Urban high-intensity district security assessment method based on multi-source space element coupling

Abstract

The invention discloses an urban high-intensity district security assessment method based on multi-source space element coupling, and relates to the technical field of environment assessment, and the method comprises the steps: selecting a high-intensity district of a research city, and obtaining a plurality of observation points; generating a street view panorama of the position of each observation point; for the streetscape panorama of each observation point, space elements are extracted from each streetscape panorama based on image semantic recognition to obtain objective perception indexes, and subjective perception indexes of sight lines of different participating individuals facing different space elements are obtained based on a sight line capture technology; and coupling the objective perception index and the subjective perception index obtained at each observation point based on a multiple linear regression equation to obtain a security evaluation value. According to the method, the physiological basis of subjective psychological security is directly quantified by acquiring the physiological data of the participants facing the spatial elements, objective spatial features and subjective physiological reactions are coupled, and the complex perception of people on the spatial security is reflected more truly.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental assessment, and in particular to a method for assessing sense of security in high-intensity urban areas based on coupling of multi-source spatial elements. Background Art

[0002] Urban high-intensity areas are densely populated, three-dimensionally complex, frequently interactive, and environmentally complex areas in the city. They exhibit the significant characteristics of high construction intensity, high functional intensity, and high traffic intensity. They are key areas for urban development, and their public spaces are the core areas that showcase urban characteristics and reflect urban vitality. In the development and construction of rapid urbanization, the diverse and dense flow of people, the organization of diverse and overlapping activities, and the multi-dimensional and complex use of space have gradually led to multiple safety hazards in high-intensity areas, such as physiological health, public safety, and disasters and accidents. These issues have always been the focus of multiple disciplines such as urban planning, urban management, and environmental psychology.

[0003] Early studies usually used the method of constructing a comprehensive indicator system for evaluation. The problem with this approach is that the indicator selection and weight setting are relatively subjective, which leads to inaccurate evaluation results.

[0004] In recent years, existing evaluation technologies have begun to use street view image data and adopt machine learning methods to conduct street safety perception assessment. The assessment is mainly based on objective analysis of image features and does not directly consider the subjective feelings of participants. In addition, street view image data only considers static spatial characteristics and ignores the dynamic impact of environmental complexity on individual psychological perception. Summary of the invention

[0005] 1) Technical issues solved

[0006] The present invention provides a method for evaluating sense of security in high-intensity urban areas based on the coupling of multi-source spatial elements, which solves the problem that the prior art only considers static spatial characteristics and ignores the dynamic impact of environmental complexity on individual psychological perception.

[0007] 2) Technical solution

[0008] To achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements, comprising the following steps;

[0009] Select a high-intensity area in the study city, obtain road network information of the high-intensity area, and select multiple observation points;

[0010] Using the API of the street view map, batch-acquire panoramic images by programming, and generate a street view panoramic image of the location of each observation point;

[0011] For each street view panorama of the observation point, extracting spatial elements of each street view panorama based on image semantic recognition and calculating a plurality of spatial indicators, wherein the spatial indicators are greening rate, sky openness and sight accessibility;

[0012] For each of the street view panoramas, based on the gaze capture technology, the gaze trajectory of the participating individuals is recorded, and the corresponding physiological data is obtained for the spatial elements of the gaze; the physiological data includes the average power of a specific frequency band of brain waves, the baseline level of GSR, the mean value of GSR, and the time domain and frequency domain parameters of HRV; wherein the specific frequency band of brain waves includes the frequency bands of alpha waves, beta waves and theta waves of brain waves;

[0013] According to the spatial indicators of each street view panorama and the physiological data of the participating individuals, the objective perception indicators and subjective perception indicators of each observation point are generated; wherein the objective perception indicators include greening rate, sky openness and sight accessibility; the subjective perception indicators include the average power of a specific frequency band of brain waves, the baseline level of GSR, the mean value of GSR, and the time domain and frequency domain parameters of HRV;

[0014] Based on the multivariate linear regression equation, the objective perception index and the subjective perception index obtained at each observation point are coupled to obtain the sense of security evaluation value of the high-intensity area; wherein different spatial indicators are used as independent variables, and the weight of the spatial element is calculated based on the physiological indicators obtained when the participating individuals face different spatial elements.

