Quantitative analysis method of spatial elements in urban high-intensity areas based on image data

Abstract

The present invention discloses a method for quantitative analysis of spatial elements in urban high-intensity areas based on image data, which relates to the field of image element recognition technology. The method comprises: obtaining a street view panorama from an observation point in a selected urban high-intensity area; obtaining shallow features of the street view panorama based on visual detection technology, and obtaining deep features of the street view panorama based on a deep learning model; based on topological analysis, aligning each of the shallow features and deep features with set logical rules to construct a topological map; inputting the topological map into a pre-trained classification model, and finally outputting the classification results of each spatial element in the street view image and its spatial distribution map. The present invention realizes multi-dimensional feature fusion by simultaneously utilizing the shallow features and deep features of the image, which not only improves the accuracy of classification, but also enhances the adaptability to multi-category elements in complex scenes.

Description

Technical Field

[0001] The present invention relates to the technical field of image element recognition, and in particular to a method for quantitatively analyzing spatial elements of urban high-intensity areas based on image data. Background Art

[0002] With the rapid development of computer technology, the classification of spatial elements in urban imagery has gradually shifted from manual interpretation to automated processing. Currently, automated classification methods based on image processing technology are becoming increasingly diverse, primarily including supervised classification, unsupervised classification, and object-oriented classification.

[0003] Supervised classification relies on high-quality, pre-labeled samples, guiding the classification model to classify features in unknown areas by learning from known samples. However, the results of supervised classification are highly dependent on the representativeness and accuracy of the samples. When the number of samples is insufficient or the labeling quality is poor, the classification accuracy can drop significantly, and classification bias is prone to occur. Unsupervised classification uses an algorithm to automatically cluster image pixels without pre-labeling samples. However, the high computational resource requirements of unsupervised classification limit its application in image processing of large-scale urban areas.

[0004] Object-oriented classification is a method based on image segmentation. It divides an image into a series of small, homogeneous regions (i.e., "objects") and then classifies these objects. This method effectively combines spectral and geometric information, but the classification results are susceptible to the quality of the image segmentation algorithm. Over-segmentation or under-segmentation can directly affect the final classification accuracy, causing the classification results to deviate from the actual situation.

[0005] In contrast, deep learning-based classification methods can deeply explore the deep features in images and show significant advantages in urban image classification. Through multi-scale feature extraction and contextual information fusion techniques, deep learning models can effectively distinguish spatial elements such as buildings, vegetation, and roads in complex scenes. However, deep learning methods mainly focus on deeply exploring the characteristics of spatial elements themselves, and insufficiently utilize the association knowledge between elements and specific identification features. In scenes with multiple elements, this can easily lead to poor classification accuracy and inconsistent classification results. Summary of the Invention

[0006] 1) Technical issues solved

[0007] The present invention provides a quantitative analysis method of spatial elements in urban high-intensity areas based on image data, which can solve the problems of poor classification accuracy and inconsistent classification results in scenes with mixed multiple elements.

[0008] 2) Technical solution

[0009] To achieve the above-mentioned object, the present invention provides the following technical solution: a method for quantitatively analyzing spatial elements of urban high-intensity areas based on image data, comprising the following steps:

[0010] Obtain street view panoramas from selected observation points in high-intensity urban areas;

[0011] Based on edge detection, color analysis, and texture analysis, shallow feature extraction is performed on the street view panorama, and spatial elements in the street view panorama are preliminarily identified in combination with predefined shallow feature templates;

[0012] For each of the spatial elements initially identified, obtaining its deep features through a deep learning model; wherein the deep features include the geometric shape and spatial position of each spatial element, and the semantic association relationship between different spatial elements;

[0013] Based on topological analysis, a topological graph between spatial elements is constructed, wherein each node of the topological graph represents each spatial element that has been initially identified, and each node carries its shallow and deep features, and the edges of the nodes represent the semantic associations between the spatial elements. The structure of the topological graph is optimized using preset logical rules;

[0014] The constructed topological map is input into a pre-trained classification model, and the classification model outputs the classification result of each spatial element in the street view panorama and its spatial distribution map according to the node features and the topological structure.

