Urban landscape design-oriented skyline analysis method and system

Abstract

The invention discloses a skyline analysis method and system for urban landscape design, and relates to the field of urban planning and design, and the method comprises the steps: collecting urban landscape historical images of a target city, generating an image data set, and dividing the image data set into a training set and a test set; establishing an SAM image segmentation model, and training and testing the SAM image segmentation model; acquiring an urban landscape image of a target city, and inputting the urban landscape image into the tested SAM image segmentation model to obtain a segmented binary image; performing edge extraction on the binarized image to obtain skyline discrete points and calculating urban landscape evaluation system indexes; and comprehensively analyzing the indexes of the urban landscape evaluation system to obtain an analysis result of the skyline of the target city. According to the method, the city skyline can be accurately extracted, large-scale model training can be avoided, so that a large amount of cost is saved, a scientific reference basis is provided for city planning and design, and optimization and improvement of city landscapes are promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of urban planning and design, and more particularly to a skyline analysis method and system for urban landscape design. Background Art

[0002] At present, with the acceleration of global urbanization, the city's skyline, as an important part of the city's image, has attracted more and more attention. The skyline not only reflects the city's architectural style and cultural characteristics, but also affects the quality of life of residents and the sustainable development of the city. Therefore, how to effectively analyze and optimize the city's skyline has become an important topic in the field of urban planning and design.

[0003] Existing skyline analysis methods mainly rely on manual observation and empirical judgment, which usually have the following problems:

[0004] 1. Highly subjective: Traditional skyline analysis often relies on the personal experience and aesthetics of designers or planners and lacks objective evaluation criteria. This subjectivity may lead to large differences in the evaluation of the same city skyline by different analysts, affecting the scientific nature of planning decisions.

[0005] 2. Inefficiency: Manual analysis of the skyline requires a lot of time and effort, especially when facing large-scale cities or complex landscapes. The analysis process is often cumbersome and time-consuming. This inefficient analysis method is difficult to meet the needs of rapidly developing urban planning.

[0006] 3. Lack of systematicity: Existing methods usually only focus on one aspect of the skyline, such as building height or shape, while ignoring the overall and multi-dimensional characteristics of the skyline. This one-sided analysis method cannot fully reflect the aesthetic value and functionality of the city skyline.

[0007] 4. Insufficient technical means: Although computer vision and image processing technologies have made significant progress in recent years, their application in the field of urban skyline analysis is still relatively limited. Existing technologies are often unable to achieve automated extraction and analysis of complex urban landscapes, resulting in insufficient accuracy and reliability of analysis results.

[0008] 5. Lack of quantitative indicators: Current analysis methods often lack systematic quantitative indicators and cannot comprehensively evaluate the skyline. This makes it difficult for planners to make decisions based on scientific data when designing cities, affecting the overall coordination and aesthetics of the urban landscape.

[0009] Therefore, there is an urgent need for a new skyline analysis method that can overcome the shortcomings of existing technologies and provide a more scientific and effective reference for urban planning and design through automation and quantification. Summary of the invention

[0010] In view of this, the present invention provides a skyline analysis method and system for urban landscape design to solve the problems existing in the background technology.

[0011] In order to achieve the above object, the present invention adopts the following technical solution:

[0012] On the one hand, a skyline analysis method for urban landscape design is provided, including:

[0013] Collect historical images of the urban landscape of the target city, generate an image dataset and divide it into a training set and a test set;

[0014] Establishing a SAM image segmentation model, training and testing the SAM image segmentation model using a training set and a test set to obtain a tested SAM image segmentation model;

[0015] Collect the urban landscape image of the target city and input it into the tested SAM image segmentation model to obtain the binary image after segmentation;

[0016] Performing edge extraction on the binary image to obtain skyline discrete points;

[0017] Based on the obtained skyline discrete points, the indicators of the urban landscape evaluation system are calculated;

[0018] A comprehensive analysis of the indicators of the urban landscape assessment system is conducted to obtain the analysis results of the target city skyline.

[0019] Optionally, the SAM image segmentation model includes a SAM encoder, a SAM decoder, a linear network, and an edge extractor;

[0020] The historical images of urban landscape in the training set are input into the SAM encoder to obtain the embedded feature image. At the same time, the historical images of urban landscape are processed by the linear network to generate prompt points and sent to the SAM decoder.

