Method for simulating and predicting comfort of urban high-intensity district

Abstract

The invention discloses an urban high-intensity district comfort simulation prediction method, which relates to the technical field of environment evaluation, and comprises the following steps: acquiring a street view panorama of each observation point, and extracting a space element of each street view panorama and physiological data of participating individuals corresponding to the space element through an image semantic recognition technology; generating objective perception indexes and subjective perception indexes, constructing a comfort simulation prediction model, establishing a mapping relation between the objective perception indexes and comfort scores and between the subjective perception indexes and the comfort scores, inputting the perception indexes of the observation points into the model, predicting the corresponding comfort scores, and generating a comfort distribution map in the research area through a spatial interpolation algorithm. According to the method, the mapping relation between the objective perception indexes and the subjective perception indexes is established, the perception indexes of the observation points are input into the model for prediction, and the comfort distribution diagram is generated in combination with the spatial interpolation algorithm, so that the comfort distribution condition of the high-intensity area can be intuitively presented.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental assessment, and in particular to a method for simulating and predicting comfort in high-intensity areas of a city. Background Art

[0002] High-intensity areas, as important, densely populated public spaces within cities, serve as the primary environmental carriers for residents' daily lives, social interactions, and various commercial activities. The comfort of these areas directly impacts residents' daily experiences and behavior patterns, and is closely related to a city's social stability, economic development, and vitality. Therefore, how to scientifically and comprehensively assess environmental comfort in high-intensity areas has become a crucial issue in urban planning and management.

[0003] Previous research has shown that the relationship between environmental factors and comfort in high-intensity areas is complex and multidimensional. This means that the public's perceived comfort level is influenced not only by visual factors but also by a variety of factors, including the area's spatial layout, transportation planning, and greenery. However, traditional methods for assessing environmental comfort often rely on user questionnaires, which rely too heavily on subjective experience and fail to fully and accurately reflect residents' actual experiences in high-intensity areas.

[0004] In recent years, with the rapid development of street scene recognition technology, environmental comfort assessment methods based on street scene images have gradually attracted attention. This method analyzes various visual features of the neighborhood environment, such as street cleanliness, building appearance, greenery ratio, and sky visibility, and considers the impact of different proportions of visual features on residents' comfort. However, this method is limited to objective evaluation of the environment and ignores the psychological and physiological reactions of residents in different spatial environments. The comfort of high-intensity areas is not only influenced by objective factors of the external environment, but also closely related to the subjective perception of individual residents. Summary of the Invention

[0005] 1) Technical issues solved The present invention provides a method for simulating and predicting the sense of comfort in high-intensity urban areas, which can comprehensively consider objective environmental factors and multi-dimensional evaluation of individual subjective perception.

[0006] 2) Technical solution To achieve the above objectives, the present invention provides the following technical solutions: a method for simulating and predicting comfort in high-intensity urban areas, comprising: Obtaining street view panoramas of multiple observation points in the selected high-intensity area, and extracting spatial elements of each of the street view panoramas using image semantic recognition technology; The gaze capture technology is used to record the gaze trajectory of the participants when viewing each of the street view panoramas, and the gaze duration of each of the spatial elements is obtained, and the physiological data of the participants are obtained at the same time; the physiological data include the average power of a specific frequency band of the brain wave, the baseline level of the GSR, the mean value of the GSR, and the time domain and frequency domain parameters of the HRV, wherein the specific frequency band includes Wave, wave and Frequency band of the wave; Calculating the fixation ratio of each spatial element in the street view panorama according to the fixation duration of each spatial element, selecting a preset number of spatial elements with the highest fixation duration ratio as relevant spatial elements, and obtaining physiological data of the relevant spatial elements; The relevant spatial elements of each acquired street view panorama are used as objective perception indicators, and the corresponding physiological data are used as subjective perception indicators; Constructing a comfort simulation prediction model and establishing a mapping relationship between the objective perception index and the subjective perception index and the comfort score; wherein the objective perception index and the subjective perception index are used as input variables of the prediction model, and the prediction model is trained using the participants' subjective ratings of the street view panorama as target values to predict the comfort score of the input variables; The input variables of each observation point are obtained and input into the model to predict the corresponding comfort score, and a comfort distribution map within the study area is generated through a spatial interpolation algorithm.