[0015] Furthermore, before shallow feature extraction is performed on the acquired street view panorama, image preprocessing is performed on the street view panorama. The image preprocessing includes denoising, image enhancement and image segmentation operations, corresponding to:

[0016] removing random noise from the street view panorama and retaining edge information;

[0017] The brightness and contrast of the street view panorama after noise removal are analyzed, histogram equalization is performed on the local area of the image, and contrast enhancement is performed on the low-contrast area.

[0018] Furthermore, edge detection, color analysis, and texture analysis are performed on the street view panorama to extract the shallow features. Specifically:

[0019] Perform Canny edge detection on the acquired street view panorama. By calculating the intensity of grayscale changes in the street view panorama, the boundaries of objects in the street view panorama are located using gradient information.

[0020] Extracting the color distribution characteristics of each of the objects located by HSV color space;

[0021] The gray-level co-occurrence matrix is used to extract the surface texture information of each located object.

[0022] Furthermore, each of the objects located in the street view panorama and the shallow features extracted therefrom are matched with the preset shallow feature template, and the spatial elements represented by each object are preliminarily identified through a matching algorithm.

[0023] Furthermore, for each of the spatial elements preliminarily identified, the corresponding deep features are extracted through a deep learning semantic segmentation model, including the geometric shape and spatial position of each spatial element, the relative distance between different spatial elements, and the contextual semantic association relationship between different spatial elements.

[0024] Furthermore, the preset logic rules include:

[0025] Adjusting the weight values of the edges based on the geometric shapes and spatial positions of the spatial elements preliminarily identified, as well as the relative distances between different spatial elements; wherein the projected distances between different spatial elements are calculated, a distance threshold is set, and the weights are adjusted proportionally;

[0026] The weight value of the edge is adjusted according to the semantic association relationship between different spatial elements; if two spatial elements have a functional association in the real scene, the weight is increased; if the two spatial elements have no functional association, the edge of the node represented by the two spatial elements is deleted.

[0027] Furthermore, by optimizing the preset logical rules, the weights of the edges are updated, and a graph convolution operation is performed on the topological graph, thereby optimizing the node features and spatial distribution in the topological graph.

[0028] Furthermore, the pre-trained classification model classifies the spatial elements of the high-intensity area based on deep learning technology. During the training process of the classification model, the topological map is input and the corresponding feature vector of each node is generated according to the shallow features and deep features of each node.

[0029] The classification model uses a message passing mechanism to aggregate the features of the target node's neighbor nodes based on the adjacency relationship and edge weight of the target node to be identified, and fuses them with the current features of the target node to generate the target node features containing context information;

[0030] The feature vector of each node is input into the defined classifier to output the classification result of each node. The classification result of each node is combined with the spatial position relationship in the topological map to generate a spatial distribution map of spatial elements in the street view panorama.

[0031] 3) Beneficial effects:

[0032] Compared with the prior art, this invention has the following beneficial effects:

[0033] The present invention preliminarily determines the spatial elements in the image by jointly extracting shallow and deep features in the street view panorama, constructs a topological map of the spatial elements, and adjusts the edge weights of the topological map through preset logical rules. This enhances the model's adaptability to the complex correlations between spatial elements in actual street scenes and is suitable for complex and diverse spatial structures in high-intensity areas of cities.

[0034] Through the pre-trained deep learning classification model, the characteristics of spatial elements are combined with their topological relationships, and the message passing mechanism of adjacent node features and weights is used to optimize the node feature expression, so that the classification model can effectively capture contextual information, thereby improving the reliability and accuracy of the classification results. Finally, the output is a combination of the classification results and the spatial distribution map, which intuitively shows the distribution of each spatial element. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A schematic diagram of the process of a method for quantitatively analyzing spatial elements of high-intensity urban areas based on image data provided by an embodiment of the present invention;

[0036] Figure 2 A schematic diagram of multiple target identification spatial elements in a high-intensity area in the method for quantitative analysis of spatial elements in urban high-intensity areas based on image data provided by an embodiment of the present invention;

[0037] Figure 3 A schematic diagram of a topological map of multiple spatial elements in an image constructed using topological technology in a method for quantitatively analyzing spatial elements in high-intensity urban areas based on image data provided by an embodiment of the present invention;

[0038] Figure 4 This is a schematic diagram of a topological map optimized based on preset logical rules in the quantitative analysis method of urban high-intensity area spatial elements based on image data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0040] In the description of the present invention, it should be understood that the terms "longitudinal", "transverse", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0041] In addition, the terms "first", "second", etc., if used, are merely used to distinguish and describe, and should not be understood as indicating or implying relative importance.