[0015] Furthermore, generating a street view panorama of each observation point comprises the following steps:

[0016] Obtaining road network information of the selected high-intensity area;

[0017] Generate a plurality of observation points using road network nodes and road sections; wherein intersections and other areas with large traffic volume are selected as observation points;

[0018] Obtaining the latitude and longitude coordinates of each of the observation points;

[0019] According to the latitude and longitude of the selected observation point, the API is called to obtain the street view panorama of each observation point in turn.

[0020] Furthermore, the extraction of spatial elements of each street view panorama based on image semantics specifically includes the following steps:

[0021] Unifying the size of each of the street view panoramas, and then performing image enhancement processing on the images;

[0022] Through the deep learning image semantic segmentation model, the different spatial elements in each processed street view panorama are classified and labeled.

[0023] Furthermore, for the street view panorama after the spatial element classification, the greening rate, sky openness and sight accessibility in each street view panorama are calculated; specifically:

[0024] The greening rate indicates the proportion of green plants in the street view panorama, specifically, the ratio of the number of green pixels in the image to the total number of pixels in the street view panorama;

[0025] The sky openness indicates the proportion of the sky in the image, reflecting the degree of open space, and is specifically the ratio of the number of sky pixels in the image to the total number of pixels in the street view panorama;

[0026] The sky openness measures the range of the spatial area that can be directly seen from the current observation point, specifically the ratio of the sum of pixels in the non-occluded area of ​​the image to the total number of pixels in the street view panorama.

[0027] Furthermore, the acquisition of the average power of a specific frequency band of the EEG of the participating individuals, the baseline level of GSR, the mean value of GSR, and the time domain and frequency domain parameters of HRV is specifically as follows:

[0028] Obtaining specific EEG frequency bands of the participating individuals through a multi-channel EEG device, including collecting frequency bands of alpha waves, beta waves, and theta waves, calculating their power spectrum density and obtaining the average power of each frequency band;

[0029] Through the GSR device, the baseline level of the skin conductance of the participating individuals is first selected as the reference time value, and then the average conductivity level of the skin of the participating individuals when they are looking at each spatial element is obtained as the mean value;

[0030] The heart rate waveform data of the participants when they are looking at each spatial element is obtained through the HRV device, and then the standard deviation and mean of the RR interval are calculated based on the waveform data to extract the low-frequency and high-frequency power; wherein the RR interval is the time interval between two adjacent R waves.

[0031] Furthermore, the street view panorama of each observation point is played through a VR device. During the playback, each participating individual wears an eye tracking device to record in real time the line of sight of the participating individual when he or she looks at different spatial elements in the street view panorama, and record the gaze time of the participating individual when looking at different spatial elements.

[0032] Furthermore, the timestamps of the multi-channel EEG device, GSR device, HRV device and eye tracking device are synchronized.

[0033] Furthermore, the eye tracking device is used to record the gaze trajectory of the participating individuals when observing the street view panorama, and the trajectory is mapped to the spatial element area after semantic segmentation. The total time and number of times each spatial element is gazed at are counted, and the gaze proportion is calculated as the gaze gaze weight corresponding to each spatial element.

[0034] Furthermore, when the participating individuals watch the street view panorama, the physiological data of the participating individuals are synchronously collected, the physiological data are matched with the spatial elements captured by the line of sight of the participating individuals, and the mean physiological responses of different spatial elements are calculated.

[0035] Furthermore, the comprehensive weight of each spatial element is calculated by combining the gaze weight and the physiological response mean; specifically:

[0036] Normalize the gaze weight W of the participating individuals 注视,j and the mean physiological response R 生理反应,j ;

[0037] Calculate the comprehensive weight β of each spatial element i :

[0038]

[0039] Where n is the total number of spatial elements identified in the street view panorama.

[0040] 3) Beneficial effects:

[0041] Compared with the prior art, the invention has the following beneficial effects:

[0042] The present invention directly quantifies the physiological basis of subjective psychological security by obtaining the physiological data of participants when facing spatial elements (specific frequency bands of EEG, baseline level and mean of GSR, and time domain and frequency domain parameters of HRV), couples objective spatial characteristics with subjective physiological reactions, and more realistically reflects people's complex perception of spatial security.