[0007] Furthermore, the eye movement trajectory of the participants is tracked using gaze capture technology to obtain the gaze duration of each of the spatial elements. Specifically, by recording the gaze points of the participants when viewing the street view panorama, each time the participants gaze at one of the spatial elements, a gaze start time and end time are generated based on the timing information, and the gaze start time and end time are used as the gaze period.

[0008] Furthermore, the gaze capture and the timestamp of the physiological numerical control are obtained synchronously, so as to align the gaze capture data and the physiological data. For the gaze period of each spatial element obtained, the average power of the specific frequency band of all brain waves in the gaze period, the baseline level of GSR, the mean of GSR, and the time domain and frequency domain parameters of HRV are obtained as the corresponding physiological data.

[0009] Furthermore, before constructing the comfort simulation prediction model, the input objective perception indicators and subjective perception indicators are standardized, and all input data are converted into the same scale to ensure that the numerical ranges of indicators of different dimensions are consistent.

[0010] Furthermore, the MLP model is used to construct the comfort prediction model, and its model structure is: in, is the predicted comfort score, is the activation function, are the standardized objective perception indicators and subjective perception indicators, is the weight, is the bias term; In the MLP model, each layer of the network calculates the output through weighted input, and a nonlinear activation function is applied between each layer. The score of the output layer is the final comfort score prediction value.

[0011] Furthermore, the subjective ratings of the participants on the street view panorama were used as target values to train the prediction model, and the weights were optimized using the gradient descent algorithm. and bias , so that the error between the comfort score of the prediction model and the subjective score of the participants is minimized; specifically, In the process of simulating the prediction model training, the weight parameters of the prediction model are adjusted by minimizing the loss function. Defined as: in, is the comfort score predicted by the model, The subjective ratings of the participants. is the sample size; Through the gradient descent algorithm, the comfort simulation prediction model will gradually optimize the weights and bias terms so that the error between the predicted comfort score and the subjective score reaches the historical minimum difference.

[0012] Furthermore, after the comfort simulation prediction model is trained, the objective perception index and subjective perception index of the new observation point are standardized to form input variables, which are input into the model for prediction. The output prediction result is the comfort score: in, and are the final weights and biases obtained through training.

[0013] Furthermore, the comfort distribution in the study area is generated by the spatial interpolation algorithm, specifically, Obtaining the spatial positions and comfort scores of the plurality of observation points in the selected high-intensity area; The comfort score and spatial location of each observation point are used as known data points and input into the selected interpolation algorithm to calculate the comfort score of each selected location in the entire area; Through spatial interpolation, a continuous comfort score map is obtained, which shows the spatial distribution of comfort in the entire high-intensity area.

[0014] 3) Beneficial effects: Compared with the prior art, this invention has the following beneficial effects: The present invention calculates the proportion of participants' gaze duration on each identified spatial element, screens out the most relevant spatial elements, and combines the participants' physiological data such as brain waves, GSR and HRV of relevant spatial elements to ensure that the model can effectively capture the impact of spatial elements on comfort and reduce dependence on irrelevant factors.

[0015] By combining objective and subjective perception indicators, the comfort simulation prediction model can comprehensively reflect the environmental characteristics of high-intensity areas and the physiological responses of participants. By establishing a mapping relationship between objective and subjective perception indicators, it can capture the relationship between the environmental characteristics of the high-intensity area and individual subjective perceptions, avoiding the one-sidedness caused by a single indicator or single perception dimension, making the prediction of comfort more comprehensive and accurate. By inputting the perception indicators of the observation points into the model for prediction and combining it with a spatial interpolation algorithm to generate a comfort distribution map, the comfort distribution in the high-intensity area can be intuitively presented. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flowchart of a method for simulating and predicting comfort in high-intensity areas of an urban area provided by an embodiment of the present invention; Figure 2 A flowchart of a method for training a prediction model in a method for simulating and predicting comfort in high-intensity areas of an urban area provided by an embodiment of the present invention; Figure 3 A flowchart of a method for generating the spatial distribution of comfort in a high-intensity area in a method for simulating and predicting comfort in a high-intensity area of an city provided by an embodiment of the present invention; Figure 4 A schematic diagram of capturing the sight lines of participants viewing a street view panorama in a method for simulating and predicting comfort in a high-intensity urban area provided by an embodiment of the present invention; Figure 5 A schematic diagram of obtaining physiological data while capturing the participant's line of sight in a method for simulating and predicting comfort in a high-intensity urban area provided by an embodiment of the present invention.