[0042] It should be noted that, in the absence of conflict, the features in the embodiments of the present invention may be combined with each other.

[0043] With the rapid development of computer technology, the classification of spatial elements in urban imagery has gradually shifted from manual interpretation to automated processing. Methods for automated classification using image processing techniques are becoming increasingly diverse, primarily including supervised classification, unsupervised classification, and object-oriented classification. Each of these methods has its own unique characteristics in practical applications.

[0044] Among them, supervised classification is one of the most widely used methods. It relies on high-quality, pre-labeled samples. That is, it guides the classification model to classify features in unknown areas by learning from samples of known spatial features. However, the results of supervised classification are highly dependent on the representativeness and accuracy of the samples. When the sample quality is inaccurate or the sample quantity is insufficient, classification bias is prone to occur.

[0045] Unsupervised classification uses an algorithm to automatically cluster image pixels, without pre-labeling samples, making it a relatively independent classification method. However, unsupervised classification algorithms require a pre-set number of categories and rely heavily on the quality of the resulting clustering. Furthermore, the complexity of unsupervised classification, its slow convergence rate, and the long computational time required by computers limit its application in large-scale urban image processing.

[0046] In addition to the two classification methods mentioned above, object-oriented classification is a commonly used method based on image segmentation. It divides an image into a series of small, homogeneous regions (i.e., "objects") and then classifies these objects. This method can effectively combine the spectral and geometric information of the image, but the classification results are easily affected by the quality of the image segmentation algorithm. Over-segmentation or under-segmentation will directly affect the final classification accuracy, causing the classification results to deviate from the actual situation. Therefore, object-oriented classification methods are also not suitable for processing images in large urban areas.

[0047] With the rapid development of artificial intelligence, traditional image processing techniques are gradually being replaced by machine learning methods in the classification of spatial elements in urban area images.

[0048] Traditional machine learning methods primarily employ algorithms such as decision trees, random forests, or maximum likelihood methods. These methods rely primarily on shallow image features, leveraging directly observable features such as color, texture, and shape within an image region to perform classification. However, these methods have limitations in extracting and utilizing deeper features within an image, resulting in insufficient adaptability of classification results for complex scenes, particularly when there are fuzzy boundaries or complex relationships between spatial elements.

[0049] In contrast, deep learning-based methods can fully exploit the deep features in images and have been widely used in urban image classification in recent years. By constructing a multi-layer network structure, deep learning can analyze the high-level semantic information of images and significantly improve classification accuracy. Among them, network structures such as U-Net, Pyramid Scene Parsing Network (PSP Net), and DeepLabv3+ are widely used in spatial feature classification tasks. Through techniques such as multi-scale feature extraction and contextual information fusion, they can effectively distinguish between elements such as buildings, vegetation, and roads in complex scenes. However, deep learning methods mainly focus on in-depth exploration of the characteristics of the elements themselves and lack the use of knowledge related to the relationship between specific elements and their specific identification features. This may lead to insufficient interpretability and consistency of classification results in scenes with mixed features.

[0050] In summary, the evolution of spatial feature classification technology in urban imagery, from traditional image processing to artificial intelligence methods, has greatly improved the automation and accuracy of classification. However, each method still faces limitations and challenges, and further improving classification efficiency, accuracy, and applicability remains an important research direction.

[0051] In order to solve the above problems, the inventors proposed a method of combining Figures 1 to 4 The quantitative analysis method of spatial elements in urban high-intensity areas based on image data is shown. This method not only improves the accuracy and efficiency of quantitative analysis of spatial elements in urban high-intensity areas, but also has strong applicability and scalability, providing more intelligent technical support for urban planning and management.

[0052] Specifically, first, perform S1: obtain a street view panorama from an observation point in a selected high-intensity area of the city.

