A method for analyzing architectural space design using VR to collect gaze points

Abstract

The present invention discloses a method for analyzing architectural space design by collecting gaze points using VR. A three-dimensional model matching the real architectural space is established, and a VR scene is constructed. When an experimenter tours the VR scene, the experimenter's gaze line is collected through VR equipment, the intersection of the gaze line and the VR scene is calculated, the experimenter's gaze point on the interior landscape of the building is obtained, and the gaze duration is counted. Finally, all the experimenter's gaze points and corresponding gaze durations are mapped and visualized to obtain intuitive visualization results, which can accurately, clearly, and intuitively evaluate the architectural space design. During the analysis process of the present invention, the stability and consistency of the spatial coordinate system are guaranteed, the collected gaze points match the architectural space, and the gaze points also correspond to the architectural space, making the association between the analysis results and the architectural space more intuitive.

Description

Technical Field

[0001] The present invention relates to the technical field of architectural space design evaluation, and in particular to a method for analyzing architectural space design by using VR to collect gaze points. Background Art

[0002] In traditional architectural design, the spatial experience is primarily judged by the architect's personal experience, often diverging from the actual user experience of the built space. Architectural ergonomics aims to establish a correlation between human factors measurement data and the quality of architectural spaces. By collecting and analyzing human factors data on how people experience architectural spaces, it is possible to objectively and quantitatively assess the quality of architectural spaces.

[0003] The distribution of gaze points and gaze duration when people are touring an architectural space reflect their attention to different interfaces and landmarks in the interior during the tour. By superimposing and analyzing the gaze points and gaze duration of a large number of subjects, it is possible to reflect the attention attraction and visual guidance of subjects by interior design elements, test whether the architectural space design and interior environment design are consistent with the design expectations, and assist in the design optimization of interior furnishings, guide signs, etc.

[0004] As virtual reality technology matures, VR scenes can be built during the design phase, allowing users to experience unfinished spaces through VR devices. This process is combined with eye-tracking technology, which tracks and records the user's eye movements. Based on this data, the user's gaze coordinates are calculated in real time to obtain detailed behavioral data.

[0005] Currently, most research using eye-tracking technology in architectural design focuses on two-dimensional images, such as a single rendering or video. This research overlays gaze data onto the two-dimensional image for data visualization and design analysis. This approach typically involves overlaying gaze data onto the two-dimensional image and using a Gaussian blur algorithm to create a saliency map. However, these methods don't directly map gaze points to the actual observed object. Therefore, researchers typically need to determine the specific object of attention through observation and analysis, or use deep learning techniques to identify elements in the image to assist in data analysis. Furthermore, designers need to convert the gaze analysis results based on the two-dimensional image into three-dimensional space to guide the optimization of spatial design.

[0006] The aforementioned application methods suffer from unintuitive application scenarios and inaccurate analysis results. Furthermore, they are subject to significant human influence and rely heavily on analysts, resulting in their inability to meet the requirements of quantitative analysis in practice. This present invention addresses these issues by providing a method for analyzing architectural space design using VR to capture gaze points. Summary of the Invention

[0007] The present invention provides a method for analyzing architectural space design by using VR to collect gaze points. The method adopts a three-dimensional model identical to the real architectural structure and clearly marks the gaze points, thereby achieving intuitive and quantitative analysis.

[0008] The technical solution adopted by the present invention to solve the above technical problems is:

[0009] A method for analyzing architectural space design by using VR to collect gaze points includes the following steps:

[0010] S1, setting up the model and VR scene: Build a 3D model of the building, then perform lightweight processing on the 3D model to ensure rendering efficiency and visual quality; import the processed 3D model, and then adjust the parameters, visual elements and interior layout to form the required VR scene;

[0011] S2, conduct VR tour: set the starting point of the VR tour, define the VR navigation range, configure the data recording program, and then the experimenter conducts the VR tour and records the raw data of the tour through the VR device. The raw data includes the gaze point coordinate data and the corresponding timestamp data;

[0012] S3, raw data import: import the raw data into the data processing model, create the original fixation point with the x, y, z coordinate data of the original trajectory point, convert the timestamp data corresponding to the original fixation point, and calculate the time difference as the original fixation duration of the original fixation point;

[0013] S4, raw data screening: set a threshold, and filter out raw gaze points that do not meet the requirements according to the set threshold to obtain stable data, which includes stable gaze points and stable gaze duration;

[0014] S5, data export: process the stable data of each experimenter according to the operation of S4, and write the stable data into the data table after processing;

[0015] S6, data import: read the stable data of all experimenters, determine the center point of the stable gaze point, and calculate the total duration of stable gaze;

[0016] S7, data visualization: Draw a visualization sphere based on the gaze duration. Draw the sphere with the center point of the stable gaze point as the center and the relevant data of the total stable gaze duration as the radius. The sphere is colored by the color value mapped by the relevant data of the total stable gaze duration to achieve data visualization.

