Walking space restorative environment factor optimization method based on VR and deep learning

Abstract

The embodiment of the invention provides a VR and deep learning-based walking space restoration environment factor optimization method. Relates to the visual perception field. The method comprises the following steps: screening visual restorative environment factors of a walking space by using an eye tracker experiment and a semantic difference questionnaire; establishing VR walking spaces corresponding to different combination visual recovery environment factors according to an orthogonal experiment method, carrying out a VR orthogonal experiment, and recording the recovery state change of the testee in different VR walking spaces; training a prediction model according to different combined visual recovery environment factors and the state changes of the testees in different VR walking spaces; and inputting each visual restorative environment factor in the real walking space into the prediction model, obtaining the recovery state change of the person in the real walking space, and optimizing each visual restorative environment factor in the real walking space by using a genetic algorithm. Therefore, the influence of the visual recovery environmental factors on attention recovery can be revealed, the recovery of the walking space is predicted, the environmental factors are optimized, and the psychological health of residents is improved.

Description

Technical Field

[0001] The present disclosure relates to the fields of visual perception technology and restorative environment, and in particular to a method for optimizing restorative environmental factors of a walking space based on VR and deep learning. Background Art

[0002] The fast-paced lifestyle and highly competitive work environment of modern society have brought people problems such as long working hours, information overload, endless interference from social media, and high expectations for success. People need to deal with a large amount of information and tasks in their daily lives, which often leads to people feeling stressed and distracted, resulting in psychological fatigue and mental burnout. Studies have shown that restorative environments can effectively help individuals restore their depleted attention, relieve psychological stress, and bring about a series of positive changes in physiology and psychology. Pedestrian space is an important part of modern cities, and its spatial vitality and attractiveness have a significant impact on the health of residents. Improving environmental characteristics and visual conditions can reduce the stress level of users, while exposure to tense visual environments may reduce efficiency and concentration. A large number of studies have shown that people's attention recovery is greatly affected by environmental visual perception, and the quality of visual perception is positively correlated with recovery benefits. Therefore, it is necessary to help improve residents' attention recovery and overall health by optimizing the visual restorability of pedestrian spaces.

[0003] In terms of research methods, the existing research methods are mainly interviews and questionnaires. The subjective data obtained by these methods often cannot truly reflect the effective relationship between visual perception and spatial elements. There are great limitations in studying the impact of different characteristic parameters on the visual environment in a real physical environment. At the same time, VR technology has provided accurate and effective data support for environmental vision research in recent years, but the research on spatial elements is mainly concentrated in the basic recognition and extraction stages.

[0004] In addition, in the field of public health research, some relatively mature conclusions have been drawn on the restorativeness of urban walking spaces, but the classification of spatial features is too general, it cannot deeply understand the contribution of specific physical dimensions (such as fences and orientation), and lacks and neglects to consider the impact of walking activities on environmental perception. At the same time, there is also the problem of small sample size, and it has not yet reached a comprehensive understanding of the impact mechanism of the restorative visual environment elements of walking spaces on attention restoration. Summary of the invention

[0005] The present disclosure provides a method, device, equipment and storage medium for optimizing restorative environmental factors of walking space based on VR and deep learning.

[0006] According to a first aspect of the present disclosure, a method for optimizing restorative environmental factors of a walking space based on VR and deep learning is provided. The method comprises:

[0007] The visual restorative environmental factors of the walking space were screened out using an eye tracker experiment and a semantic differential questionnaire; wherein the visual restorative environmental factors included building height, tree height, shrub area in the field of vision, ground color and ground texture;

[0008] According to the orthogonal experimental method, different combinations of visual restorative environmental factors are carried out to establish VR walking spaces corresponding to different combinations of visual restorative environmental factors.

[0009] VR orthogonal experiments were conducted based on VR walking spaces corresponding to different combinations of visual restorative environmental factors, and the changes in participants' recovery status in different VR walking spaces were recorded;

[0010] The prediction model is trained based on different combinations of visual restorative environmental factors and the changes in the recovery status of the participants in the VR walking space corresponding to the different combinations of visual restorative environmental factors;

[0011] The visual restorative environmental factors of the real walking space are input into the trained prediction model to obtain the change value of the recovery state of the personnel in the real walking space; according to the change value of the recovery state of the personnel in the real walking space, the genetic algorithm is used to optimize the visual restorative environmental factors of the real walking space.

[0012] In some implementations of the first aspect, the VR orthogonal experiment includes:

[0013] Through the pre-experiment, the participants were subjected to a stress experiment and their first recovery state was evaluated using the recovery state quantification table;

[0014] After the participants wore the VR device and walked freely in the VR walking space corresponding to different combinations of visual restorative environmental factors for a preset period of time, the recovery state quantification table was used to evaluate the participants' second recovery state corresponding to different VR walking spaces;

[0015] The difference between the first recovery state and the second recovery state corresponding to the participant after walking recovery in different VR walking spaces is used as the recovery state difference corresponding to different VR walking spaces.