[0053] Based on the characteristics of high-intensity urban areas, such as high building density, large population density, and heavy traffic flow, observation points are selected in combination with the following factors:

[0054] Geographical location distribution: Priority will be given to representative areas such as commercial areas, transportation hubs, and densely populated areas.

[0055] Scene coverage: Observation points must be evenly distributed to ensure that the street view panorama covers the main roads, squares and key facilities in the area.

[0056] Functional requirements: Based on specific analysis needs, select areas that can fully reflect various spatial elements such as buildings, roads, plants, and street lights.

[0057] For example, in a city's commercial center, subway station entrances and exits, main road intersections, and commercial plazas are selected as observation points to ensure that crowded areas and key traffic nodes are covered, and to obtain street view panoramic images of the corresponding observation points.

[0058] Before extracting features from the acquired street view panorama, image preprocessing is required.

[0059] Specifically, random noise (such as Gaussian noise and salt-and-pepper noise) is removed from street view panoramas, while preserving edge information and preventing it from affecting the accuracy of subsequent feature extraction. In some embodiments of the present invention, noise is removed by calculating local image similarity and using a weighted average method, while preserving edge and texture information.

[0060] For example, in the aforementioned street view image of a commercial center, the non-local mean filter is applied to remove noise, and the clarity of the edge area is improved, thereby retaining the main details of the building outline.

[0061] The image enhancement operation is to improve the brightness, contrast and visual clarity of the image, improve the visibility of low-light or local low-contrast areas, and provide higher-quality input for subsequent feature extraction. In some embodiments of the present invention, global histogram equalization is performed on the entire image to redistribute the grayscale and improve the image contrast, especially the details of the highlights and shadows. Alternatively, adaptive histogram equalization is used to process local areas of the image to avoid over-enhancement that may be caused by global equalization. For areas affected by shadows in street view images, gamma correction is used to adjust the brightness. Laplace filtering or high-pass filtering is performed on the enhanced image to highlight the edge features in the image and improve the clarity of structures such as building outlines and road dividing lines.

[0062] For example, after CLAHE processing of a street view panorama along a road, the street lights and road signs that were originally unclear in the shadows become visible, the contrast is improved, and more usable information is provided for the classification model input.

[0063] In some embodiments of the present invention, the noise-filtered and enhanced street view panorama is segmented into multiple small blocks, ensuring that each block contains local information, facilitating subsequent feature extraction and classification analysis of specific spatial elements. A grid partitioning method is used to divide the street view panorama into multiple blocks of a fixed size (e.g., 128×128 pixels), with each block serving as the basic unit for feature extraction. Alternatively, a K-means clustering algorithm is applied to the street view image, clustering different regions based on pixel values (color, brightness) and spatial distribution, ensuring that pixels within each region are similar and possess specific local features.

[0064] In summary, it can be understood that through the above preprocessing and segmentation operations, the processed street view panorama has high quality and clarity, and the local segmentation blocks can accurately represent the spatial elements in the image, providing reliable input data for subsequent feature extraction and classification models.

[0065] After preprocessing the street view panorama, S2 is performed: shallow feature extraction is performed on the street view panorama based on edge detection, color analysis, and texture analysis. Combined with predefined shallow feature templates, the spatial elements in the street view panorama are initially identified. And S3: For each initially identified spatial element, the deep features are obtained through a deep learning model. The deep features include the geometric shape and spatial location of each spatial element, as well as the semantic relationships between different spatial elements.

[0066] First, edge detection, color analysis, and texture analysis are performed on the street view panorama to extract shallow features. These shallow features primarily extract basic visual information related to objects, including edges, color, and texture, from the pixel level of the street view panorama. These shallow features can reflect the outlines and local structure of objects in the image and serve as an important foundation for subsequent spatial feature classification.

[0067] Specifically, by detecting areas of grayscale abrupt changes in an image, contour information reflecting the boundaries and shapes of objects is extracted. In some embodiments of the present invention, the Canny algorithm is used to calculate the gradient magnitude and direction of each pixel to identify areas with the greatest grayscale change. Non-maximum suppression is applied to preserve local extreme values of edges, and dual-threshold segmentation (high and low thresholds) is used to connect edges, thereby extracting the edges of object elements in the image.