[0017] Furthermore, in step S2, the operation of determining the gaze point is as follows:

[0018] The center coordinates of the virtual camera representing the VR device and the direction vector of the virtual camera's front direction relative to the world coordinate system are extracted frame by frame. Then, the direction vector of the human eye's line of sight captured by the eye tracking sensor relative to the virtual camera coordinate system is extracted. The above two direction vectors are added together to obtain the direction vector of the human eye's line of sight relative to the world coordinate system. The direction vector of the gaze direction is determined. Starting from the camera center point, a ray is drawn along the gaze direction. The first intersection of the ray and the architectural visual model is the gaze point.

[0019] Furthermore, in step S4, the saccade gaze points in the original gaze points are filtered out. The specific filtering operation is:

[0020] The original fixation points were grouped with a distance of 300 mm as the threshold, and the total fixation duration corresponding to each group of fixation points was superimposed. The fixation points with a total fixation duration of less than 500 ms were grouped and removed as a whole.

[0021] Furthermore, in step S6, the specific operation of determining the center point of the stable gaze point is as follows: using 300mm as the threshold, group the stable gaze points of all the experimenters into F groups, and superimpose the x, y, and z coordinates of all the stable gaze points in the group and calculate the average value. The average value of the x, y, and z coordinates is the center point P of each group of stable gaze points. f The x, y, and z coordinates of

[0022] When calculating the total duration of stable fixation, the stable fixation duration corresponding to each group of stable fixation points is grouped in the same way, and all stable fixation durations in the group are superimposed to obtain the total duration of stable fixation in each group, t f .

[0023] Furthermore, in step S7, the radius R of the visualized sphere is f The calculation formula is:

[0024]

[0025] K is the coefficient, which is determined according to the proportion of building space.

[0026] Furthermore, in step S7, the total duration of each group of stable gaze duration t f The color value mapped by the numerical value is the color of the visualization sphere.

[0027] The beneficial effects of the present invention are as follows:

[0028] VR modeling is based on real architectural spaces, and eye movement data is used to calculate the user's gaze coordinates in real time. These gaze points are then imported into the 3D model space for analysis and visualization directly in the 3D environment. The gaze points are clearly marked, and after calculation, the gaze duration is clearly indicated with a visual sphere, enabling quantitative analysis of architectural space design and direct comparison.

[0029] The stability and consistency of the spatial coordinate system are ensured during the analysis process. The collected fixations match the architectural space, and the fixations also correspond to the architectural space, making the relationship between the analysis results and the architectural space more intuitive.

[0030] It can overlay data from multiple people for comprehensive analysis, making the results more accurate. Data overlay and analysis can be achieved without complicated settings. It is not affected by human factors and can achieve accurate quantitative analysis without relying on analysts. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Schematic diagram of the process steps of the present invention;

[0032] Figure 2 Schematic diagram of the original gaze point distribution of the present invention;

[0033] Figure 3 Schematic diagram of the stable gaze point distribution of the present invention;

[0034] Figure 4 This is a top view schematic diagram of the architectural space after visualization processing of the present invention;

[0035] Figure 5 This is a schematic diagram of architectural space gaze after visualization processing of the present invention. DETAILED DESCRIPTION

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0037] In the description of the present invention, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship 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 cannot be understood as limiting the present invention.

[0038] like Figures 1 to 5 As shown in the figure, a method for analyzing architectural space design by using VR to collect gaze points is proposed. A three-dimensional model matching the real architectural structure is established, and a VR scene is built. When the experimenter tours in the VR scene, the experimenter's gaze line is collected through the VR equipment, and the intersection of the gaze line and the VR scene is calculated to obtain the experimenter's gaze point on the internal landscape of the building, and the gaze duration is counted. Finally, the gaze points and corresponding gaze durations of all the experimenters are mapped and visualized to obtain intuitive visualization results, which can accurately, clearly and intuitively evaluate the architectural space design.