[0016] In some implementations of the first aspect, training a prediction model based on different combinations of visual restorative environmental factors and changes in recovery states of participants in VR walking spaces corresponding to the different combinations of visual restorative environmental factors includes:

[0017] Different combinations of visual restorative environmental factors are used as training sample sets, and the changes in the recovery status of participants in the VR walking space corresponding to different combinations of visual restorative environmental factors are used as training output sets;

[0018] According to the training sample set and the training output set, a convolutional neural network is used to learn the relationship between each visual restorative environmental factor and the change in the recovery state of the participant and to train a prediction model for predicting the restorability of walking space.

[0019] In some implementations of the first aspect, using a convolutional neural network to learn the relationship between each visual restorative environmental factor and the change in the recovery state of the participant includes:

[0020] The correlation between each visual restorative environmental factor and the changes in the recovery status of the participants was obtained by Spearman correlation analysis method, and the convolutional neural network was used to learn the relationship between each visual restorative environmental factor and the changes in the recovery status of the participants.

[0021] In some implementations of the first aspect, according to the change value of the recovery state of the person in the real walking space, the visual restorative environmental factors of the real walking space are optimized using a genetic algorithm, including:

[0022] The change value of the recovery state of the person in the real walking space is used as the fitness index of the genetic algorithm, and the extreme values ​​of each visual restorative environmental factor in different combinations are adjusted to obtain each visual restorative environmental factor in the combination that meets the preset fitness;

[0023] According to each visual restorative environmental factor in the combination that satisfies the preset fitness, the parameter range of each visual restorative environmental factor in the real walking space is optimized.

[0024] In some implementations of the first aspect, the visual restorative environmental factors of the walking space are screened out using an eye tracker experiment and a semantic differential questionnaire, including:

[0025] The participants wore an eye tracker and calibrated it, pressed the button of the eye tracker to start recording data, and then walked freely in the preset real walking space for a preset time. The environmental components that the participants paid attention to in the preset real walking space were determined based on the eye movement data of the participants captured by the eye tracker, and a semantic differential questionnaire was designed based on the environmental components. The visual restorative environmental factors of the walking space were screened out through the semantic differential questionnaire.

[0026] In some implementations of the first aspect, the visual restorative environmental factors of the walking space are screened out by a semantic differential questionnaire, including:

[0027] According to the environmental elements that the participants pay attention to in the preset real walking space, neutral factors affecting spatial visual recovery, adjective pairs corresponding to the neutral factors, and a semantic differential table are set, so that the participants can score the adjective pairs on the semantic differential table according to their own recovery status and the neutral factors;

[0028] The scoring results of the participants were collected and analyzed according to the Spearman correlation analysis method to obtain various visual restorative environmental factors.

[0029] In some implementations of the first aspect, before establishing a VR orthogonal experiment, the steps include:

[0030] Establish a VR walking space that is consistent with the preset real walking space, so that participants can walk freely in the VR walking space that is consistent with the preset real walking space for a preset time after wearing the VR device, and record eye movement data based on the VR eye tracker worn;

[0031] By comparing the eye movement data recorded by the VR eye tracker with the eye movement data of the participants captured by the eye tracker, it was determined that the VR orthogonal experiment was feasible.

[0032] According to a second aspect of the present disclosure, a device for optimizing restorative environmental factors of a walking space based on VR and deep learning is provided. The device comprises:

[0033] A visual restorative environmental factor determination module is used to screen out visual restorative environmental factors of walking space using an eye tracker experiment and a semantic differential questionnaire; wherein the visual restorative environmental factors include building height, tree height, shrub area in the field of vision, ground color and ground texture;

[0034] A VR walking space establishment module is used to make different combinations of visual restorative environmental factors according to the orthogonal experimental method to establish VR walking spaces corresponding to different combinations of visual restorative environmental factors;

[0035] The recovery state change recording module is used to conduct VR orthogonal experiments based on VR walking spaces corresponding to different combinations of visual restorative environmental factors, and record the changes in the recovery state of participants in different VR walking spaces;

[0036] A prediction model training module is used to train the prediction model according to different combinations of visual restorative environmental factors and changes in the recovery status of the participants in the VR walking space corresponding to the different combinations of visual restorative environmental factors;

[0037] The visual restorative environmental factor optimization module is used to input various visual restorative environmental factors of the real walking space into the trained prediction model to obtain the change value of the recovery state of the personnel in the real walking space; according to the change value of the recovery state of the personnel in the real walking space, the genetic algorithm is used to optimize the various visual restorative environmental factors of the real walking space.

[0038] According to a third aspect of the present disclosure, an electronic device is provided. The electronic device includes: at least one processor; and a memory connected to the at least one processor in communication; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the method described above.

[0039] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, where the computer instructions are used to cause a computer to execute the method as described above.

[0040] In the present disclosure, the visual restorative environmental factors of the walking space are screened out by using an eye tracker experiment and a semantic differential questionnaire; the orthogonal experimental method is used to make different combinations of the visual restorative environmental factors, and VR walking spaces corresponding to different combinations of visual restorative environmental factors are established; VR orthogonal experiments are conducted according to the VR walking spaces corresponding to different combinations of visual restorative environmental factors, and the changes in the recovery status of the participants in different VR walking spaces are recorded; the prediction model is trained according to the changes in the recovery status of the participants in the VR walking spaces corresponding to different combinations of visual restorative environmental factors; the visual restorative environmental factors of the real walking space are input into the trained prediction model to obtain the change value of the recovery status of the personnel in the real walking space; according to the change value of the recovery status of the personnel in the real walking space, the genetic algorithm is used to optimize the visual restorative environmental factors of the real walking space. In this way, the influence of the visual restorative environmental factors of the walking space on the recovery of attention can be revealed, the restorativeness of the walking space can be predicted by using the visual restorative environmental factors and the convolutional neural network, and the parameter range of the visual restorative environmental factors of the walking space can be optimized by using the genetic optimization algorithm to improve the attention and mental health of urban residents.

