban visual landscapes, such as those with animals, food, vehicles, etc. as the subject, in order to avoid interference with the results when analyzing non-urban visual landscape pictures; finally, pictures with the number of likes, reposts, and comments below 1 will be eliminated, mainly to prevent personal preferences rather than reflect public preferences. For pictures that have not been viewed, the quality of the landscape they reflect should not be considered.

[0050] S2, based on deep convolutional neural network, quantifies urban visual landscape elements. The deeplabV3+ model is used to perform semantic segmentation on the cleaned images in the S1 dataset. The pixel ratios of various elements in the urban visual landscape images are counted to obtain the urban visual landscape element indicators.

[0051] The open source dataset is used as the training set of deeplabV3+ to ensure the accuracy of identifying urban visual landscape elements. Ultimately, 21 elements can be identified, including the sky, motorcycles, mountains, stones, vegetation, buildings, traffic lights, traffic signs, crowds, trains, cars, trucks, land, fences, roads, sidewalks, buses, floors, walls, pillars, riders, etc. The above urban visual landscape elements are measured by their pixel ratio in the picture to obtain urban visual landscape element indicators.

[0052] S3, determine the relationship between the quantified urban visual landscape element indicators and landscape characteristics, summarize and classify them through factor analysis of SPSS software, and find out the corresponding relationship between urban visual landscape element indicators and landscape characteristics;

[0053] When the factors were summarized and classified by SPSS software, the Kaiser-Meyer-Olkin (KMO) and Bartlett's sphericity test (K 2 ) to determine the reliability of the collected data; then the maximum variance method was used to extract the correlation matrix, obtain the rotated component matrix, and eliminate useless urban visual landscape element indicators; the calculation formula of KMO is:

[0054]

[0055] where r ij represents the simple correlation coefficient, represents the partial correlation coefficient.

[0056] Bartlett's test of sphericity (K 2 ) is calculated as:

[0057]

[0058]

[0059] Where r is the number of groups, and the standard deviation of each group is

[0060] When the correspondence between the urban visual landscape element index and the landscape characteristics is obtained, the correspondence between the urban visual landscape element index and the landscape characteristics is determined according to the total variance explanation and related parameters, and the following is obtained: the landscape naturalness corresponds to the vegetation visibility, water surface visibility, mountain visibility and land visibility; the landscape richness corresponds to the visibility of landscape elements such as fences, pillars, and walls, and the element richness; the landscape coherence corresponds to the crowd visibility, building visibility and sky visibility; the landscape interference corresponds to the traffic sign visibility, vehicle visibility and road visibility;

[0061] The calculation process of urban visual landscape element index is shown in Table 1:

[0062] Table 1: Calculation table of urban visual landscape elements index

[0063]

[0064]

[0065] Among them, A t_i is the sum of the pixels in image i, G t_i is the sum of vegetation pixels in image i, M t_i is the sum of the mountain pixels in image i, W t_i is the total number of water pixels in image i, T t_i is the sum of land pixels in image i, S t_i is the sum of sky pixels in image i, P t_i is the sum of the crowd pixels in image i, B t_i is the sum of building pixels in image i, TS t_i is the sum of the traffic sign pixels in image i, TL t_i is the total number of traffic light pixels in image i, CA t_i is the sum of the car pixels in image i, TR t_i is the sum of train pixels in image i, BUt_i is the total number of bus pixels in image i, MO t_i is the sum of the motorcycle pixels in image i, R t_i is the sum of road pixels in image i, E i is the total number of landscape elements in image i, WA t_i is the sum of wall pixels in image i, PO t_i is the total number of column pixels in image i, RO t_i is the total number of rock pixels in image i, FE t_i is the sum of fence pixels in image i.

[0066] S4, based on the correspondence between urban visual landscape indicators and urban visual landscape characteristics, a preliminary structural equation model for urban visual landscape evaluation was constructed;

[0067] The model building process includes:

[0068] S41, using AMOS software, the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics was imported into AMOS software to achieve data matching. Based on the standardized path coefficients between various landscape characteristics in AMOS software, the influence relationship between landscape naturalness, landscape richness, landscape coherence, and landscape disturbance on public preference was obtained;

[0069] S42, the influence of landscape naturalness, landscape richness, landscape coherence and landscape disturbance on public preference was tested respectively, and the relationship between different landscape characteristics was tested. The correspondence between urban visual landscape indicators and urban visual landscape characteristics was obtained, and a preliminary structural equation model for urban visual landscape evaluation was constructed.