[0039] The implementation principle of the present invention is to first build a VR tour scene based on a real architectural design plan. After the experimenter conducts a VR tour, the VR equipment can collect the gaze points and gaze duration that match the real architectural space, cluster and group the gaze points, and calculate the total gaze duration in groups. The corresponding data is then converted into a visual image, thereby forming an intuitive visual quantitative evaluation of the architectural space. The method adopted by the present invention is to directly place the gaze points on a three-dimensional model that matches the real architectural structure. The determination of the gaze points is accurate and accurate. Compared with the existing spatial analysis through two-dimensional images, it has the advantages of accurate data collection, no need for multiple conversions, and no need for manual analysis. The final visualization result is directly displayed on the three-dimensional model and displayed proportionally according to the total gaze duration, which can evaluate the architectural space design at a glance.

[0040] like Figure 1 As shown, a specific implementation method of a method for analyzing architectural space design by using VR to collect gaze points is as follows.

[0041] S1, Setting up the model and VR scene: Build a 3D model of the building, then perform lightweight processing on the 3D model to ensure smoothness in the VR scene, improve rendering efficiency, and ensure visual quality. "Visibility" is used as the judgment standard. Specifically, the operation adopted is to remove all parts of the model that are not visible in the final VR experience in the modeling software. For objects with smooth surfaces, they are converted into multiple continuous planes, such as converting a cylinder into a polygonal column.

[0042] Build a VR scene, import the processed 3D model into the modeling software, adjust the model's material parameters according to the architectural design plan, and arrange elements such as lighting, tables and chairs, furniture, and plant decorations. At the same time, set the ambient lighting parameters to optimize the scene's realism and immersion, and finally build the required VR scene.

[0043] S2, conduct VR tour: set the starting point of the VR tour, define the VR navigation range according to the expected navigable area in the VR scene, configure the data recording program, and then have the experimenter conduct the VR tour. While conducting the tour, the VR device records the original data of the experimenter's tour frame by frame. After the tour, the data is saved and exported through the VR device. The original data includes the gaze point coordinate data and the corresponding timestamp data.

[0044] The gaze point is determined in the following way: the center coordinates of the virtual camera representing the VR device and the direction vector of the virtual camera's front direction relative to the world coordinate system are extracted frame by frame, and then the direction vector of the human eye's line of sight captured by the eye tracking sensor relative to the virtual camera coordinate system is extracted. The above two direction vectors are added together to obtain the direction vector of the human eye's line of sight relative to the world coordinate system. The direction vector of the gaze direction is determined, and a ray is drawn along the gaze direction with the camera center point as the starting point. The first intersection of the ray and the building visual model is the gaze point, and the timestamp of this gaze point is recorded at the same time.

[0045] like Figure 2 As shown, S3, raw data import: import the raw data into the data processing model, create the original gaze point with the x, y, z coordinate data of the original trajectory point, convert the timestamp data corresponding to the original gaze point into milliseconds, and calculate the time difference between each frame and the previous frame, which is used as the original gaze duration of the original gaze point.

[0046] like Figure 3 As shown in S4, raw data screening: Since the entire process of human observation is continuous, in addition to relatively stable gazes, there are also aimless observations such as scans. It is necessary to screen out the non-gaze process and leave only the gaze process. Since the gaze duration in the scan process is extremely short and the scan process is fast, the distance between the gaze points is also large. The scan process can be screened out by setting the screening conditions of the gaze point distance and the gaze duration. The screening operation is to group the original gaze points by setting the distance threshold, and screen out the original gaze points that do not meet the requirements to obtain stable data. The stable data includes stable gaze points and stable gaze duration.

[0047] The original fixation points that are in the scanning state are filtered out. The specific filtering operation is to group the original fixation points with a distance threshold of 300 mm, calculate the total fixation duration corresponding to each group of fixation points, and group the fixation points with a total fixation duration of less than 500 ms and delete the entire group. Only the fixation points with a certain fixation duration are left.

[0048] S5, data export: process the stable data of each experimenter according to the operation of S4, write the stable data into the data table after processing, and proceed to the next step of visual analysis.

[0049] S6, data import: read the stable data of all experimenters, determine the center point of the stable gaze point, and calculate the total duration of stable gaze.