[0017] Figure 6 This is a schematic diagram of the heart rate variability amplitude of participants measured in a method for simulating and predicting comfort in a high-intensity urban area provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] 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.

[0019] 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.

[0020] 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.

[0021] 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.

[0022] While developing a comfort evaluation index for high-intensity urban areas, the inventors discovered that comfort in these areas is influenced not only by objective environmental factors but also by individual subjective perceptions. Further research revealed that comfort assessments are not simply a response to visual factors but also involve more complex psychological and physiological responses.

[0023] Therefore, in order to solve the problem of inaccuracy of single visual assessment in current technical methods, Figures 1 to 5As shown, an embodiment of the present invention provides a method for simulating and predicting comfort in high-intensity urban areas. This method integrates objective environmental indicators and individual subjective perceptions into a multi-dimensional assessment system. This method not only overcomes the subjectivity inherent in traditional assessments but also integrates physiological data. For example, modern physiological sensing technology can capture the psychological changes of residents in different environments, such as brain waves, galvanic skin response (GSR), and heart rate variability (HRV). This physiological data not only provides a deeper understanding of people's responses to high-intensity environments but also provides more accurate subjective perception data for comfort assessment. For example, a higher GSR value can reflect an individual's anxiety or stress, while HRV can reflect an individual's emotional state and level of relaxation. Combining these physiological data with objective environmental indicators forms a more comprehensive comfort assessment framework. This allows comfort assessment in high-intensity areas to combine environmental characteristics with individual subjective perceptions, providing a more reliable basis for predicting comfort in high-intensity urban areas.

[0024] Specifically, refer to Figure 1 First, data on objective environmental factors is obtained, and then S10 is performed: multiple observation points are obtained in the selected high-intensity area, street view panoramas of each observation point are obtained, and spatial elements of each street view panorama are extracted using image semantic recognition technology. In this step, multiple typical observation points within the study area need to be selected. These observation points are usually selected based on the geographical characteristics, urban structure, population activity density, and areas representative of comfort in the high-intensity area. By selecting these observation points with typical environmental characteristics of the high-intensity area, the environmental characteristics of the entire high-intensity area can be comprehensively reflected.

[0025] Regarding obtaining street view panoramas of each observation point, in some embodiments, street view map API tools, such as Google Street View or Baidu Street View, can be used to obtain street view panoramas of these observation points. Among them, street view panoramas generally provide 360-degree panoramic views, which can display comprehensive visual information of the location, such as whether there are spatial elements such as roads, buildings, green spaces, and streets, and the locations of these spatial elements. In addition, these street view panoramas are generally converted into two-dimensional image data of a uniform scale, thereby providing comparative information for subsequent environmental analysis, that is, semantic recognition of the image.

[0026] In summary, it can be understood that image semantic recognition extracts different spatial element features from street view panoramic images by applying deep learning and computer vision technologies. In some embodiments of the present invention, deep learning algorithms such as convolutional neural networks (CNNs) are used to train a CNN classification model based on semantic segmentation, enabling it to automatically extract different spatial elements from street view panoramic images, such as buildings, sky, roads, streetlights, traffic paths, plants, motor vehicles, and non-motor vehicles.

[0027] For example, deep learning algorithms, including but not limited to DeepLabV3+ or Mask R-CNN, are used to perform pixel-level semantic segmentation on street view panoramic images, identify the element features of different areas in the image, and classify the spatial elements. The output obtained is the category of each spatial element in the street view panoramic image and their position in the image. For example, the spatial elements mentioned above, such as buildings, sky, roads, street lights, traffic roads, plants, motor vehicles and non-motor vehicles, each identified element will be individually identified and marked with its corresponding regional position. On this basis, by further calculating the proportion of these spatial elements in the image, spatial indicators related to environmental comfort, such as greening rate, sky openness and line of sight accessibility, can be obtained.

[0028] Assume that a commercial center area is selected as the research object of the high-intensity area. First, multiple observation points are determined, such as the intersection of commercial streets, the square in front of major shopping malls, and the side of street parks. These points are typical public spaces that can represent the environmental characteristics of the area. Through street view map APIs such as Google Street View, 360-degree street view panoramic images of these observation points are obtained. Among them, the intersection of commercial streets provides rich information on roads, buildings, green belts and streets, and the street view next to street parks provides more information on green elements and open spaces.