[0068] Taking the above example again, after applying Canny edge detection to the street view image of a commercial street, the building outlines, street light poles, traffic light poles and road boundaries were successfully extracted. The edge detection results were visually clear and had little noise impact.

[0069] By extracting color distribution features, the color information of objects in street view panoramas is described, providing support for classification. In some embodiments of the present invention, the image is converted from the RGB color space to the HSV color space, separating the image's hue (H), saturation (S), and brightness (V). The HSV space better aligns with human color perception and is easier to process. The color histograms of each image region are then calculated, and the distribution of different color components is statistically analyzed. Regarding the histogram dimension, each channel can be divided into 256 grayscale levels. A sliding window technique is then used to extract the color distribution features of the local region.

[0070] For example, HSV analysis is performed on the green belt area in a street scene to identify the main hue (green) of the plant area, and the saturation and brightness differences of different plant types are quantified through histograms.

[0071] By analyzing local pixel patterns in an image, texture features reflecting information such as surface roughness and structural complexity are extracted. In some embodiments of the present invention, a gray-level co-occurrence matrix (GLCM) is used to calculate the spatial relationships between grayscale value pairs in an image. For example, texture analysis of a road area in a street view image using the GLCM reveals that the road surface texture is highly uniform (homogeneous), while the sidewalk texture has high contrast (brick-like structure).

[0072] In summary, it can be understood that the acquired street view panorama is subjected to denoising and enhancement operations, followed by edge detection, color analysis, and texture analysis to extract the contours of objects such as buildings, plants, roads, street lights, traffic lights, motor vehicles, and non-motor vehicles; extract the color distribution characteristics of objects such as buildings, plants, roads, street lights, traffic lights, motor vehicles, and non-motor vehicles; and extract the texture patterns of objects such as building surfaces and road surfaces.

[0073] Through the above steps, shallow features reflecting the outline, color and local structure of objects in the street view panorama are obtained, laying the foundation for subsequent feature registration and classification based on topological analysis.

[0074] After extracting shallow features, we extract deep features of the street view panorama based on the semantic segmentation model. Deep features represent high-level semantic information in the image and typically include:

[0075] Geometric shape: The overall outline and three-dimensional structure of an object, such as the shape of a building or the outline of a motor vehicle;

[0076] Spatial position: the specific location of an object in the image and its relative positional relationship with other objects;

[0077] Contextual relationships: The semantic connection between an object and its surroundings, such as the relationship between traffic lights and roads, or buildings and people.

[0078] Deep features are automatically learned from data using deep learning models. Deep learning models typically rely on convolutional neural networks (CNNs) and their extended architectures, such as semantic segmentation models, to extract these features. Common semantic segmentation model architectures include, but are not limited to, U-Net, DeepLabv3+, and PSP Net.

[0079] It can be understood that the geometric shape, spatial location and contextual relationship of deep features provide the input of the classification model for subsequent topological analysis and logical rule alignment.

[0080] The following is a table for reference Figure 2 , the target spatial elements of high-intensity urban areas will be integrated with shallow features, deep features and set logical rules.

[0081]

[0082] Then S4 is performed: Based on topological analysis, a topological graph between spatial elements is constructed, where each node of the topological graph represents each spatial element that is initially identified, and each node carries its shallow and deep features. The edges of the nodes represent the semantic association relationship between spatial elements, and the structure of the topological graph is optimized through preset logical rules.

[0083] Specifically, first, a topological graph is constructed, where each node represents an object element in the street view panorama. Figure 3 and Figure 4 ,It should be noted that the types of object elements are ,more than the types of target spatial elements to be ,recognized., Nodes contain shallow features of objects (such as edges, ,color, texture, etc.) and deep features (geometric shape, spatial position, ,contextual relationships), which are represented by specific data ,structures.

[0084] The edges between nodes represent the spatial relationship between object elements. The edge weight reflects the strength of the topological relationship and is calculated by combining the following identified features:

[0085] Spatial position: the distance between adjacent object elements (such as the shortest distance between a building and a road);

[0086] Color similarity: measures the closeness of colors by color space distance (such as HSV or Lab space);

[0087] Semantic relevance: obtained through semantic matching of object categories by semantic segmentation models. For example, buildings are usually adjacent to roads, and street lights are usually close to roads.