[0070] S5, adjusting the indicators of the preliminary structural equation model for urban visual landscape evaluation to obtain an urban visual landscape evaluation model;

[0071] like Figure 2 As shown in the figure, the model is adjusted according to the fit index of the structural equation model, where the Degrees of Freedom (DF) value must be less than 5, the Goodness of Fit Index (GFI) value must be greater than 0.9, the Adjusted GFI (AGFI) value must be greater than 0.9, and the Root-Mean-Square Error of Approximation (RMSEA) value must be less than 0.8; the operation logic of the model is adjusted according to the above indicators, and the paths that do not show significant correlation are removed until they are fully adapted to obtain the final urban visual landscape evaluation model; the specific process is as follows:

[0072] First, the influencing paths between the characteristics in the model were determined. If the influencing paths were significant (p-value < 0.005), they were retained; if the influencing paths were not significant, they were removed. Secondly, the standardized path coefficients in the model fitness index were used to determine the highest standardized path coefficient in the landscape characteristic influencing path as the final influencing path of the landscape characteristic. Finally, the relevant residuals were corrected according to the correction index in the model to obtain a model that met the above indicators.

[0073] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0074] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A method for evaluating urban visual landscape based on a structural equation model, characterized in that: The following steps are involved: Define the distribution range of urban visual landscapes, crawl urban visual landscape images and accompanying information according to the range, clean the data, and establish an urban visual landscape image dataset; Based on the deep convolutional neural network, the urban visual landscape elements are quantified. The deeplabV3+ model is used to perform semantic segmentation on the urban visual landscape images in the dataset, and the pixel ratio of each element in the urban visual landscape images is counted to obtain the urban visual landscape element index. The relationship between the quantified urban visual landscape element indicators and landscape characteristics was determined, and the corresponding relationship between the urban visual landscape element indicators and landscape characteristics was found through factor analysis using SPSS software. Based on the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics, a preliminary structural equation model for urban visual landscape evaluation is constructed; Adjusting the indicators of the structural equation model to obtain an urban visual landscape evaluation model; The data cleaning process is as follows: first, ensure that there is only one image for each user ID; second, remove images that are not urban visual landscapes; and finally remove images with less than 1 like, forward, or comment count; When the correspondence between the urban visual landscape element index and the landscape characteristics is obtained, the correspondence between the urban visual landscape element index and the landscape characteristics is determined according to the total variance explanation and the related parameters, and the following is obtained: the landscape naturalness corresponds to the vegetation visibility rate, the water surface visibility rate, the mountain visibility rate and the land visibility rate; the landscape richness corresponds to the landscape sketch visibility rate and the element richness; Landscape continuity corresponds to the visibility of people, buildings and the sky; landscape interference corresponds to the visibility of traffic signs, vehicles and roads. The calculation formula of urban visual landscape index is as follows: Vegetation visibility GVI: Mountain visibility MVI: Water surface visibility WVI: Land visibility TVI: Sky visibility SVI: Crowd visibility PVI: Building visibility UCL: Traffic sign visibility TSVI: Vehicle visibility VVI: Road visibility rate RVI: Element richness LEC: Landscape sketch visibility OVI: In the formula, A t_i is the sum of the pixels in image i, G t_i is the sum of vegetation pixels in image i, M t_i is the sum of the mountain pixels in image i, W t_i is the total number of water pixels in image i, T t_i is the total number of land pixels in image i, S t_i is the sum of sky pixels in image i, P t_i is the sum of the crowd pixels in image i, B t_i is the sum of building pixels in image i, TS t_i is the sum of the traffic sign pixels in image i, TL t_i is the sum of the traffic light pixels in image i, CA t_i is the sum of the car pixels in image i, TR t_i is the sum of train pixels in image i, BU t_i is the total number of bus pixels in image i, MO t_i is the sum of the motorcycle pixels in image i, R t_i is the sum of road pixels in image i, E i is the total number of landscape elements in image i, WA t_i is the sum of wall pixels in image i, PO t_i is the total number of column pixels in image i, RO t_i is the total number of rock pixels in image i, FE t_i is the sum of fence pixels in image i; The steps of constructing the structural equation model include: S41, using AMOS software, the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics was imported into AMOS software to achieve data matching. Based on the standardized path coefficients between various landscape characteristics in AMOS software, the influence relationship between landscape naturalness, landscape richness, landscape coherence and landscape disturbance on public preference was obtained; S42, the influence of landscape naturalness, landscape richness, landscape coherence and landscape disturbance on public preference was tested respectively, and the relationship between different landscape characteristics was tested to obtain a preliminary structural equation model for urban visual landscape evaluation.

2. The urban visual landscape evaluation method based on the structural equation model according to claim 1 is characterized in that: Using the open source dataset as the training set of the deeplabV3+ model, 21 urban visual landscape elements can be identified. The urban visual landscape element indicators are obtained by measuring their pixel ratio in the picture.