[0029] Deep learning algorithms are used to perform semantic recognition on these street view panoramic images. Specifically, a pre-trained Mask R-CNN model is used to identify and segment pre-trained spatial elements within the street view panoramic images. For example, for an image of a commercial street intersection, the Mask R-CNN model is pre-trained to segment and identify spatial elements such as main roads, side roads, sidewalks, buildings, motor vehicles, non-motor vehicles, pedestrians, street lights, and greenery.

[0030] In summary, using image semantic recognition technology to automatically extract spatial elements from street view panoramas not only significantly improves data acquisition efficiency but also enhances the accuracy of analysis results through the precise recognition of deep learning algorithms. Comprehensive analysis of street view images from different observation points in high-intensity areas accurately reflects the environmental characteristics of the area, providing a reliable data foundation for comfort simulation and prediction models.

[0031] After obtaining the street view panorama at the observation point and identifying its spatial elements, S20 is performed: gaze capture technology is used to record the gaze trajectory of the participant viewing each street view panorama. Combined with the spatial elements of gaze, the corresponding physiological data of brain waves, GSR, and HRV are obtained. The purpose of this step is to obtain different individuals' subjective perception data of the areas where the aforementioned observation points are located. Gaze capture technology is used to track the gaze trajectory of the participant while viewing each street view panorama, and combined with the spatial elements of the participant's gaze, their physiological responses are recorded and analyzed. Gaze capture technology typically relies on eye tracking equipment, which captures the direction and point of an individual's gaze by accurately measuring the trajectory of eye movements.

[0032] In some embodiments, eye tracking technology can track the participant's eye movements through eye sensors on a participant's head-mounted device, such as an infrared camera. Eye tracking equipment can accurately record the movement and focus position of the pupil and combine it with various gaze parameters to create an eye gaze trajectory. Common parameters include gaze point location, gaze duration, scan path, and eye movement velocity. Using these parameters, the participant's gaze trajectory can be obtained while viewing the street view panorama.

[0033] Combined with the above-mentioned eye tracking device to obtain the eye trajectory of the participant when watching the street view panorama, it can be understood that when the participant watches the street view panorama, the eye tracking device records the spatial elements of his gaze in real time. These spatial elements are extracted by the above-mentioned image semantic recognition technology. By corresponding the participant's eye trajectory with these spatial elements, we can gain a deeper understanding of the participant's attention to different environmental features. Figure 4 The yellow areas in the figure represent the areas where a participant gazed for a prolonged period of time while viewing the street view panorama at each observation point. It can be understood that if a participant gazes at a spatial element for a prolonged period of time and the collected physiological data also fluctuates, then this spatial element can be determined to have a significant impact on comfort.

[0034] Regarding the collection of physiological data, in some embodiments, EEG technology uses electrodes to monitor brainwave activity, extracting EEG signals of different frequency bands, such as alpha waves (8-13Hz), beta waves (13-30Hz), and theta waves (4-8Hz). These frequencies are closely related to an individual's psychological state. For example, alpha waves are typically associated with states of relaxation and concentration, while beta waves are associated with high concentration and anxiety. By analyzing changes in participants' EEG while viewing street view panoramas, the impact of the environment on an individual's psychological state can be assessed.

[0035] In addition, you can refer to Figure 5 This figure shows that while capturing the participant's gaze, other physiological data is obtained by measuring skin conductivity (GSR), body temperature, heart rate, pulse amplitude, and respiration. Among them, galvanic skin response (GSR) can reflect an individual's physiological response by measuring changes in the skin's electrical conductivity. Psychological states such as emotions and stress can cause sweat gland activity and change skin conductivity, so GSR is an effective tool for monitoring emotional responses. In this work, GSR can be used to measure participants' emotional fluctuations when viewing street scenes, such as tension, comfort, or relaxation.

[0036] The heart rate variability (HRV) of the participants is obtained by measuring the heart rate. HRV reflects the activity of the heart's autonomic nervous system. By analyzing the frequency and time domain indicators of heart rate changes, the physiological stress response and relaxation level of the participants can be revealed. Figure 6 This graph shows a measurement of a participant's heart rate variability (HRV). HRV values are obtained by measuring the RR pulse interval. Lower HRV values are generally associated with higher stress levels or discomfort, while higher HRV indicates a more relaxed state.