[0041] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, among which:

[0043] Figure 1 A flowchart of a method for optimizing restorative environmental factors of a walking space based on VR and deep learning provided by an embodiment of the present disclosure is shown;

[0044] Figure 2A structural diagram of a walking space restorative environmental factor optimization device based on VR and deep learning provided by an embodiment of the present disclosure is shown;

[0045] Figure 3 A block diagram of an exemplary electronic device capable of implementing an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0047] In addition, the term "and / or" in this article is only a description of the association relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0048] In response to the problems in the background technology, the embodiments of the present disclosure provide a method and device for optimizing restorative environmental factors of walking space based on VR and deep learning. Specifically, the visual restorative environmental factors of the walking space are screened out using eye tracker experiments and semantic differential questionnaires; the visual restorative environmental factors are combined in different ways using the orthogonal experimental method to establish VR walking spaces corresponding to different combinations of visual restorative environmental factors; VR orthogonal experiments are conducted based on the VR walking spaces corresponding to different combinations of visual restorative environmental factors, and the changes in the recovery status of participants in different VR walking spaces are recorded; a prediction model is trained based on different combinations of visual restorative environmental factors and the changes in the recovery status of participants in VR walking spaces corresponding to different combinations of visual restorative environmental factors; each visual restorative environmental factor of the real walking space is input into the trained prediction model to obtain the change value of the recovery status of the person in the real walking space; and each visual restorative environmental factor of the real walking space is optimized using a genetic algorithm based on the change value of the recovery status of the person in the real walking space. In this way, we can reveal the impact of visual restorative environmental factors in walking spaces on attention recovery, use visual restorative environmental factors to predict the restorativeness of walking spaces and optimize the parameter range of visual restorative environmental factors in walking spaces to improve the attention and mental health of urban residents.

[0049] In the following, in conjunction with the accompanying drawings, the method and device for optimizing the restorative environmental factors of walking space based on VR and deep learning provided by the embodiments of the present disclosure are described in detail through specific embodiments.

[0050] Figure 1 A flowchart of a method for optimizing restorative environmental factors of a walking space based on VR and deep learning provided by an embodiment of the present disclosure is shown. The method 100 includes the following steps:

[0051] S110, using an eye tracker experiment and a semantic differential questionnaire to screen out visual restorative environmental factors of the walking space; wherein the visual restorative environmental factors include building height, tree height, shrub area in the field of vision, ground color tone and ground texture.

[0052] In some embodiments, the visual restorative environmental factors of the walking space are screened out using an eye tracker experiment and a semantic differential questionnaire, including:

[0053] The participants wore an eye tracker and calibrated it, and then pressed the button of the eye tracker to start recording data, and then walked freely in a preset real walking space for a preset time. The environmental components that the participants paid attention to in the preset real walking space were determined based on the eye movement data of the participants captured by the eye tracker, and a semantic differential questionnaire was designed based on the environmental components. The visual restorative environmental factors of the walking space were screened out through the semantic differential questionnaire.

[0054] Furthermore, the real walking space is preset as the walking space of a certain university campus; and the preset duration is 2 minutes.

[0055] In some embodiments, determining the environmental elements that the participant is paying attention to in the preset real walking space based on the participant's eye movement data captured by the eye tracker includes:

[0056] The eye tracker captures the eye movement data of the participants; the eye movement data includes the gaze duration, gaze time ratio, fixation duration ratio, total fixation time and number of visits to each environment area in the space;

[0057] The environmental area of ​​interest to the participant is determined based on the eye movement data, and then the environmental elements that the participant pays attention to in the preset real walking space are determined based on the participant's eye movement data in the area of ​​interest; wherein the environmental elements include the ground, buildings, trees, shrubs, sky, facilities, and pedestrians.

[0058] In some embodiments, the visual restorative environmental factors of the walking space are screened out by a semantic differential questionnaire, including:

[0059] According to the environmental elements that the participants pay attention to in the preset real walking space, neutral factors affecting spatial visual recovery, adjective pairs corresponding to the neutral factors, and a semantic differential table are set, so that the participants can score the adjective pairs on the semantic differential table according to their own recovery status and the neutral factors;

[0060] The scoring results of the participants were collected and analyzed according to the Spearman correlation analysis method to obtain various visual restorative environmental factors.

[0061] In some embodiments, a comparison table of neutral factors and adjective pairs corresponding to the neutral factors is shown in Table 1:

[0062] Table 1 Comparison table of neutral factors and adjective pairs corresponding to neutral factors

[0063]

[0064] In some embodiments, the semantic differential table has five dimensions, namely very bad, poor, no feeling, good, and very good, and the scores corresponding to these five dimensions are -2, -1, 0, 1, and 2, respectively.