[0021] The SAM encoder performs feature encoding on the historical image of the urban landscape to obtain an embedded feature image; and sends the embedded feature image to the SAM decoder;

[0022] The SAM decoder receives the embedded feature image and combines it with the cue points generated by the linear network to decode the historical image of the urban landscape and obtain a segmented binary image.

[0023] The edge extractor extracts the edges of the binary image and obtains discrete points of the skyline.

[0024] Optionally, the linear network is a fully connected network model, used to generate prompt points in the original image, the linear network input layer is set to 1024 neurons, the hidden layer is set to 2048 neurons, and the output layer is 1 neuron; when retraining the linear network, the number of iterations is set to 20,000 times, and the loss function adopts the Binary Cross Entropy loss function. Points are selected from the original image, and an area with a side length of 32 centered on the point is used as input to the linear network. In the training set image, if the corresponding selected area is white, it is a positive sample, otherwise it is considered to be a negative sample; after inputting the linear network output and the training set information into the loss function, the loss value is obtained to update the neural network parameters until the requirements for the number of iterations are met.

[0025] Optionally, the indicators of the urban landscape evaluation system include the tortuosity of the landmark building outline, the influence range of the landmark building outline, the primacy of the landmark building, the outline tortuosity, and the outline rhythm index.

[0026] Optionally, the calculation process of the landmark building outline tortuosity is:

[0027] After obtaining the discrete points of the skyline, the low points are used to divide the segments, and the two adjacent low points and the middle high point form a band. The percentage of the sum of the difference between the low point and the high point ΔH = Δh1 + Δh2 and the horizontal distance ΔL between the two low points is used to calculate the contour tortuosity T. The formula is as follows:

[0028]

[0029] Optionally, the calculation process of the influence range of the landmark building outline is:

[0030] Calculate the horizontal distances Δd1 and Δd2 between the landmark building as the maximum point and the minimum points on both sides, Δd = Δd1 + Δd2, and take the sum of these two distance values ​​as the influence interval of the landmark building. W represents the overall width of the field of view. The formula is as follows:

[0031]

[0032] Optionally, the calculation process of the primacy of the landmark building is:

[0033] The highest point H within the determined skyline max = max(H) minus the average building height H of the towers within the range mean =mean(H), then divided by the height of the highest building H max , where H represents the skyline discrete point vector, the formula is as follows:

[0034]

[0035] Optionally, the calculation process of the contour tortuosity is:

[0036] The contour tortuosity is expressed as skyline variance, and the formula is as follows:

[0037]

[0038] Where N is the number of discrete points on the skyline.

[0039] Optionally, the calculation process of the contour rhythm index is:

[0040] The contour rhythm index R expresses the state of fluctuation and takes the difference between adjacent skyline discrete points as an evaluation index. A contour tortuosity greater than 0.5 times is considered an undulating point, and a contour tortuosity less than 0.5 times is considered a flat point. The ratio of undulating points to the total number of discrete points N is the contour rhythm index.

[0041] On the other hand, a skyline analysis system for urban landscape design is provided, including:

[0042] The dataset generation module collects historical images of the urban landscape of the target city, generates an image dataset and divides it into a training set and a test set;

[0043] A model building module is used to build a SAM image segmentation model, and train and test the SAM image segmentation model using a training set and a test set to obtain a tested SAM image segmentation model;

[0044] The skyline discrete point acquisition module collects the urban landscape image of the target city and inputs it into the tested SAM image segmentation model to obtain a binary image after segmentation; performs edge extraction on the binary image to obtain skyline discrete points;

[0045] The index calculation module calculates the urban landscape evaluation system index based on the obtained skyline discrete points;

[0046] The result analysis module conducts a comprehensive analysis of the indicators of the urban landscape evaluation system to obtain the analysis results of the target city skyline.