[0088] In addition, some embodiments of the present invention can also calculate the relative distance between different spatial elements. Relative distance generally refers to the geometric distance between different spatial elements or the distance measured by some metric. The deep learning semantic segmentation model can further calculate the relative distance between each spatial element by accurately locating the position of each spatial element.

[0089] For example, if there is a clear gap between the building and the road in the image, or there is a specific layout relationship between the building and the street lamp, by extracting the geometric information and spatial position of the spatial elements, the relative distance between the building and the road can be calculated, and it can serve as an important reference information in topological analysis.

[0090] In summary, after extracting shallow and deep features between spatial elements to construct the nodes of the topological graph, pre-set logical rules are used to define the connection relationships between spatial elements in the topological graph. These rules are then used to optimize the graph structure. Specifically, the core of the pre-set logical rules is to adjust the edge weights between nodes in the topological graph based on the geometric shape, spatial position, relative distance, and semantic association of spatial elements. For reference, the logical rules are summarized as follows.

[0091] First, the geometric shape and spatial position of the spatial elements extracted by the deep learning model are used to calculate the projection distance between different spatial elements. The projection distance refers to the distance between two spatial element objects on a plane, usually considering the distance between the closest edges or center points of the two objects.

[0092] After obtaining the projected distance between spatial elements, some embodiments consider setting a distance threshold to control whether to adjust edge weights. For example, if the projected distance between two spatial elements is less than the threshold, it indicates a strong spatial association between them, and the edge weight will be increased. Conversely, if the distance is greater, it indicates a weak association between the two spatial elements, and the corresponding edge weight will be reduced.

[0093] In addition to spatial location, the semantic relationships between spatial elements are also important factors in adjusting edge weights. Semantic relationships refer to the functional connections between different spatial elements in the real environment. For example, referring to the table above, buildings and roads are generally closely semantically related, as buildings are often located along roads, and roads provide traffic access for buildings. Streetlights also have a semantic relationship with roads, as they are generally installed on both sides of roads to provide nighttime illumination.

[0094] For spatial elements with functional associations, the weight of their connecting edges should be increased. For spatial elements without functional associations, such as buildings and street lights, if there is no direct functional connection, the edge between the two spatial elements should be deleted. Figure 3 and Figure 4 ,There are 13 nodes in the initial topology graph, which represent different spatial elements, such as buildings, roads, green spaces, street lights, etc. The goal of the optimization is to adjust the weights of the edges and delete the edges between nodes that have no functional association, thereby streamlining the graph structure and improving the practical significance and classification efficiency of the topology graph.

[0095] Assume that the initial topology is as shown in the following table.

[0096]

[0097] During the topology optimization process, the functional context relationship is determined:

[0098] There is a functional relationship between the building and the road. The building needs to provide transportation connection through the road, so this edge is maintained.

[0099] There is usually no direct functional relationship between buildings and streetlights (except in special cases, such as building lighting relying on streetlights), so this side should be deleted.

[0100] The road and the parking lot have a functional relationship, and the parking lot needs to be connected to the road, so this edge is maintained.

[0101] The streetlight has a functional relationship with the square (it usually illuminates the square), so this side is retained.

[0102] The bus stop has a functional relationship with the road, and the bus stop is usually located next to the road, keep this side.

[0103] The topology nodes after optimization are shown in the following table.

[0104]

[0105] In some feasible embodiments of the present invention, a rule set may be established in advance to guide the adjustment of weights. For example, the rule set stipulates that the connection between buildings, roads, and streetlights has a high weight, while the connection between vehicles and streetlights has no significant functional relationship and therefore has a zero or low weight.

[0106] Under the influence of pre-set logical rules, edge weights in the topological graph are adjusted to reflect the true relationships between spatial elements. On this basis, a graph convolutional network (GCN) operation is performed to further optimize the node features and spatial distribution in the topological graph. The graph convolution operation updates each node's features by weightedly aggregating the features of its neighboring nodes. This process helps improve the representational capabilities of the topological graph, especially for spatial element relationships in large, high-density urban areas.