[0065] S120, performing different combinations of visual restorative environmental factors according to an orthogonal experimental method, and establishing VR walking spaces corresponding to different combinations of visual restorative environmental factors.

[0066] In some embodiments, based on common urban ground paving, building features, and landscape characteristics, the value ranges of various visual restorative environmental factors are set as follows:

[0067] The height difference HDB between building heights is between 0 and 60mm; the height difference HT between tree heights is 8mm, 12mm, 16mm, 20mm, 24mm, and 28mm; the area difference AS between shrub areas in the field of view is 0%, 3%, 6%, 9%, 12%, and 15%; the ground color tone HG is 0 degree, 72 degrees, 144 degrees, 216 degrees, 288 degrees, and 360 degrees; the ground texture TG is 100mm*100mm, 200mm*100mm, 300mm*300mm, 600mm*300mm, 600mm*600mm, and 1000mm*1000mm.

[0068] S130, conducting a VR orthogonal experiment based on the VR walking spaces corresponding to different combinations of visual restorative environmental factors, and recording changes in the recovery status of the participants in different VR walking spaces.

[0069] In some embodiments, before performing the VR orthogonal experiment, the following steps are performed:

[0070] Establish a VR walking space that is consistent with the preset real walking space, so that participants can walk freely in the VR walking space that is consistent with the preset real walking space for a preset time after wearing the VR device, and record eye movement data based on the VR eye tracker worn;

[0071] Compare the eye movement data recorded by the VR eye tracker with the eye movement data of the participants captured by the eye tracker to confirm that the VR orthogonal experiment is feasible;

[0072] Specifically, the Unity3D software platform was used to establish a VR walking space that was consistent with the preset real walking space, colliders were added to the ground, and a suitable skybox was set up to give the participants a real and vivid walking visual experience; the participants wore HTC VIVE PRO EYE (a device that includes VR device functions and VR eye tracker functions) and walked freely in the VR walking space for 2 minutes to obtain eye movement data; the eye movement data generated by the participants walking in the VR walking space were processed through the eye movement analysis module of the Unity3D software platform, and the participants' gaze characteristics were analyzed. After analysis, it was found that there was almost no difference between the eye movement data generated by the participants walking in the VR walking space and the eye movement data of the participants captured by the eye tracker, which proved the feasibility and reliability of the VR orthogonal experiment. Therefore, the experimental values ​​obtained from the VR orthogonal experiment can be used to train the prediction model, and the prediction model can be applied to the restorative prediction of the real walking space.

[0073] In some embodiments, the VR orthogonal experiment includes:

[0074] Through the pre-experiment, the participants were subjected to a stress experiment and their first recovery state was evaluated using the recovery state quantification table;

[0075] After the participants wore the VR device and walked freely in the VR walking space corresponding to different combinations of visual restorative environmental factors for a preset period of time, the recovery state quantification table was used to evaluate the participants' second recovery state corresponding to different VR walking spaces;

[0076] The difference between the first recovery state and the second recovery state corresponding to the participant after walking recovery in different VR walking spaces is used as the recovery state difference corresponding to different VR walking spaces;

[0077] Furthermore, a pre-experiment was conducted with the stressor developed on the E-Studio platform to consume the participants’ directed attention and reduce their recovery state;

[0078] The recovery state quantification table uses a 5-point system. Participants score their stress states in different VR walking spaces according to the recovery state quantification table. For the same VR walking space, the weighted average of all participants' scoring results is used to obtain the recovery state difference of the VR walking space.

[0079] S140, training a prediction model based on different combinations of visual restorative environmental factors and changes in recovery status of participants in VR walking spaces corresponding to different combinations of visual restorative environmental factors.

[0080] In some embodiments, the prediction model is trained based on different combinations of visual restorative environmental factors and changes in the recovery status of the participants in the VR walking space corresponding to the different combinations of visual restorative environmental factors, including:

[0081] Different combinations of visual restorative environmental factors are used as training sample sets, and the changes in the recovery status of participants in the VR walking space corresponding to different combinations of visual restorative environmental factors are used as training output sets;

[0082] According to the training sample set and the training output set, a convolutional neural network is used to learn the relationship between each visual restorative environmental factor and the change in the recovery state of the participant and to train a prediction model for predicting the restorability of walking space.

[0083] In some embodiments, using a convolutional neural network to learn the relationship between each visual restorative environmental factor and the change in the recovery state of the participant includes:

[0084] The correlation between each visual restorative environmental factor and the change of the recovery status of the participants was analyzed by the Spearman correlation analysis method, and the relationship between each visual restorative environmental factor and the change of the recovery status of the participants was learned by using a convolutional neural network;

[0085] Furthermore, the correlation between each visual restorative environmental factor and the change in the recovery state of the participants obtained by the Spearman correlation analysis method is as follows:

[0086] The correlation of height difference HDB between building heights is 0.35, the correlation of height difference HT between tree heights is 0.38, the correlation of area difference AS between shrub areas in the field of view is 0.36, the correlation of ground hue HG is 0.23, and the correlation of ground texture TG is 0.43;