[0047] Through the above technical solutions, it can be known that compared with the prior art, the present invention discloses a skyline analysis method and system for urban landscape design, including: collecting historical images of urban landscape of the target city, generating an image data set and dividing it into a training set and a test set; establishing a SAM image segmentation model, using the training set and the test set to train and test the SAM image segmentation model, and obtaining a tested SAM image segmentation model; collecting urban landscape images of the target city, and inputting them into the tested SAM image segmentation model to obtain a binarized image after segmentation; extracting edges from the binarized image to obtain skyline discrete points; calculating urban landscape evaluation system indicators based on the obtained skyline discrete points; and comprehensively analyzing the urban landscape evaluation system indicators to obtain the analysis results of the target city skyline. The present invention can not only accurately extract the urban skyline, but also avoid large-scale model training, thereby saving a lot of costs, but also provide a scientific reference basis for urban planning and design, and promote the optimization and improvement of urban landscapes. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] 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, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0049] Figure 1 A flow chart of the method provided by the present invention;

[0050] Figure 2 The schematic diagram of the SAM image segmentation model provided by the present invention;

[0051] Figure 3 A linear network schematic diagram provided by the present invention;

[0052] Figure 4 A schematic diagram of calculating the tortuosity of the outline of a landmark building provided by the present invention;

[0053] Figure 5 This is a schematic diagram of the influence range of calculating the landmark building outline provided by the present invention.

[0054] Figure 6 This is a schematic diagram of the contour rhythm index calculation method provided by the present invention. DETAILED DESCRIPTION

[0055] 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.

[0056] On the one hand, an embodiment of the present invention discloses a skyline analysis method for urban landscape design. Figure 1 As shown, including:

[0057] Collect historical images of the urban landscape of the target city, generate an image dataset and divide it into a training set and a test set;

[0058] Establish a SAM image segmentation model, use the training set and the test set to train and test the SAM image segmentation model, and obtain the tested SAM image segmentation model;

[0059] Collect the urban landscape image of the target city and input it into the tested SAM image segmentation model to obtain the binary image after segmentation;

[0060] Perform edge extraction on the binary image to obtain discrete points of the skyline;

[0061] Based on the obtained skyline discrete points, the indicators of the urban landscape evaluation system are calculated;

[0062] A comprehensive analysis of the indicators of the urban landscape assessment system is conducted to obtain the analysis results of the target city skyline.

[0063] In a specific embodiment, Figure 2 As shown, the SAM image segmentation model includes a SAM encoder, a SAM decoder, a linear network, and an edge extractor;

[0064] The historical images of urban landscape in the training set are input into the SAM encoder to obtain embedded feature images. At the same time, the historical images of urban landscape are processed by a linear network to generate prompt points and sent to the SAM decoder. Among them, the points 5% away from the edge of the upper left and upper right corners of the image are selected as prompt points.

[0065] The SAM encoder performs feature encoding on the historical image of the urban landscape to obtain an embedded feature image; and sends the embedded feature image to the SAM decoder;

[0066] The SAM decoder receives the embedded feature image and combines it with the cue points generated by the linear network to decode the historical image of the urban landscape and obtain a segmented binary image.

[0067] The edge extractor extracts the edges of the binary image and obtains discrete points of the skyline.

[0068] In a specific embodiment, Figure 3 As shown in the figure, the linear network is a fully connected network model, which is used to generate prompt points in the original image. The input layer of the linear network is set to 1024 neurons, the hidden layer is set to 2048 neurons, and the output layer is 1 neuron. When retraining the linear network, the number of iterations is set to 20,000 times, and the loss function uses the Binary CrossEntropy loss function. Points are selected from the original image, and the area with a side length of 32 and the point as the center is used as input to the linear network. In the training set image, if the corresponding selected area is white, it is a positive sample, that is, the sky area; otherwise it is considered a negative sample; the linear network output and the training set information are input into the loss function to obtain the loss value to update the neural network parameters until the requirements of the number of iterations are met.

[0069] In a specific embodiment, the indicators of the urban landscape evaluation system include the tortuosity of the outline of the landmark building, the influence range of the outline of the landmark building, the primacy of the landmark building, the outline tortuosity, and the outline rhythm index.

[0070] In a specific embodiment, the calculation process of the landmark building outline tortuosity is:

[0071] After obtaining the skyline discrete points, such as Figure 4 As shown, the paragraphs are divided by low points, and the two adjacent low points and the middle high point form a band. The percentage of the sum of the difference between the low point and the high point ΔH=Δh1+Δh2 and the horizontal distance ΔL between the two low points is used to calculate the contour tortuosity T. The formula is as follows:

[0072]

[0073] The higher the tortuosity index of the landmark building outline, the richer the changes in the single band outline and the more prominent the image of the landmark building.