[0043] Based on the participants' physiological reactions and gaze capture technology, subjective physiological indicators are used as the basis of weights, and the weight distribution of spatial elements is dynamically adjusted from a human-centered perspective. This can better reflect the impact of spatial elements on the real sense of security of participating individuals when facing different spatial elements in high-intensity areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A flow chart of a method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements provided by an embodiment of the present invention;

[0045] Figure 2A schematic diagram of a scene for generating a street view panorama of the location of an observation point in a method for evaluating sense of security in a high-intensity urban area based on coupling of multi-source spatial elements provided by an embodiment of the present invention;

[0046] Figure 3 In the method for evaluating the sense of security in high-intensity urban areas based on the coupling of multi-source spatial elements provided in the embodiment of the present invention, a schematic diagram of a scene is generated showing the sight line trajectories of participating individuals when they gaze at different spatial elements in a street view panorama;

[0047] Figure 4 A schematic diagram of a scenario in which the physiological data of the participating individuals are synchronously collected when the participating individuals watch the street view panoramic image in the method for assessing the sense of security in high-intensity urban areas based on the coupling of multi-source spatial elements provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] 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.

[0049] In the description of the present invention, it should be understood that the terms "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside" and "outside" etc., indicating orientations or positional relationships, are 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, and do not indicate or imply 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 a limitation on the present invention.

[0050] 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.

[0051] Combination Figures 1 to 4 The method for assessing the sense of security in high-intensity urban areas based on the coupling of multi-source spatial elements shown in the figure first requires the selection of high-intensity urban areas. The definition of high-intensity areas is based on urban planning data, and selects areas with high population density, dense buildings, complex functions, and busy traffic. The data source can be statistical data from government planning departments, satellite image data, or socioeconomic data.

[0052] Obtain road network data through open data platforms (such as OpenStreetMap, OSM), and use their API to download vector road network information (including road nodes, paths, etc.) of high-intensity areas. If more detailed data is needed, detailed maps published by the city transportation department can be used.

[0053] Use GIS software to convert the road network format (such as GeoJSON, Shapefile) to ensure data compatibility. After the format conversion, perform topological analysis to identify road intersections, path lengths, and road grades, providing a basis for subsequent observation point selection.

[0054] Regarding the principles for selecting observation points, it is necessary to ensure that the observation points are evenly distributed and cover the entire high-intensity area, and give priority to areas with dense traffic (such as subway stations, commercial centers) and key intersections.

[0055] Based on the above, in some feasible embodiments of the invention, reference is made to Figure 2 Based on the road network, the selected high-intensity area is divided into several grids, and one or more intersections are selected as observation points in each grid. This process can be achieved by using GIS tools or Python libraries (such as GeoPandas). Then, the road network nodes are weighted according to the node connectivity, road grade (main road priority) or traffic flow, and high-weight nodes are randomly selected.

[0056] Regarding how to obtain panoramic images of each observation point, in some feasible embodiments of the present invention, the API interface of a street view map service (such as Google Street View, Baidu Map, Mapillary) can be used to obtain panoramic image data at the observation point.

[0057] Specifically, the latitude and longitude coordinates of the selected observation point are converted into query parameters supported by the street view service, and then the API is called, that is, a script is written to send HTTP requests in batches through the API, specify the required parameters (coordinates, viewing angle, image resolution, etc.), and download the returned panoramic images, store the obtained street view panoramic images of the observation points, and record metadata (coordinates, azimuth).

[0058] The following uses Baidu Maps as an example to demonstrate how to use Baidu Maps' API to batch obtain street view images for each observation point through some Python programming.

[0059]

[0060]

[0061] #Get multi-directional images for each observation point

[0062] headings = [0,90,180,270] # Get views of 0°, 90°, 180°, 270°

[0063] for i,loc in enumerate(locations):

[0064] for heading in headings:

[0065] filename="streetview_{i}_{heading}.jpg"

[0066] save_path=os.path.join(SAVE_DIR,filename)

[0067] download_panorama(location=loc,heading=heading,save_path=save_path)

[0068] Through the above technical process, the entire process from area selection to panoramic image generation is completed, providing a data basis for subsequent spatial element extraction and sense of security assessment.

[0069] For the street view panorama of each observation point obtained above, spatial elements are extracted from each street view panorama based on image semantic recognition, and multiple spatial indicators are calculated. The spatial indicators obtained at each observation point are used as objective perception indicators for judging the sense of security in the high-intensity area.