[0087] A correlation greater than 0 and less than 0.2 indicates poor or no correlation, a correlation greater than 0.2 and less than 0.5 indicates fair correlation, a correlation greater than 0.5 and less than 0.7 indicates moderate correlation, a correlation greater than 0.7 and less than 0.9 indicates strong correlation, and a correlation greater than 0.9 and less than 1 indicates perfect correlation;

[0088] Therefore, the correlations between the various visual restorative environmental factors and the changes in the participants' recovery status obtained by the Spearman correlation analysis method are as follows from high to low: ground texture TG, height difference between tree heights HT, area difference between shrub areas in the field of vision AS, height difference between building heights HDB, ground color tone HG;

[0089] The relationship between each visual restorative environmental factor and the change in the recovery state of the participants satisfies the quadratic nonlinear fitting formula, which is as follows:

[0090] Y=ax 2 +bx+c

[0091] Among them, x is the visual restorative environmental factor, a, b, c are fitting coefficients, and Y is the change in the recovery state of the participants;

[0092] Specifically, the fitting coefficient a between the height difference HDB between the building heights and the change in the recovery state of the participants is 0.283. Before the critical point, the change value of the recovery state of the participants decreases with the increase of HDB. After the critical point, the change value of the recovery state of the participants increases with the increase of HDB. When HDB is between 3.42 and 15 meters, the change value of the recovery state of the participants is the largest, and the maximum value is 5.

[0093] The fitting coefficient a between the height difference HT between the tree heights and the change in the recovery state of the participants is -0.225. Before the critical point, the change in the recovery state of the participants increases with the increase in HT. After the critical point, the change in the recovery state of the participants decreases with the increase in HT. When HT is between 16 and 20.9 meters, the change in the recovery state of the participants is the largest.

[0094] The fitting coefficient a between the ground color tone HG and the change of the participant's recovery state is -0.16. Before the critical point, the change value of the participant's recovery state increases with the increase of HG. After the critical point, the change value of the participant's recovery state decreases with the increase of HG. When HG is between 136 degrees and 224 degrees, the change value of the participant's recovery state is the largest.

[0095] The fitting coefficient a between the ground texture TG and the participants’ recovery state change is -0.225. Before the critical point, the participants’ recovery state change value increases with the increase of TG. After the critical point, the participants’ recovery state change value decreases with the increase of TG. When TG is at 0.31m 2 ~0.62m 2 When the value of participants' recovery status changes is the largest;

[0096] The fitting coefficient a between the area difference AS between the shrub areas in the field of view and the change in the recovery status of the participants is -0.213. Before the critical point, the change value of the participants' recovery status increases with the increase of AS. After the critical point, the change value of the participants' recovery status decreases with the increase of AS. When AS is between 6.7% and 11.6%, the change value of the participants' recovery status is the largest.

[0097] In some embodiments, the convolutional neural network structure includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer;

[0098] The images of different VR walking spaces were input into the convolutional neural network (the images were RGB three-channel images of 224×224 pixels), so that the convolutional neural network could obtain the visual restorative environmental factors corresponding to different VR walking spaces and fit the relationship between the visual restorative environmental factors and the changes in the recovery status of the participants.

[0099] Specifically, the image input into the convolutional neural network is convolved through a convolutional layer to extract the visual restorative environmental factors corresponding to different VR walking spaces, where the stride of the convolutional layer is 1; the visual restorative environmental factors corresponding to different VR walking spaces are activated using the ReLU activation function and then pooled through a maximum pooling layer. During the convolution and pooling process, the filter size is fixed to 3×3, and the stride of the pooling layer is 3; after multiple convolution and pooling operations, a flattened feature map is obtained; according to the relationship between the information fitting on the feature map and the changes in the recovery state of the participants, the overfitting is reduced through the first fully connected layer with an output dimension of 512 and the Dropout layer with a discarding probability of 0.09, and the nonlinear characteristics are increased using the ReLU activation function; the fitted data stream passes through a second fully connected layer with an output dimension of 256, and the ReLU activation function is used again to increase the nonlinear characteristics, and the final result is output after passing through a third fully connected layer with an output dimension of 1 and the output layer.

[0100] In some embodiments, the prediction model is trained using a decision tree, a K-nearest neighbor algorithm, and an artificial neural network, respectively, to obtain three corresponding prediction models, and the three prediction models are compared with the prediction model obtained by training the prediction model using a convolutional neural network, and the prediction model with the best learning effect is selected as the prediction model for predicting the real walking space recovery state;

[0101] After verifying the above four prediction models, it was found that the prediction accuracy of the prediction model based on decision tree was 78%, the prediction accuracy of the prediction model based on K-nearest neighbor algorithm was 90.5%, the prediction accuracy of the prediction model based on artificial neural network was 94.4%, and the prediction accuracy of the prediction model based on convolutional neural network was 95.2%; therefore, the prediction model based on convolutional neural network was selected as the prediction model for the final application.

[0102] S150, inputting various visual restorative environmental factors of the real walking space into the trained prediction model to obtain the change value of the recovery state of the person in the real walking space; according to the change value of the recovery state of the person in the real walking space, using the genetic algorithm to optimize the various visual restorative properties of the real walking space.