[0074] In a specific embodiment, the calculation process of the influence range of the landmark building outline is:

[0075] like Figure 5 As shown, the horizontal distances Δd1 and Δd2 between the landmark building as the maximum point and the minimum points on both sides are calculated, Δd = Δd1 + Δd2, and the sum of these two distance values ​​is taken as the influence interval of the landmark building. W represents the overall width of the field of view. The formula is as follows:

[0076]

[0077] The larger the impact interval value of a landmark building and the larger the ratio, the more significant the landmark building is in terms of visual impact; conversely, the smaller the impact interval value of a landmark building and the smaller the ratio, the less significant the landmark building is.

[0078] In a specific embodiment, the calculation process of the primacy of the landmark building is as follows:

[0079] The highest point H within the determined skyline max = max(H) minus the average building height H of the towers within the range mean =mean(H), then divided by the height of the highest building H max , where H represents the skyline discrete point vector, the formula is as follows:

[0080]

[0081] The primacy of a landmark building is used to describe the centrality of the height of a landmark building. The larger the index, the more prominent the landmark building is and the stronger the centrality of the skyline.

[0082] In a specific embodiment, the calculation process of the profile tortuosity is:

[0083] The contour tortuosity is expressed as skyline variance, and the formula is as follows:

[0084]

[0085] Where N is the number of discrete points on the skyline.

[0086] The higher the contour tortuosity, the better the rhythm of the skyline, and the observer's preference for it increases.

[0087] In a specific embodiment, the calculation process of the contour rhythm index is:

[0088] The contour rhythm index R expresses the state of fluctuation and has a strong descriptive value for the overall rhythm of the skyline. The contour rhythm index uses the difference between adjacent skyline discrete points as an evaluation indicator. A contour tortuosity greater than 0.5 times is considered an undulating point, and a contour tortuosity less than 0.5 times is considered a flat point. The ratio of the undulating point to the total number of discrete points N is the contour rhythm index. The calculation method is as follows: Figure 6 shown.

[0089] This example is based on the SAM image segmentation model, and selects the points 5% away from the edge of the upper left and upper right corners of the image as indicator points to automatically detect the skyline of the urban landscape image in the form of discrete points. Based on the skyline discrete points, an urban landscape evaluation system is proposed, which includes five city skyline indicators: landmark building outline tortuosity, landmark building influence range, landmark building primacy, outline tortuosity, and outline rhythm index, providing a reference for urban design.

[0090] On the other hand, a skyline analysis system for urban landscape design is provided, including:

[0091] The dataset generation module collects historical images of the urban landscape of the target city, generates an image dataset and divides it into a training set and a test set;

[0092] A model building module is used to build a SAM image segmentation model, and train and test the SAM image segmentation model using a training set and a test set to obtain a tested SAM image segmentation model;

[0093] The skyline discrete point acquisition module collects the urban landscape image of the target city and inputs it into the tested SAM image segmentation model to obtain a binary image after segmentation; the edge of the binary image is extracted to obtain the skyline discrete points;

[0094] The index calculation module calculates the urban landscape evaluation system index based on the obtained skyline discrete points;

[0095] The result analysis module conducts a comprehensive analysis of the indicators of the urban landscape evaluation system to obtain the analysis results of the target city skyline.

[0096] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0097] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A skyline analysis method for urban landscape design, characterized in that: include: Collect historical images of the urban landscape of the target city, generate an image dataset and divide it into a training set and a test set; Establishing a SAM image segmentation model, training and testing the SAM image segmentation model using a training set and a test set to obtain a tested SAM image segmentation model; Collecting the urban landscape image of the target city and inputting it into the tested SAM image segmentation model to obtain a binary image after segmentation; performing edge extraction on the binary image to obtain skyline discrete points; Based on the obtained skyline discrete points, the indicators of the urban landscape evaluation system are calculated; A comprehensive analysis of the indicators of the urban landscape assessment system is conducted to obtain the analysis results of the target city skyline.