[0107] Specifically, the edge weight is updated by the feature correlation of neighboring nodes, so that edges with stronger semantic and spatial correlation have higher weights. The update formula is:

[0108]

[0109] in, For the Layer nodes The eigenvector of For nodes The set of neighbor nodes of is the learnable weight matrix of the graph convolutional layer; For nodes and degree; is a nonlinear activation function, such as ReLU.

[0110] For each node , by aggregating its neighbor nodes The features of the node are updated based on the features of the node. Aggregation methods typically include weighted average and weighted sum. For each edge, the weight of the edge affects the degree of aggregation of neighboring node features. Edges with larger weights have a greater impact on neighboring node features, while edges with smaller weights have a reduced impact on neighboring node features. Based on the aggregation results and combined with the original node features, a nonlinear activation function is used to update the node features. The updated node features can better capture the relationships and topological structure between spatial elements.

[0111] The results of graph convolution update the node features in the topological graph, thereby optimizing the structure of the topological graph. Specifically, after the graph convolution operation, the updated node features can make the topological graph more consistent with the actual spatial relationships. For example, the connections between spatial elements such as buildings, roads, and streetlights will become clearer. The adjustment and update of edge weights ensure that the actual semantic connections between these spatial elements are properly reflected.

[0112] The optimized topological map obtained after graph convolution can provide more accurate input features for subsequent classification tasks, helping the classification model to better identify spatial elements and generate accurate spatial distribution maps.

[0113] In summary, it can be understood that by adjusting the weights of edges based on geometric shapes, spatial positions, semantic associations, etc., and optimizing the topological map in combination with graph convolution operations, the modeling of the relationships between spatial elements can be effectively improved, ensuring that the topological map can accurately reflect the spatial structure in the actual scene. The optimized topological map not only improves the classification accuracy of spatial elements, but also enhances the interpretability and consistency of the classification results, providing strong support for subsequent urban street scene analysis and spatial element distribution prediction.

[0114] Finally, S5: Input the constructed topology map into a pre-trained classification model. The classification model outputs the classification results of each spatial element in the street view panorama and its spatial distribution map based on the node features and topology structure. The core goal of this step is to use the pre-trained classification model. The model learns the relationship between node features and topology structure, combines the feature space and the contextual semantic information of the spatial elements, and classifies each node. Ultimately, it outputs the category label of each spatial element in the street view panorama and its spatial distribution in the image.

[0115] The pre-trained classification model combines deep learning techniques with topological information to efficiently classify spatial elements based on both shallow and deep features. Using a multi-layer neural network and message passing mechanism, the model accurately identifies and classifies spatial elements in street view panoramas and generates a spatial distribution map of these elements.

[0116] Regarding the training process of the classification model, it starts with the input topology map. Through the deep learning classification model, the feature vector of each node is extracted from the original spatial element description. This feature vector is passed to the classification model as the representation vector of the node.

[0117] Specifically, referring to the above-mentioned graph neural network, the adjacency relationship and node features in the graph structure are used to aggregate information layer by layer. The output of each layer updates the node features. During the message transmission process, the edge weights are used to adjust the contribution of neighbor nodes to the target node, gradually integrating shallow information into high-level semantics.

[0118] Each node's feature vector is then fed into a predefined classifier. The classifier classifies the spatial element based on each node's feature vector and outputs the probability distribution of the node's category through a fully connected layer and a soft max function, thereby outputting the node's category label.

[0119] Once all node classification results are output, the classification model combines each node's classification label with its spatial location to generate a spatial distribution map of the spatial elements in the street view panorama. This spatial distribution map not only reflects the classification of each spatial element, but also shows how it is distributed in geographic space.

[0120] In summary, it can be understood that this deep learning-based classification model, combined with node features and adjacency relationships in the topological map, allows spatial feature classification to not only accurately identify individual features but also efficiently classify them based on spatial location and contextual information. In this process, the message passing mechanism and feature fusion technology effectively capture the relationships between spatial features, ultimately outputting a more accurate spatial distribution map. This method not only improves classification accuracy but also enhances the spatial consistency and interpretability of the classification results, providing strong technical support for urban streetscape analysis.

[0121] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. The scope of patent protection of the present invention shall be based on the claims. Any equivalent structural changes made using the description and drawings of the present invention shall be included in the scope of protection of the present invention.