[0103] In some embodiments, according to the change value of the recovery state of the person in the real walking space, the genetic algorithm is used to optimize the visual recovery properties of the real walking space, including:

[0104] The change value of the recovery state of the personnel in the real walking space is used as the fitness index of the genetic algorithm, and the extreme values ​​of each visual restorative environmental factor in different combinations are adjusted to obtain each visual restorative environmental factor in the combination that meets the preset fitness (i.e., the fitness that is most conducive to the recovery state of the personnel in the real walking space);

[0105] According to each visual restorative environmental factor in the combination that satisfies the preset fitness, the parameter range of each visual restorative environmental factor in the real walking space is optimized.

[0106] In some embodiments, the change value of the recovery state of a person who meets the preset fitness in the real walking space is 4. The parameter value range of each visual restorative environmental factor of the real walking space is optimized according to the value of each visual restorative environmental factor corresponding to the change value of the recovery state of the person in the real walking space being 4, that is, the value range of the height difference HDB between the building heights is 20 to 28m, the value range of the height difference HT between the tree heights is 8.5 to 15.5m, the value range of the area difference AS between the shrub areas in the field of vision is 12% to 15%, the value range of the ground hue HG is 140 degrees to 245 degrees, and the value range of the ground texture TG is 0.2m 2 ~0.5m 2 .

[0107] According to the embodiment of the present disclosure, the visual restorative environmental factors of the walking space are screened out by using an eye tracker experiment and a semantic differential questionnaire; different combinations of the visual restorative environmental factors are made according to the orthogonal experimental method to establish VR walking spaces corresponding to different combinations of visual restorative environmental factors; VR orthogonal experiments are conducted according to VR walking spaces corresponding to different combinations of visual restorative environmental factors, and the changes in the recovery status of the participants in different VR walking spaces are recorded; a prediction model is trained according to different combinations of visual restorative environmental factors and the changes in the recovery status of the participants in VR walking spaces corresponding to different combinations of visual restorative environmental factors; each visual restorative environmental factor of the real walking space is input into the trained prediction model to obtain the change value of the recovery status of the person in the real walking space; and each visual restorative environmental factor of the real walking space is optimized according to the change value of the recovery status of the person in the real walking space using a genetic algorithm. In this way, the influence of the visual restorative environmental factors of the walking space on the recovery of attention can be revealed, the restorativeness of the walking space can be predicted by the visual restorative environmental factors, and the parameter range of the visual restorative environmental factors of the walking space can be optimized, so as to improve the attention and mental health of urban residents.

[0108] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present disclosure is not limited by the order of the actions described, because according to the present disclosure, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present disclosure.

[0109] The above is an introduction to the method embodiment. The following is a further explanation of the scheme disclosed in the present invention through an apparatus embodiment.

[0110] Figure 2 The structure diagram of a walking space restorative environmental factor optimization device based on VR and deep learning provided by an embodiment of the present disclosure is shown. The device 200 includes:

[0111] The visual restorative environmental factor determination module 210 is used to screen out the visual restorative environmental factors of the walking space by using an eye tracker experiment and a semantic differential questionnaire; wherein the visual restorative environmental factors include building height, tree height, shrub area in the field of vision, ground color and ground texture.

[0112] In some embodiments, module 210 is specifically configured to:

[0113] Eye tracker experiment and semantic differential questionnaire were used to screen out visual restorative environmental factors of walking space, including:

[0114] The participants wore an eye tracker and calibrated it, pressed the button of the eye tracker to start recording data, and then walked freely in the preset real walking space for a preset time. The environmental components that the participants paid attention to in the preset real walking space were determined based on the eye movement data of the participants captured by the eye tracker, and a semantic differential questionnaire was designed based on the environmental components. The visual restorative environmental factors of the walking space were screened out through the semantic differential questionnaire.

[0115] In some embodiments, module 210 is further configured to:

[0116] The semantic differential questionnaire was used to screen out the visual restorative environmental factors of the walking space, including:

[0117] According to the environmental elements that the participants pay attention to in the preset real walking space, neutral factors affecting spatial visual recovery, adjective pairs corresponding to the neutral factors, and a semantic differential table are set, so that the participants can score the adjective pairs on the semantic differential table according to their own recovery status and the neutral factors;

[0118] The scoring results of the participants were collected and analyzed according to the Spearman correlation analysis method to obtain various visual restorative environmental factors.

[0119] The VR walking space establishment module 220 is used to perform different combinations of visual restorative environmental factors according to the orthogonal experimental method to establish VR walking spaces corresponding to different combinations of visual restorative environmental factors.

[0120] The recovery state change recording module 230 is used to conduct VR orthogonal experiments according to VR walking spaces corresponding to different combinations of visual restorative environmental factors, and record the changes in the recovery state of the participants in different VR walking spaces.

[0121] In some embodiments, module 230 is specifically configured to:

[0122] Before setting up a VR orthogonal experiment, include:

[0123] Establish a VR walking space that is consistent with the preset real walking space, so that participants can walk freely in the VR walking space that is consistent with the preset real walking space for a preset time after wearing the VR device, and record eye movement data based on the VR eye tracker worn;

[0124] By comparing the eye movement data recorded by the VR eye tracker with the eye movement data of the participants captured by the eye tracker, it was determined that the VR orthogonal experiment was feasible.