[0070] Specifically, in an embodiment of the present invention, spatial elements in a street view panorama are identified by a semantic segmentation model. Semantic segmentation is to classify each pixel in an image into a predefined category (such as buildings, roads, vegetation, sky, etc.). Commonly used deep learning models include: UNet: lightweight and suitable for small data sets; DeepLab series: supports multi-scale feature extraction and is suitable for complex scenes; and MaskR-CNN: combined with instance segmentation, it can handle semantic segmentation and object detection tasks at the same time.

[0071] It should be noted that the street view image is a circular view, and the panoramic image needs to be expanded into a plane or sliced, such as decomposing into front, back, left, and right views. Here, it is necessary to ensure that the semantic segmentation model can correctly segment each pixel in the circular view. In some feasible embodiments of the present invention, the street view image is unified into the standard size required by the deep learning model (such as 512×512 or 1024×1024), and then random cropping, brightness adjustment, rotation and other data enhancement operations are performed to improve the robustness of the model.

[0072] The deep learning model is used to perform semantic segmentation on the street view panoramic images. Based on the semantic segmentation results, binary mask images of different category areas are generated, and the number of pixels in each category is counted as the basis for classification. Then, the greening rate, sky openness, and line of sight accessibility in each street view panoramic image are calculated based on the image processing results. Among them, the greening rate is the ratio of the number of pixels in the segmented vegetation area to the number of pixels in the whole image; the sky openness is the ratio of the number of pixels in the segmented sky area to the number of pixels in the whole image; the line of sight accessibility is to evaluate the degree of occlusion of the line of sight by the distribution of obstacles in the street view image by combining the depth estimation model with the object segmentation results.

[0073] The following is some Python code to calculate the greening rate, sky openness and line of sight accessibility in each street view panorama.

[0074] #Load the pre-trained DeepLabV3 model

[0075] model=deeplabv3_resnet101(pretrained=True)

[0076] model.eval()

[0077] # Image preprocessing

[0078] def preprocess_image(image_path):

[0079] input_image=Image.open(image_path).convert("RGB")

[0080] preprocess=transforms.Compose([

[0081] transforms.Resize((512,512)),

[0082] transforms.ToTensor(),

[0083] transforms.Normalize(mean=[0.485,0.456,0.406],std=[0.229,0.224,0.225]), ])

[0085] return preprocess(input_image).unsqueeze(0)

[0086] #Semantic Segmentation

[0087] def semantic_segmentation(model,image_tensor):

[0088]

[0089] print(f"Greenery rate:{greenery_rate:.2%}")

[0090] print(f"Sky openness:{sky_openness:.2%}")

[0091] # Display segmentation results

[0092] plt.imshow(segmentation_map)

[0093] plt.title("Semantic Segmentation Map")

[0094] plt.show()

[0095] In summary, it can be understood that richer spatial elements, including visual greening rate, sky openness and visual accessibility, are extracted based on image semantic segmentation, combined with the dynamic physiological responses obtained in the subsequent present invention, so as to achieve a comprehensive assessment of the impact of the spatial environment on psychological security.

[0096] Regarding how to obtain the dynamic physiological response of personnel as the subjective perception index of each observation point, in an embodiment of the present invention, by combining virtual reality (VR) technology and sight capture technology, the sight trajectory of participants when watching the street view panorama is recorded to extract their gaze behavior data on different spatial elements (such as greenery, buildings, sky, etc.). At the same time, by synchronizing physiological signal monitoring equipment, specific frequency bands of electroencephalogram (EEG), galvanic skin response (GSR), and heart rate variability (HRV) parameters are collected to analyze the psychological and physiological states of the participating individuals when they are looking at different spatial elements.

[0097] More specifically, an immersive panoramic scene generated from a street view panorama is played through a VR device, providing a realistic environment for participants to experience naturally. In some feasible implementations of the present invention, the street view panorama can be converted into a 360° spherical view through projection mapping, the downloaded street view image is mapped into a spherical image in a quirectangular form, the converted image is adapted to the projection format of the VR device, the observation point is set, and the initial position of the participating individual is fixed.