2. The skyline analysis method for urban landscape design according to claim 1, characterized in that: The SAM image segmentation model includes a SAM encoder, a SAM decoder, a linear network, and an edge extractor; The historical urban landscape images in the training set are input into the SAM encoder to obtain embedded feature images. At the same time, the historical urban landscape images are processed using a linear network to generate cue points and sent to the SAM decoder. The SAM encoder performs feature encoding on the historical image of the urban landscape to obtain an embedded feature image; and sends the embedded feature image to the SAM decoder; The SAM decoder receives the embedded feature image and combines it with the cue points generated by the linear network to decode the historical image of the urban landscape and obtain a segmented binary image. The edge extractor extracts the edges of the binary image and obtains discrete points of the skyline.

3. The skyline analysis method for urban landscape design according to claim 2 is characterized in that: The linear network is a fully connected network model, which is used to generate prompt points in the original image. The input layer of the linear network is set to 1024 neurons, the hidden layer is set to 2048 neurons, and the output layer is 1 neuron. When retraining the linear network, the number of iterations is set to 20,000 times, and the loss function adopts the Binary Cross Entropy loss function. Points are selected from the original image, and an area with a side length of 32 centered on the point is used as input to the linear network. In the training set image, if the corresponding selected area is white, it is a positive sample, otherwise it is identified as a negative sample. After the linear network output and the training set information are input into the loss function, the loss value is obtained to update the neural network parameters until the requirements set for the number of iterations are met.

4. The skyline analysis method for urban landscape design according to claim 1, characterized in that: The indicators of the urban landscape evaluation system include the tortuosity of the outline of the landmark building, the influence range of the outline of the landmark building, the primacy of the landmark building, the outline tortuosity, and the outline rhythm index.

5. The skyline analysis method for urban landscape design according to claim 4, characterized in that: The calculation process of the landmark building outline tortuosity is as follows: After obtaining the discrete points of the skyline, the low points are used to divide the segments, and the two adjacent low points and the middle high point form a band. The percentage of the sum of the difference between the low point and the high point ΔH = Δh1 + Δh2 and the horizontal distance ΔL between the two low points is used to calculate the contour tortuosity T. The formula is as follows:

6. The skyline analysis method for urban landscape design according to claim 4, characterized in that: The calculation process of the influence range of the landmark building outline is as follows: Calculate the horizontal distances Δd1 and Δd2 between the landmark building as the maximum point and the minimum points on both sides, Δd = Δd1 + Δd2, and take the sum of these two distance values ​​as the influence interval of the landmark building. W represents the overall width of the field of view. The formula is as follows:

7. The skyline analysis method for urban landscape design according to claim 4, characterized in that: The calculation process of the primacy of the landmark building is as follows: The highest point H within the determined skyline max = max(H) minus the average building height H of the towers within the range mean =mean(H), then divided by the height of the highest building H max , where H represents the skyline discrete point vector, the formula is as follows:

8. The skyline analysis method for urban landscape design according to claim 7, characterized in that: The calculation process of the profile tortuosity is: The contour tortuosity is expressed as skyline variance, and the formula is as follows: Where N is the number of discrete points on the skyline.

9. The skyline analysis method for urban landscape design according to claim 4, characterized in that: The calculation process of the contour rhythm index is: The contour rhythm index R expresses the state of fluctuation and takes the difference between adjacent skyline discrete points as an evaluation index. A contour tortuosity greater than 0.5 times is considered an undulating point, and a contour tortuosity less than 0.5 times is considered a flat point. The ratio of undulating points to the total number of discrete points N is the contour rhythm index.

10. A skyline analysis system for urban landscape design, characterized in that: A skyline analysis method for urban landscape design according to any one of claims 1 to 9 is applied, comprising: The dataset generation module collects historical images of the urban landscape of the target city, generates an image dataset and divides it into a training set and a test set; A model building module is used to build a SAM image segmentation model, and train and test the SAM image segmentation model using a training set and a test set to obtain a tested SAM image segmentation model; The skyline discrete point acquisition module collects the urban landscape image of the target city and inputs it into the tested SAM image segmentation model to obtain a binary image after segmentation; performs edge extraction on the binary image to obtain skyline discrete points; The index calculation module calculates the urban landscape evaluation system index based on the obtained skyline discrete points; The result analysis module conducts a comprehensive analysis of the indicators of the urban landscape evaluation system to obtain the analysis results of the target city skyline.

Images