[0125] In some embodiments, module 230 is further configured to:

[0126] VR orthogonal experiments, including:

[0127] Through the pre-experiment, the participants were subjected to a stress experiment and their first recovery state was evaluated using the recovery state quantification table;

[0128] After the participants wore the VR device and walked freely in the VR walking space corresponding to different combinations of visual restorative environmental factors for a preset period of time, the recovery state quantification table was used to evaluate the participants' second recovery state corresponding to different VR walking spaces;

[0129] The difference between the first recovery state and the second recovery state corresponding to the participant after walking recovery in different VR walking spaces is used as the recovery state difference corresponding to different VR walking spaces.

[0130] The prediction model training module 240 is used to train the prediction model according to different combinations of visual restorative environmental factors and changes in the recovery status of the participants in the VR walking space corresponding to the different combinations of visual restorative environmental factors.

[0131] In some embodiments, module 240 is specifically configured to:

[0132] The prediction model is trained based on different combinations of visual restorative environmental factors and the changes in the recovery status of the participants in the VR walking space corresponding to the different combinations of visual restorative environmental factors, including:

[0133] Different combinations of visual restorative environmental factors are used as training sample sets, and the changes in the recovery status of participants in the VR walking space corresponding to different combinations of visual restorative environmental factors are used as training output sets;

[0134] According to the training sample set and the training output set, a convolutional neural network is used to learn the relationship between each visual restorative environmental factor and the change in the recovery state of the participant and to train a prediction model for predicting the restorability of walking space.

[0135] In some embodiments, module 240 is further configured to:

[0136] The convolutional neural network is used to learn the relationship between various visual restorative environmental factors and the changes in the recovery status of the participants, including:

[0137] The correlation between each visual restorative environmental factor and the changes in the recovery status of the participants was obtained by Spearman correlation analysis method, and the convolutional neural network was used to learn the relationship between each visual restorative environmental factor and the changes in the recovery status of the participants.

[0138] The walking space visual restorative optimization module 250 is used to input various visual restorative environmental factors of the real walking space into the trained prediction model to obtain the change value of the recovery state of the person in the real walking space; according to the change value of the recovery state of the person in the real walking space, the various visual restorative environmental factors of the real walking space are optimized using a genetic algorithm.

[0139] In some embodiments, module 250 is specifically configured to:

[0140] According to the change value of the recovery state of personnel in the real walking space, the genetic algorithm is used to optimize the visual recovery environment factors of the real walking space, including:

[0141] The change value of the recovery state of the person in the real walking space is used as the fitness index of the genetic algorithm, and the extreme values ​​of each visual restorative environmental factor in different combinations are adjusted to obtain each visual restorative environmental factor in the combination that meets the preset fitness;

[0142] According to each visual restorative environmental factor in the combination that satisfies the preset fitness, the parameter range of each visual restorative environmental factor in the real walking space is optimized.

[0143] Understandably, Figure 2 Each module / unit in the detection device 200 shown has the function of implementing each step in the detection method 100 provided in the embodiment of the present disclosure and can achieve its corresponding technical effect. For the sake of brevity, it will not be repeated here.

[0144] Figure 3 A block diagram of an exemplary electronic device capable of implementing an embodiment of the present disclosure is shown. The electronic device 300 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device 300 may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0145] like Figure 3As shown, the electronic device 300 includes a computing unit 301, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 302 or a computer program loaded from a storage unit 308 into a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 can also be stored. The computing unit 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An I / O interface 305 is also connected to the bus 304.

[0146] A number of components in the electronic device 300 are connected to the I / O interface 305, including: an input unit 306, such as a keyboard, a mouse, etc.; an output unit 307, such as various types of displays, speakers, etc.; a storage unit 308, such as a disk, an optical disk, etc.; and a communication unit 309, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 309 allows the electronic device 300 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0147] The computing unit 301 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 301 performs the various methods and processes described above, such as method 100. For example, in some embodiments, the method 100 may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 308. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 300 via the ROM 302 and / or the communication unit 309. When the computer program is loaded into the RAM 303 and executed by the computing unit 301, one or more steps of the method 100 described above may be performed. Alternatively, in other embodiments, the computing unit 301 may be configured to perform the method 100 in any other appropriate manner (e.g., by means of firmware).

[0148] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0149] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0150] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0151] It should be noted that the present disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute method 100 and achieve the corresponding technical effect achieved by executing the method in the embodiment of the present disclosure. For the sake of concise description, they are not repeated here.

[0152] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0153] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0154] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0155] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0156] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for optimizing restorative environmental factors of walking space based on VR and deep learning, characterized in that: include: The visual restorative environmental factors of the walking space were screened out using an eye tracker experiment and a semantic differential questionnaire; wherein the visual restorative environmental factors included building height, tree height, shrub area in the field of vision, ground color and ground texture; According to the orthogonal experimental method, different combinations of visual restorative environmental factors are carried out to establish VR walking spaces corresponding to different combinations of visual restorative environmental factors. VR orthogonal experiments were conducted based on VR walking spaces corresponding to different combinations of visual restorative environmental factors, and the changes in participants' recovery status in different VR walking spaces were recorded; The prediction model is trained based on different combinations of visual restorative environmental factors and the changes in the recovery status of the participants in the VR walking space corresponding to the different combinations of visual restorative environmental factors; The visual restorative environmental factors of the real walking space are input into the trained prediction model to obtain the change value of the recovery state of the personnel in the real walking space; according to the change value of the recovery state of the personnel in the real walking space, the genetic algorithm is used to optimize the visual restorative environmental factors of the real walking space.