[0037] Combining the above, gaze tracking technology and physiological data collection can provide important subjective physiological data for comfort analysis in high-intensity areas. By combining eye tracking technology with physiological indicators such as EEG, GSR, and HRV, we can more comprehensively assess the impact of environmental characteristics on individual physiological and psychological states, providing strong data support for subsequent comfort simulation and prediction models.

[0038] After obtaining the eye movement data and physiological data of the participants, S30 is performed: based on the fixation duration of each spatial element, its fixation proportion in the street view panorama is calculated, and a specific number of spatial elements with the highest fixation duration proportions are selected as relevant spatial elements, and physiological data of the relevant spatial elements is obtained. And S40: the relevant spatial elements of each street view panorama obtained are used as objective perception indicators, and the corresponding physiological data are used as subjective perception indicators.

[0039] Regarding the acquisition of gaze proportion, specifically, in some embodiments, the eye movement trajectory of the participants is tracked by gaze capture technology, such as the above-mentioned eye tracker, and their gaze points are recorded. Whenever a participant gazes at a certain spatial element, the system will generate a gaze start time and end time, which is usually accurate to the millisecond level, and it is the same as the millisecond unit of the heart rate. For example, if a participant gazes at a building in a street view panorama for 550 milliseconds, the system will record the gaze start time and end time, and mark this spatial element as "Building A". In this way, the gaze duration of each participant on each spatial element is accurately marked as a time period.

[0040] By synchronizing timestamps, the gaze capture data and physiological data are aligned. It is important to ensure that the physiological signals at each moment, such as brain waves, GSR, and HRV data, match the participant's gaze activity.

[0041] For example, during the EEG acquisition process, suppose a time series of data is obtained with timestamps from T1 to T2. At the same time, during this time period, the participant is looking at the spatial element "Building A." Therefore, the EEG data during this time period is associated with the gaze data for "Building A." This can be understood as retrieving all data in the physiological data acquisition system for the gaze period of each spatial element. The EEG data, GSR data, and HRV data in this time period are labeled as data for this spatial element. Direct matching can then be performed using timestamps to ensure that each data point belongs to a specific spatial element.

[0042] Finally, each relevant spatial element and the corresponding physiological data are combined together to prepare for subsequent analysis. In some embodiments, after the data collection is completed, a data table is generated for subsequent integration of the physiological data of multiple participating individuals. The columns of the table include:

[0043] Related spatial elements: spatial element name; Timestamp: the start and end time of the time period for looking at the relevant spatial feature; Fixation duration: the length of time the participant fixated on the relevant spatial element; Physiological data: physiological data recorded during this time period, namely, the average power of the EEG frequency band, the baseline and mean of GSR, and HRV time-domain and frequency-domain parameters; Gaze ratio: The ratio of the gaze time of the relevant spatial element to the total gaze time.

[0044] Finally, step S50 is performed: based on the objective perception indicators and the subjective perception indicators, a comfort simulation prediction model is constructed, and a mapping relationship between the objective perception indicators, the subjective perception indicators and the comfort scores is established; wherein the objective perception indicators and the subjective perception indicators are used as model input variables, and the subjective comfort scores of the participating individuals for the street view panorama are used as target values to establish a mapping relationship between the input indicators and the comfort scores.

[0045] It's important to note that before building a comfort simulation prediction model, the input objective and subjective perception metrics must be standardized. This ensures that metrics of different dimensions are calculated on the same scale to prevent certain large-scale metrics from dominating the model training process. In some embodiments, all input data is converted to datasets with zero mean and unit variance to ensure that each input feature contributes to the same range of model training.

[0046] After standardizing the data, an MLP model is used to construct a comfort prediction model. MLP is a feedforward neural network with at least one hidden layer, where each layer calculates the output through weighted input and a nonlinear activation function is applied between each layer. The structure of the model can be expressed as follows: in, is the predicted comfort score, which is the output of the model, The activation function is a nonlinear function. Common activation functions include ReLU, Sigmoid, and Tanh. The activation function adds nonlinearity to the output of each neuron in the neural network, increasing the model's expressiveness and enabling the activation function of each layer to learn complex nonlinear mapping relationships.