[0098] During the playback, each participant wears an eye tracking device to record the eye trajectory of the participant when looking at different spatial elements in the street view panorama in real time, that is, the eye movement trajectory on the panorama. Figure 3 , and records the gaze time of the participants when looking at different spatial elements, that is, the time the line of sight stays stably in a certain area. Eye tracking technology records the movement trajectory of the gaze point in the VR scene by capturing the dynamic position changes of the eyeball.

[0099] More specifically, before the experiment, the device was calibrated using a five-point or nine-point calibration method to ensure that the gaze point was accurately mapped to the spatial position of the panorama. When watching the street view, the gaze trajectory data of the participating individuals was recorded in real time, and the coordinates of each gaze point were matched with the spatial elements in the street view panorama. The gaze point was matched with the semantic segmentation results (such as whether the gaze was on vegetation, buildings or the sky) to obtain the gaze time distribution of each spatial element.

[0100] While recording the eye movement data of the participants, the physiological signals of the participants are also recorded synchronously, including: using multi-channel EEG equipment to record brain wave activities in different frequency bands; using GSR equipment to measure skin conductance levels (reflecting the degree of emotional arousal); using a heart rate belt (such as Polar H10) or a bracelet to collect heart rate data.

[0101] More specifically, the brainwave activities recorded include:

[0102] Alpha waves (8Hz–12Hz): associated with a state of relaxation;

[0103] Beta waves (13Hz–30Hz): associated with focused attention and alertness;

[0104] Theta waves (4Hz–7Hz): associated with deep thinking and emotional changes.

[0105] Here, it is important to note that the EEG signals were synchronized to the eye movement trajectory to determine the changes in EEG when the participants looked at different spatial elements, and the EEG data were filtered to extract specific frequency bands, calculate the power spectral density (PSD), and obtain the average power of each frequency band.

[0106] Using the galvanic skin response (GSR), the baseline level and mean value of skin conductance can represent the quiet state and the degree of emotional fluctuation, respectively. Specifically, the GSR data of 10 seconds before the experiment can be extracted to calculate the baseline value and the average conductance level of each spatial element when the participant is looking at it. By correlating the real-time data with the gaze behavior, the emotional response to the spatial elements can be analyzed.

[0107] Heart rate data is collected using HRV equipment, such as a heart rate belt (such as Polar H10) or a bracelet. For the waveform diagram of the acquired heart rate data, specifically, in the electrocardiogram (ECG) waveform, each heartbeat forms a periodic waveform, including P wave, QRS complex and T wave, wherein the QRS complex represents ventricular depolarization, and the R wave is the highest point in the QRS complex, and the RR interval is the time interval between two adjacent R waves, and the unit is usually milliseconds (ms). In some embodiments, the RR interval is the core data for evaluating the activity of the autonomic nervous system (sympathetic and parasympathetic nerve regulation), wherein the short-term RR interval change reflects the heart rate's ability to respond to instantaneous environmental changes; the long-term RR interval change reflects the overall state of heart health and neural regulation. It can be understood that if the heart rate is faster, the RR interval will be shorter; if the heart rate is slower, the RR interval will be longer.

[0108] In some embodiments of the present invention, in the street view sense of safety assessment, the standard deviation (SDNN) and mean (RMSSD) of the RR interval are calculated to extract HRV parameters as a quantitative indicator of the individual's physiological response to environmental stimuli (spatial elements). Specifically, when an individual looks at a spatial element (such as a green area or a densely built-up area), the HRV device records the heart rate waveform and calculates the standard deviation (SDNN) and frequency domain power (LF, HF) of the RR interval. The HRV response caused by different spatial elements can be used to evaluate the impact of the spatial element on the individual's psychological state, and combined with the line of sight weight, it is used for the final calculation of the sense of safety weight.

[0109] Among them, SDNN overall heart rate volatility, extracts low-frequency and high-frequency power through fast Fourier transform (FFT), where high frequency is related to parasympathetic nerve activity and low frequency is related to sympathetic nerve activity. Through time synchronization, the psychological state of participants at different gaze points is analyzed.