2. The method according to claim 1, characterized in that The VR orthogonal experiment includes: Through the pre-experiment, the participants were subjected to a stress experiment and their first recovery state was evaluated using the recovery state quantification table; After the participants wore the VR device and walked freely in the VR walking space corresponding to different combinations of visual restorative environmental factors for a preset period of time, the recovery state quantification table was used to evaluate the participants' second recovery state corresponding to different VR walking spaces; The difference between the first recovery state and the second recovery state corresponding to the participant after walking recovery in different VR walking spaces is used as the recovery state difference corresponding to different VR walking spaces.

3. The method according to claim 1, characterized in that The prediction model is trained based on different combinations of visual restorative environmental factors and changes in the recovery state of the participants in the VR walking space corresponding to the different combinations of visual restorative environmental factors, including: Different combinations of visual restorative environmental factors are used as training sample sets, and the changes in the recovery status of participants in the VR walking space corresponding to different combinations of visual restorative environmental factors are used as training output sets; According to the training sample set and the training output set, a convolutional neural network is used to learn the relationship between each visual restorative environmental factor and the change in the recovery state of the participant and to train a prediction model for predicting the restorability of walking space.

4. The method according to claim 3, characterized in that The method of using a convolutional neural network to learn the relationship between various visual restorative environmental factors and changes in the recovery status of the participants includes: The correlation between each visual restorative environmental factor and the changes in the recovery status of the participants was obtained by Spearman correlation analysis method, and the convolutional neural network was used to learn the relationship between each visual restorative environmental factor and the changes in the recovery status of the participants.

5. The method according to claim 1, characterized in that The optimization of various visual restorative environmental factors of the real walking space by using a genetic algorithm according to the change value of the recovery state of the person in the real walking space includes: The change value of the recovery state of the person in the real walking space is used as the fitness index of the genetic algorithm, and the extreme values ​​of each visual restorative environmental factor in different combinations are adjusted to obtain each visual restorative environmental factor in the combination that meets the preset fitness; According to each visual restorative environmental factor in the combination that satisfies the preset fitness, the parameter range of each visual restorative environmental factor in the real walking space is optimized.

6. The method according to claim 1, characterized in that The visual restorative environmental factors of the walking space screened out by using the eye tracker experiment and the semantic differential questionnaire include: The participants wore an eye tracker and calibrated it, pressed the button of the eye tracker to start recording data, and then walked freely in the preset real walking space for a preset time. The environmental components that the participants paid attention to in the preset real walking space were determined based on the eye movement data of the participants captured by the eye tracker, and a semantic differential questionnaire was designed based on the environmental components. The visual restorative environmental factors of the walking space were screened out through the semantic differential questionnaire.

7. The method according to claim 6, characterized in that The visual restorative environmental factors of the walking space screened out by the semantic differential questionnaire include: According to the environmental elements that the participants pay attention to in the preset real walking space, neutral factors affecting spatial visual recovery, adjective pairs corresponding to the neutral factors, and a semantic differential table are set, so that the participants can score the adjective pairs on the semantic differential table according to their own recovery status and the neutral factors; The scoring results of the participants were collected and analyzed according to the Spearman correlation analysis method to obtain various visual restorative environmental factors.

8. The method according to claim 6, characterized in that Before establishing the VR orthogonal experiment, the following steps are included: Establish a VR walking space that is consistent with the preset real walking space, so that participants can walk freely in the VR walking space that is consistent with the preset real walking space for a preset time after wearing the VR device, and record eye movement data based on the VR eye tracker worn; By comparing the eye movement data recorded by the VR eye tracker with the eye movement data of the participants captured by the eye tracker, it was determined that the VR orthogonal experiment was feasible.

9. A walking space restorative environmental factor optimization device based on VR and deep learning, characterized in that: include: A visual restorative environmental factor determination module is used to screen out visual restorative environmental factors of walking space using an eye tracker experiment and a semantic differential questionnaire; wherein the visual restorative environmental factors include building height, tree height, shrub area in the field of vision, ground color and ground texture; A VR walking space establishment module is used to make different combinations of visual restorative environmental factors according to the orthogonal experimental method to establish VR walking spaces corresponding to different combinations of visual restorative environmental factors; The recovery state change recording module is used to conduct VR orthogonal experiments based on VR walking spaces corresponding to different combinations of visual restorative environmental factors, and record the changes in the recovery state of participants in different VR walking spaces; A prediction model training module is used to train the prediction model according to different combinations of visual restorative environmental factors and changes in the recovery status of the participants in the VR walking space corresponding to the different combinations of visual restorative environmental factors; The visual restorative environmental factor optimization module is used to input various visual restorative environmental factors of the real walking space into the trained prediction model to obtain the change value of the recovery state of the personnel in the real walking space; according to the change value of the recovery state of the personnel in the real walking space, the genetic algorithm is used to optimize the various visual restorative environmental factors of the real walking space.

10. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 8.

Images