[0047] is the overall characteristics of the standardized objective perception indicators and subjective perception indicators, is the weight, It is the sum of weighted inputs, that is, for each input feature, there will be a corresponding weight for the objective perception index and the subjective perception index , the weight represents the influence of the feature on the prediction result, and the whole can be expressed as the weighted sum of all input features: in, Indicates the input features, namely the standardized objective perception indicators or subjective perception indicators. It is The weight of the input features, weight Also represents the input feature The contribution to comfort prediction is important to note, which will be optimized during the training process.

[0048] The bias term is an additional parameter in each layer of the neural network, which helps the model fit the data better. When the weighted sum of is zero, the bias term allows the output of a non-zero result. The specific data of is learned through training and no value limitation is made here.

[0049] Overall, It is the result after the activation function is applied, which represents the output of the neuron.

[0050] Assume that the input of a layer is , which is linearly related, the activation function This linear combination can be transformed into a nonlinear result. Nonlinear activation functions can help the model learn complex mapping relationships. For example, the ReLU function transmits the original value when the input feature value is greater than 0, and outputs 0 when the input feature value is less than 0. This helps the neural network learn the nonlinear relationship between different features.

[0051] During the training of the MLP model, the weights of the network need to be adjusted and bias To minimize the gap between the predicted value and the actual value. To this end, in some embodiments, a loss function can be used to measure the error between the predicted value and the actual value. The loss function is often represented by the mean square error (MSE), and its definition formula is:

[0052] in, is the comfort score predicted by the model, that is, the comfort score calculated by the MLP model based on the input features. Rate the actual comfort level, is the number of samples. During the training process, the loss function calculates the average error of all samples, so is the total number of samples involved in training.

[0053] The goal of training a model is to adjust the model parameters, i.e. weights and bias , so that the loss function value As small as possible, to achieve this goal, a gradient descent algorithm is used in some embodiments.

[0054] First, input data through the current weights and bias terms Calculate predicted values , and then calculate the loss function . Loss function The gradient of the loss function with respect to the weights and bias terms is the derivative, also known as the partial derivative. The gradient indicates the sensitivity of the loss function to parameter changes. Specifically, we need to calculate each parameter, that is, the weight and bias Contribution to the loss function, that is, to obtain the gradient of the loss function with respect to the parameters:

[0055] The gradient descent algorithm is used to update the weights and bias terms based on the calculated gradient. The gradient descent algorithm adjusts the parameters according to the gradient value and gradually reduces the loss function value, so that the predicted value is as close to the actual value as possible.

[0056] After training is complete, a set of final weights and biases are obtained. These are the optimal parameters obtained through continuous optimization during the training process. For new observation points, that is, the objective and subjective perception data of new street view panoramas and their spatial elements, these data can be input into the trained MLP model for prediction and output of the corresponding comfort score:

[0057] in, is the final weight obtained through training, is the final bias term obtained through training.

[0058] The above process demonstrates how the trained MLP model is used to predict the comfort level for a new street view panorama and its spatial features. This process includes standardizing the input data, calculating a weighted sum, generating a prediction using an activation function, and outputting a new comfort score. The trained weights and biases help the model map different perceptual metrics to a final comfort score, thus providing effective comfort predictions for new observation points.

[0059] In summary, it can be understood that in order to simulate and predict the sense of comfort in high-intensity urban areas, the overall method steps are as follows. First, multiple observation points are selected in the selected high-intensity urban area to obtain a street view panorama of each observation point. A street view panorama is a 360-degree image obtained from a specific observation point that can truly reflect the visual characteristics of the street environment. Through image semantic recognition technology, the spatial elements in each street view panorama are extracted, among which different spatial elements have different effects on the emotions and comfort of the participants. Using eye gaze capture technology, the gaze trajectory of the participants when viewing each street view panorama is recorded, and the gaze duration of each spatial element is calculated. At the same time, the physiological data of the participants are collected. The physiological data reflects the psychological and physiological states of the participants when viewing different spatial elements, such as relaxation, tension or pleasure.