[0110] In some feasible embodiments of the present invention, the overall experimental steps for obtaining subjective perception indicators are:

[0111] Ask participants to watch the street view panoramas from different observation points one by one, and each viewing time is controlled within 1-2 minutes;

[0112] Participants wore VR headsets and eye-tracking devices, while connected to EEG, GSR, and HRV monitoring devices;

[0113] Record eye tracking and physiological signals to generate time series data;

[0114] The spatial elements (such as vegetation and sky) classified by semantic segmentation are matched with the gaze trajectory, and the proportion of the fixation time of each element is calculated;

[0115] The EEG, GSR, and HRV parameters when looking at specific elements were extracted to analyze the impact of spatial elements on psychological and physiological states; for example, when looking at street scenes with a high greening rate, the proportion of participants' alpha waves increased (relaxation); when looking at traffic areas (low visual accessibility), GSR data increased (stress).

[0116] Use multiple regression analysis or machine learning methods to model the impact of different spatial elements on psychological safety.

[0117] The following is some Python code to record the physiological data of participants when they watch the street view panorama.

[0118]

[0119] combined_data.append(data_point)

[0120] # Output integrated data

[0121] print(combined_data)

[0122] By combining VR playback, eye tracking, and physiological signal collection, we can accurately quantify the impact of different spatial elements on an individual’s sense of security, providing efficient data support for subsequent urban planning or psychological research.

[0123] In summary, it is understandable that different spatial elements, such as buildings, roads, vegetation, etc., can be extracted through image semantic recognition. However, the types of spatial elements in different street view panoramas may vary. For example, some images may contain vegetation and buildings, while other images may be dominated by roads and sky. This difference makes it not universal to directly use spatial elements as variables in statistical models, and it is also difficult to compare and model between multiple scenes. In order to solve this problem, the spatial elements are further aggregated into unified spatial indicators, namely the greening rate, sky openness, and line of sight mentioned above, so as to achieve the universality of statistical analysis.

[0124] Specifically, an eye tracker is used to record the gaze trajectory of the participating individuals when observing the street view, and the trajectory is mapped to the spatial element area after semantic segmentation. The total time and number of times the spatial element corresponding to each spatial indicator is gazed at are counted, and the gaze proportion is calculated as the gaze weight corresponding to each spatial indicator.

[0125] When the participants watched the street view panorama, physiological data were collected simultaneously:

[0126] EEG specific frequency band: extract the average power of the frequency bands of alpha, beta and theta waves, representing relaxation, tension and cognitive load states respectively;

[0127] GSR (Galvanic Skin Response): measures the level of emotional arousal, recording baseline levels and mean values;

[0128] HRV (heart rate variability): calculates time domain (such as SDNN, RMSSD) and frequency domain (such as LF / HF) parameters to assess stress status.

[0129] These physiological data are matched with the spatial elements captured by the line of sight, and the comprehensive weight of each spatial element for the participating individuals is calculated:

[0130]

[0131] Among them, R 生理数据,j It represents the physiological response value corresponding to fixating on different spatial elements (such as α wave power or GSR mean), and n is the number of samples of fixating on the spatial element.

[0132] In summary, the gaze weight W 注视,j Provides participants' attention allocation to different spatial elements and the mean physiological response W 注视,j Indicates the actual impact of the element on emotional and physiological states.

[0133] Through normalization and weighted synthesis, the comprehensive weight β of each spatial element is calculated i :

[0134] Calculate the comprehensive weight β of each spatial element i :

[0135]

[0136] Spatial indicators are the aggregation results of spatial elements. The contribution ratio of different spatial elements to a certain spatial indicator is predefined. For example, vegetation mainly affects the greening rate, and buildings and vegetation jointly affect the openness of the sky.

[0137] According to the weight of spatial elements and their contribution ratio to spatial indicators, the weight of each spatial indicator δ is calculated. k :

[0138]

[0139] Among them, σ i，k It is the contribution ratio of the i-th spatial element to the k-th spatial index, which can be calculated through historical data.

[0140] Based on multiple linear regression, the objective perception index and the subjective perception index are coupled to calculate the sense of security evaluation value:

[0141] Y=α1δ 绿化率 +α2δ 天空敞开度 +α3δ 视线可达度 +∈

[0142] Among them, α1, α2 and α3 are regression coefficients obtained by fitting historical experimental data, and ∈ is the random error.