3. The urban visual landscape evaluation method based on structural equation model according to claim 1 is characterized in that: When summarizing and classifying them through factor analysis of the SPSS software, the reliability of the collected data was judged according to KMO and Bartlett's sphericity test; then the maximum variance method was used to extract the correlation matrix, obtain the rotated component matrix, and eliminate useless urban visual landscape element indicators.

4. The urban visual landscape evaluation method based on structural equation model according to claim 1 is characterized in that: The indicators of the structural equation model are adjusted as follows: the DF value must be less than 5, the GFI value must be greater than 0.9, the AGFI value must be greater than 0.9, and the RMSEA value must be less than 0.8; and the operating logic of the model is adjusted according to the above indicators, and the paths that do not show significant correlation are removed until they are fully adapted to obtain the final urban visual landscape evaluation model.

5. An urban visual landscape evaluation system based on a structural equation model, characterized in that: include: Dataset establishment module: define the distribution range of urban visual landscape, crawl urban visual landscape pictures and accompanying information according to the range, clean the data, and establish an urban visual landscape picture dataset; Element index acquisition module: quantify urban visual landscape elements based on deep convolutional neural networks, use the deeplabV3+ model to perform semantic segmentation on urban visual landscape images in the data set, and count the pixel ratios of various elements in urban visual landscape images to obtain urban visual landscape element indicators; Correspondence determination module: Determine the relationship between the quantified urban visual landscape element indicators and landscape characteristics, summarize and classify them through factor analysis of SPSS software, and find out the correspondence between urban visual landscape element indicators and landscape characteristics; Model building module: Based on the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics, a structural equation model of the impact of urban visual landscape characteristics on public preferences is constructed; And the model adjustment module: adjust the indicators of the structural equation model to obtain the urban visual landscape evaluation model; The data cleaning process is as follows: first, ensure that there is only one image for each user ID; second, remove images that are not urban visual landscapes; and finally remove images with less than 1 like, forward, or comment count; When the correspondence between the urban visual landscape element index and the landscape characteristics is obtained, the correspondence between the urban visual landscape element index and the landscape characteristics is determined according to the total variance explanation and the related parameters, and the following is obtained: the landscape naturalness corresponds to the vegetation visibility rate, the water surface visibility rate, the mountain visibility rate and the land visibility rate; the landscape richness corresponds to the landscape sketch visibility rate and the element richness; Landscape continuity corresponds to the visibility of people, buildings and the sky; landscape interference corresponds to the visibility of traffic signs, vehicles and roads. The calculation formula of urban visual landscape index is as follows: Vegetation visibility GVI: Mountain visibility MVI: Water surface visibility WVI: Land visibility TVI: Sky visibility SVI: Crowd visibility PVI: Building visibility UCL: Traffic sign visibility TSVI: Vehicle visibility VVI: Road visibility rate RVI: Element richness LEC: Landscape sketch visibility OVI: In the formula, A t_i is the sum of the pixels in image i, G t_i is the sum of vegetation pixels in image i, M t_i is the sum of the mountain pixels in image i, W t_i is the total number of water pixels in image i, T t_i is the total number of land pixels in image i, S t_i is the sum of sky pixels in image i, P t_i is the sum of the crowd pixels in image i, B t_i is the sum of building pixels in image i, TS t_i is the sum of the traffic sign pixels in image i, TL t_i is the sum of the traffic light pixels in image i, CA t_i is the sum of the car pixels in image i, TR t_i is the sum of train pixels in image i, BU t_i is the total number of bus pixels in image i, MO t_i is the sum of the motorcycle pixels in image i, R t_i is the sum of road pixels in image i, E i is the total number of landscape elements in image i, WA t_i is the sum of wall pixels in image i, PO t_i is the total number of column pixels in image i, RO t_i is the total number of rock pixels in image i, FE t_i is the sum of fence pixels in image i; The steps of constructing the structural equation model include: S41, using AMOS software, the corresponding relationship between urban visual landscape indicators and urban visual landscape characteristics was imported into AMOS software to achieve data matching. Based on the standardized path coefficients between various landscape characteristics in AMOS software, the influence relationship between landscape naturalness, landscape richness, landscape coherence and landscape disturbance on public preference was obtained; S42, the influence of landscape naturalness, landscape richness, landscape coherence and landscape disturbance on public preference was tested respectively, and the relationship between different landscape characteristics was tested to obtain a preliminary structural equation model for urban visual landscape evaluation.

6. A computer storage medium storing a readable program, characterized in that: When the program is run, the method described in any one of claims 1 to 4 can be executed.

Images