[0060] Because street view panoramas of high-intensity areas contain a variety of spatial elements, not all of which significantly influence participants' emotions and comfort. By analyzing the fixation duration of each spatial element, we can identify the elements with the greatest impact as relevant spatial elements. Specifically, based on the fixation duration of each spatial element, its percentage of fixations in the street view panorama is calculated. The spatial elements with the highest fixation duration percentage are then selected as relevant spatial elements. These spatial elements, due to their longer fixation duration, are more likely to have a significant impact on participants' emotions and comfort. The reason for selecting fixation duration as relevant spatial elements is understandable. This is because fixation duration is often correlated with a person's interest in an object or the degree of emotional impact it has on them. For example, participants may focus more on a green space than a tall building because green space can be relaxing. Furthermore, identifying relevant spatial elements avoids processing redundant information, highlights the spatial elements most important for comfort prediction, and improves the efficiency and accuracy of the model.

[0061] Based on the collected spatial elements and corresponding physiological data, objective and subjective perception indicators are constructed. The objective perception indicator is composed of the attribute values of the relevant spatial elements, while the subjective perception indicator is composed of the physiological data of the participants when viewing the relevant spatial elements. The objective and subjective perception indicators describe comfort from the perspectives of environmental characteristics and human physiological responses, respectively.

[0062] The MLP model is used as a prediction tool, taking objective and subjective perception indicators as input variables and the participants' subjective comfort ratings of the street view panoramas as the target value. MLP is a nonlinear model that can handle complex nonlinear relationships between input features and target values. For example, the impact of greening ratio on comfort may not be a simple linear relationship, but rather exhibit a threshold effect. This threshold effect can be understood as requiring that the greenery coverage of an area exceeds a certain limit before participants in that area experience comfort.

[0063] By establishing a mapping relationship between input indicators and comfort scores during the training process, the above can be briefly described as follows: taking the objective perception indicators and subjective perception indicators as input variables, and the participants' comfort scores for street view panoramas as target values, a mapping relationship between the input variables and the comfort scores is constructed, and the model parameters are adjusted through the backpropagation algorithm to make the predicted values as close to the true values as possible.

[0064] The trained MLP model can input the perceptual indicators of new observation points and predict their corresponding comfort scores. Because the number of observation points in a target high-intensity area is limited and cannot fully cover the entire study area, some embodiments employ spatial interpolation methods, such as Kriging interpolation or IDW interpolation. This method uses common spatial interpolation methods, such as Kriging interpolation or IDW interpolation, to infer the scores of other locations in the study area based on the comfort scores of known observation points. This can be understood as generating a comfort distribution map for the entire study area through a spatial interpolation algorithm, providing a reference for urban planning.

[0065] Finally, S60 is performed: the perception indicators of each observation point are input into the model to predict the corresponding comfort score, and a spatial interpolation algorithm is used to generate a comfort distribution map within the study area. The spatial interpolation algorithm is used to estimate and generate the comfort distribution of other areas based on the known comfort score data points.

[0066] In some embodiments of the present invention, the comfort scores of known observation points are used to extrapolate these values to unknown points using the inverse distance weighting method. The formula for the inverse distance weighting method is:

[0067] in: The point to be predicted The comfort score, It is an observation point The comfort score, It is an observation point To the predicted point distance, It is the distance attenuation parameter, which controls the influence of distance on the prediction results.

[0068] After predicting the comfort scores for multiple observation points, a spatial interpolation algorithm is used to predict the comfort scores for each location within the study area. A comfort distribution map for the entire area is then generated based on these predicted scores. This map can be displayed as an image or map, using color coding to indicate the comfort levels of different areas. Assume that 10 observation points are selected within the study area, and both objective and subjective perception indicators are recorded for each point. After standardization, this data is input into the trained MLP model to predict the comfort score for each observation point.

[0069] Use the inverse distance weighted method for interpolation. Assume that for a point , which is surrounded by multiple observation points ( , ,…, ), each observation point has a corresponding comfort score The predicted comfort value of this point is calculated by the formula .