[0143] In summary, it is understandable that in the street view panorama at different observation points, the types of spatial elements may not be completely consistent, while the spatial index is a unified and universal measurement that can be applied to the statistical analysis of all observation points. Through the coupling of objective perception indicators (spatial indicators) and subjective perception indicators (physiological data), it can fully reflect the emotional and physiological reactions of individuals to different spatial environments, and finally obtain the sense of security evaluation value of high-intensity areas, which is applicable to different types of high-intensity areas, such as residential areas and commercial areas.

[0144] The following is an example of gradually collecting data and calculating the final sense of safety evaluation value, taking a street view panorama of a high-intensity area in Beijing as an example.

[0145] An observation point was selected in the CBD area of ​​​​Guomao, and a street view panorama of the observation point was taken. The spatial elements contained in the image were extracted through semantic segmentation: buildings, plants, roads, street lights, traffic lights, motor vehicles and non-motor vehicles. The image proportions are as follows:

[0146] Buildings: 30% of the viewing area;

[0147] Plants: 20% of the viewing area;

[0148] Roads: 25% of the viewing area;

[0149] Street lights: 5% of the viewing area;

[0150] Traffic lights: 3% of the viewing area;

[0151] Motor vehicles: 10% of the viewing area.

[0152] These spatial elements are converted into spatial indicators through the conversion algorithm, and finally the following is obtained:

[0153] Greening rate: 20% (small proportion of plants);

[0154] Sky openness: 42% (largely blocked by buildings);

[0155] Visual accessibility: 68% (regional roads are unobstructed and there are no obvious obstructions).

[0156] Invite 10 participants to watch the street view panorama and simultaneously record the spatial elements of gaze and physiological response data. The data after statistically averaging the total data set is shown in the following table.

[0157]

[0158] The physiological response mean of each spatial element is used, and then normalized and weighted comprehensive processing is performed. Specifically, the gaze weight and the physiological response mean are multiplied to calculate the comprehensive weight to ensure that the sum is 1. The comprehensive weights of each spatial element are shown in the following table.

[0159] Spatial elements architecture plant the way Street Lights Traffic Lights Motor Vehicles Non-motorized vehicles <![CDATA[Comprehensive weight β i > 0.199 0.186 0.152 0.060 0.030 0.067 0.062

[0160] Convert it into spatial indicator weights, and the data are shown in the following table.

[0161] Spatial indicators Greening rate Sky openness Line of sight <![CDATA[Weight δ k > 0.0186 0.741 0.281

[0162] After substituting the weights of each spatial indicator into the above multivariate regression equation, the sense of security evaluation value of the observation point is obtained:

[0163] Y=0.35·δ 绿化率 +0.42·δ 天空敞开度 +0.18·δ 视线可达度 +0.05

[0164] After inserting the corresponding data, the safety evaluation value of the observation point is 0.47 (with two decimal places), which shows that although there are no particularly disturbing factors in the area, it cannot provide a very high sense of security. There may be some regional factors, such as a relatively busy space or a low greening rate, which cause the participants' psychological and physiological reactions in this area to be slightly uneasy.

[0165] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. The patent protection scope of the present invention shall be based on the claims. All equivalent structural changes made using the contents of the description and drawings of the present invention should also be included in the protection scope of the present invention.

Claims

1. A method for assessing sense of security in high-intensity urban areas based on the coupling of multi-source spatial elements, characterized by: The steps include: Select a high-intensity area in the study city, obtain road network information of the high-intensity area, and select multiple observation points; Using the API of the street view map, batch-acquire panoramic images by programming, and generate a street view panoramic image of the location of each observation point; For each street view panorama of the observation point, extracting spatial elements of each street view panorama based on image semantic recognition and calculating a plurality of spatial indicators, wherein the spatial indicators are greening rate, sky openness and sight accessibility; For each of the street view panoramas, based on the gaze capture technology, the gaze trajectory of the participating individuals is recorded, and the corresponding physiological data is obtained for the spatial elements of the gaze; the physiological data includes the average power of a specific frequency band of brain waves, the baseline level of GSR, the mean value of GSR, and the time domain and frequency domain parameters of HRV; wherein the specific frequency band of brain waves includes the frequency bands of alpha waves, beta waves and theta waves of brain waves; According to the spatial indicators of each street view panorama and the physiological data of the participating individuals, the objective perception indicators and subjective perception indicators of each observation point are generated; wherein the objective perception indicators include greening rate, sky openness and sight accessibility; the subjective perception indicators include the average power of a specific frequency band of brain waves, the baseline level of GSR, the mean value of GSR, and the time domain and frequency domain parameters of HRV; Based on the multivariate linear regression equation, the objective perception index and the subjective perception index obtained at each observation point are coupled to obtain the sense of security evaluation value of the high-intensity area; wherein different spatial indicators are used as independent variables, and the weight of the spatial element is calculated based on the physiological indicators obtained when the participating individuals face different spatial elements.

2. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 1 is characterized in that: Generating a street view panorama at the location of each observation point specifically includes the following steps: Obtaining road network information of the selected high-intensity area; Generate a plurality of observation points using road network nodes and road sections; wherein intersections and other areas with large traffic volume are selected as observation points; Obtaining the latitude and longitude coordinates of each of the observation points; According to the latitude and longitude of the selected observation point, the API is called to obtain the street view panorama of each observation point in turn.

3. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 1 is characterized in that: The method of extracting the spatial elements of each street view panorama based on image semantics specifically includes the following steps: Unifying the size of each of the street view panoramas, and then performing image enhancement processing on the images; Through the deep learning image semantic segmentation model, the different spatial elements in each processed street view panorama are classified and labeled.

4. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 3 is characterized in that: For the street view panorama after the spatial element classification, the greening rate, sky openness and sight accessibility in each street view panorama are calculated; specifically: The greening rate indicates the proportion of green plants in the street view panorama, specifically, the ratio of the number of green pixels in the image to the total number of pixels in the street view panorama; The sky openness indicates the proportion of the sky in the image, reflecting the degree of open space, and is specifically the ratio of the number of sky pixels in the image to the total number of pixels in the street view panorama; The sky openness measures the range of the spatial area that can be directly seen from the current observation point, specifically the ratio of the sum of pixels in the non-occluded area of ​​the image to the total number of pixels in the street view panorama.

5. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 1 is characterized in that: The method of obtaining the average power of a specific frequency band of the EEG of the participating individuals, the baseline level of GSR, the mean value of GSR, and the time domain and frequency domain parameters of HRV is specifically as follows: Obtaining specific EEG frequency bands of the participating individuals through a multi-channel EEG device, including collecting frequency bands of alpha waves, beta waves, and theta waves, calculating their power spectrum density and obtaining the average power of each frequency band; Through the GSR device, the baseline level of the skin conductance of the participating individuals is first selected as the reference time value, and then the average conductivity level of the skin of the participating individuals when they are looking at each spatial element is obtained as the mean value; The heart rate waveform data of the participants when they are looking at each spatial element is obtained through the HRV device, and then the standard deviation and mean of the RR interval are calculated based on the waveform data to extract the low-frequency and high-frequency power; wherein the RR interval is the time interval between two adjacent R waves.

6. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 5 is characterized in that: The street view panorama of each observation point is played through a VR device. During the playback, each participating individual wears an eye tracking device to record in real time the line of sight of the participating individual when looking at different spatial elements in the street view panorama, and record the gaze time of the participating individual when looking at different spatial elements.

7. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 5 is characterized in that: The timestamps of the multi-channel EEG device, GSR device, HRV device and eye tracking device are synchronized.

8. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 6 is characterized in that: The eye tracking device is used to record the gaze trajectory of the participating individuals when observing the street view panorama, and the trajectory is mapped to the spatial element area after semantic segmentation. The total time and number of times each spatial element is gazed at are counted, and the gaze proportion is calculated as the gaze weight corresponding to each spatial element.

9. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 8 is characterized in that: When the participating individuals watch the street view panorama, the physiological data of the participating individuals are synchronously collected, the physiological data are matched with the spatial elements captured by the sight of the participating individuals, and the mean physiological responses of different spatial elements are calculated.

10. The method for evaluating sense of security in high-intensity urban areas based on coupling of multi-source spatial elements according to claim 9 is characterized in that: Combining the gaze weight and the physiological response mean, the comprehensive weight of each spatial element is comprehensively calculated; specifically: Normalize the gaze weight W of the participating individuals 注视,j and the mean physiological response R 生理反应,j ; Calculate the comprehensive weight β of each spatial element i : Where n is the total number of spatial elements identified in the street view panorama.

Images