Claims

1. A quantitative analysis method of spatial elements in urban high-intensity areas based on image data, characterized by: The following steps are involved: Obtain street view panoramas from selected observation points in high-intensity urban areas; Based on edge detection, color analysis, and texture analysis, shallow feature extraction is performed on the street view panorama, and spatial elements in the street view panorama are preliminarily identified in combination with predefined shallow feature templates; For each of the spatial elements initially identified, obtaining its deep features through a deep learning model; wherein the deep features include the geometric shape and spatial position of each spatial element, and the semantic association relationship between different spatial elements; Based on topological analysis, a topological graph between spatial elements is constructed, wherein each node of the topological graph represents each spatial element preliminarily identified, and each node carries its shallow and deep features, and the edges of the nodes represent the semantic association relationship between the spatial elements. The structure of the topological graph is optimized by preset logical rules; the preset logical rules include: adjusting the weight value of the edge based on the geometric shape and spatial position of the preliminarily identified spatial elements, and the relative distance between different spatial elements; wherein, the projection distance between different spatial elements is calculated, a distance threshold is set, and the weight is adjusted proportionally; the weight value of the edge is adjusted according to the semantic association relationship between different spatial elements; wherein, if two spatial elements have a functional association in a real scene, the weight is increased; if the two spatial elements have no functional association, the edge of the node represented by the two spatial elements is deleted; The constructed topological map is input into a pre-trained classification model, and the classification model outputs the classification result of each spatial element in the street view panorama and its spatial distribution map according to the node features and the topological structure.

2. The method for quantitative analysis of spatial elements of urban high-intensity areas based on image data according to claim 1 is characterized in that: Before performing shallow feature extraction on the acquired street view panorama, image preprocessing is performed on the street view panorama. The image preprocessing includes denoising, image enhancement, and image segmentation operations, corresponding to: removing random noise from the street view panorama and retaining edge information; The brightness and contrast of the street view panorama after noise removal are analyzed, histogram equalization is performed on the local area of the image, and contrast enhancement is performed on the low-contrast area.

3. The method for quantitative analysis of spatial elements of urban high-intensity areas based on image data according to claim 1 is characterized in that: Edge detection, color analysis, and texture analysis are performed on the street view panorama to extract the shallow features. Specifically: Perform Canny edge detection on the acquired street view panorama. By calculating the intensity of grayscale changes in the street view panorama, the boundaries of objects in the street view panorama are located using gradient information. Extracting the color distribution characteristics of each of the objects located by HSV color space; The gray-level co-occurrence matrix is used to extract the surface texture information of each located object.

4. The method for quantitative analysis of spatial elements of urban high-intensity areas based on image data according to claim 3 is characterized in that: Each of the objects located in the street view panorama and the shallow features extracted therefrom are matched with the preset shallow feature template, and the spatial elements represented by each object are preliminarily identified through a matching algorithm.

5. The method for quantitative analysis of spatial elements of urban high-intensity areas based on image data according to claim 1 is characterized in that: For each of the spatial elements initially identified, the corresponding deep features are extracted through a deep learning semantic segmentation model, including the geometric shape and spatial position of each spatial element, the relative distance between different spatial elements, and the contextual semantic association relationship between different spatial elements.

6. The method for quantitative analysis of spatial elements of urban high-intensity areas based on image data according to claim 1 is characterized in that: By optimizing the preset logical rules, updating the weights of the edges, and performing a graph convolution operation on the topological graph, the node features and spatial distribution in the topological graph are optimized.

7. The method for quantitative analysis of spatial elements of urban high-intensity areas based on image data according to claim 1 is characterized in that: The pre-trained classification model is based on deep learning technology to classify the spatial elements of the high-intensity area. During the training process of the classification model, the topological map is input and the corresponding feature vector of each node is generated according to the shallow features and deep features of each node; The classification model uses a message passing mechanism to aggregate the features of the target node's neighbor nodes based on the adjacency relationship and edge weight of the target node to be identified, and fuses them with the current features of the target node to generate the target node features containing context information; The feature vector of each node is input into the defined classifier to output the classification result of each node. The classification result of each node is combined with the spatial position relationship in the topological map to generate a spatial distribution map of spatial elements in the street view panorama.

Images