[0070] 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 method for simulating and predicting the comfort level in high-intensity urban areas, characterized by: include: Obtaining street view panoramas of multiple observation points in the selected high-intensity area, and extracting spatial elements of each of the street view panoramas using image semantic recognition technology; The gaze capture technology is used to record the gaze trajectory of the participants when viewing each of the street view panoramas, and the gaze duration of each of the spatial elements is obtained, and the physiological data of the participants are obtained at the same time; the physiological data include the average power of a specific frequency band of the brain wave, the baseline level of the GSR, the mean value of the GSR, and the time domain and frequency domain parameters of the HRV, wherein the specific frequency band includes Wave, wave and Frequency band of the wave; Calculating the fixation ratio of each spatial element in the street view panorama according to the fixation duration of each spatial element, selecting a preset number of spatial elements with the highest fixation duration ratio as relevant spatial elements, and obtaining physiological data of the relevant spatial elements; The relevant spatial elements of each acquired street view panorama are used as objective perception indicators, and the corresponding physiological data are used as subjective perception indicators; Constructing a comfort simulation prediction model and establishing a mapping relationship between the objective perception index and the subjective perception index and the comfort score; wherein the objective perception index and the subjective perception index are used as input variables of the prediction model, and the prediction model is trained using the participants' subjective ratings of the street view panorama as target values to predict the comfort score of the input variables; The input variables of each observation point are obtained and input into the model to predict the corresponding comfort score, and a comfort distribution map within the study area is generated through a spatial interpolation algorithm.

2. The method for simulating and predicting the sense of comfort in urban high-intensity areas according to claim 1 is characterized in that: Gaze capture technology is used to track the eye movements of participants and obtain the gaze duration of each spatial element. Specifically, by recording the gaze points of participants when viewing the street view panorama, each time the participant gazes at a spatial element, a gaze start time and end time are generated based on timing information, and the gaze start time and end time are used as the gaze period.

3. The method for simulating and predicting the sense of comfort in urban high-intensity areas according to claim 2 is characterized in that: Synchronize gaze capture and obtain the timestamp of the physiological numerical control to align the gaze capture data and the physiological data. For the gaze period of each spatial element obtained, obtain the average power of the specific frequency band of all brain waves in the gaze period, the baseline level of GSR, the mean value of GSR, and the time domain and frequency domain parameters of HRV as its corresponding physiological data.

4. The method for simulating and predicting the sense of comfort in high-intensity urban areas according to claim 1, characterized in that: Before constructing the comfort simulation prediction model, the input objective perception indicators and subjective perception indicators are standardized, and all input data are converted to the same scale to ensure that the numerical ranges of indicators in different dimensions are consistent.

5. The method for simulating and predicting the sense of comfort in urban high-intensity areas according to claim 4 is characterized in that: The MLP model is used to construct the comfort prediction model, and its model structure is: in, is the predicted comfort score, is the activation function, are the standardized objective perception indicators and subjective perception indicators, is the weight, is the bias term; In the MLP model, each layer of the network calculates the output through weighted input, and a nonlinear activation function is applied between each layer. The score of the output layer is the final comfort score prediction value.

6. A method for simulating and predicting the sense of comfort in high-intensity urban areas according to claim 5, characterized in that: The prediction model is trained using the participants' subjective ratings of comfort on the street view panoramas as target values, and the weights are optimized using the gradient descent algorithm. and bias , so that the error between the comfort score of the prediction model and the subjective score of the participants is minimized; specifically, In the process of simulating the prediction model training, the weight parameters of the prediction model are adjusted by minimizing the loss function. Defined as: in, is the comfort score predicted by the model, The subjective ratings of the participants. is the sample size; Through the gradient descent algorithm, the comfort simulation prediction model will gradually optimize the weights and bias terms so that the error between the predicted comfort score and the subjective score reaches the historical minimum difference.

7. A method for simulating and predicting the sense of comfort in urban high-intensity areas according to claim 6, characterized in that: After the comfort simulation prediction model is trained, the objective perception index and subjective perception index of the new observation point are standardized to form input variables, which are input into the model for prediction. The output prediction result is the comfort score: in, and are the final weights and biases obtained through training.

8. The method for simulating and predicting the sense of comfort in high-intensity urban areas according to claim 1, characterized in that: The comfort distribution in the study area is generated by the spatial interpolation algorithm, specifically, Obtaining the spatial positions and comfort scores of the plurality of observation points in the selected high-intensity area; The comfort score and spatial location of each observation point are used as known data points and input into the selected interpolation algorithm to calculate the comfort score of each selected location in the entire area; Through spatial interpolation, a continuous comfort score map is obtained, which shows the spatial distribution of comfort in the entire high-intensity area.

Images