[0050] The original gaze point data is large and relatively scattered. Visual conversion based on the original gaze point data will make the results too scattered and unable to be compared intuitively. The original gaze points are grouped to determine stable gaze points, and the data of stable gaze points are used for visual conversion. Stable gaze points are more concentrated and have a longer total gaze time, so they can be compared intuitively after visual conversion.

[0051] The specific operation of determining the center point of the stable gaze point is as follows: using 300mm distance as the threshold, group the stable gaze points of all the experimenters into F groups, and superimpose the x, y, and z coordinates of all the stable gaze points in the group and calculate the average value. The average value of the x, y, and z coordinates is the center point P of each group of stable gaze points. f The x, y, and z coordinates of

[0052] When calculating the total duration of stable fixation, the stable fixation duration corresponding to each group of stable fixation points is grouped in the same way, and all stable fixation durations in the group are superimposed to obtain the total duration of stable fixation in each group, t f .

[0053] like Figure 4 、 5 As shown in S7, data visualization: Draw a visualization sphere based on the fixation time, with the center point of the stable fixation point as the center, and the total stable fixation time as the radius to draw the sphere, and the total stable fixation time of each group t f The color value mapped to the numerical value is the color of the visualization sphere to achieve data visualization.

[0054] Visualize the radius R of the sphere f The calculation formula is:

[0055]

[0056] K is the coefficient, which is determined according to the proportion of building space.

[0057] The present invention maintains the consistency of the spatial coordinate system throughout the entire data analysis process, and the collected gaze points match the real architectural space while maintaining a direct correlation, so that the visualization results also correspond to the real architectural space, and ultimately the analysis results can be intuitively displayed on the architectural space. The direct correlation between the gaze points and the architectural space is maintained throughout the entire process, ensuring the reliability of the collected data and the accuracy of the visualization analysis results.

[0058] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description. It is intended that all variations within the meaning and range of equivalents of the claims be embraced herein, and any reference signs in the claims should not be construed as limiting the claims to which they relate.

Claims

1. A method for analyzing architectural space design by using VR to collect gaze points, characterized by: The following steps are involved: S1, setting up the model and VR scene: Build a 3D model of the building, then perform lightweight processing on the 3D model to ensure rendering efficiency and visual quality; import the processed 3D model, and then adjust the parameters, visual elements and interior layout to form the required VR scene; S2, conduct VR tour: set the starting point of the VR tour, define the VR navigation range, configure the data recording program, and then the experimenter conducts the VR tour and records the raw data of the tour through the VR device. The raw data includes the gaze point coordinate data and the corresponding timestamp data; S3, raw data import: Import the raw data into the data processing model, create the original fixation point with the x, y, z coordinate data of the original trajectory point, convert the timestamp data corresponding to the original fixation point, and calculate the time difference as the original fixation duration of the original fixation point; S4, raw data screening: set a threshold, and filter out raw gaze points that do not meet the requirements according to the set threshold to obtain stable data, which includes stable gaze points and stable gaze duration; S5, data export: process the stable data of each experimenter according to the operation of S4, and write the stable data into the data table after processing; S6, data import: read the stable data of all experimenters, determine the center point of the stable gaze point, and calculate the total duration of stable gaze; S7, data visualization: Draw a visualization sphere based on the fixation duration. The sphere is centered on the center point of the stable fixation point and the total stable fixation duration data as the radius. The sphere is colored by the color value mapped by the total stable fixation duration data to achieve data visualization. In step S2, the operation of determining the gaze point is as follows: The center coordinates of the virtual camera representing the VR device and the direction vector of the virtual camera's front direction relative to the world coordinate system are extracted frame by frame. Then, the direction vector of the human eye's line of sight captured by the eye tracking sensor relative to the virtual camera coordinate system is extracted. The above two direction vectors are added together to obtain the direction vector of the human eye's line of sight relative to the world coordinate system. The direction vector of the gaze direction is determined. Starting from the camera center point, a ray is drawn along the gaze direction. The first intersection of the ray and the building visual model is the gaze point, and the timestamp of this gaze point is recorded at the same time.

2. The method for analyzing architectural space design by using VR to collect gaze points according to claim 1, characterized in that: In step S4, the saccade gaze points in the original gaze points are filtered out. The specific filtering operation is as follows: The original fixation points were grouped with a distance of 300 mm as the threshold, and the total fixation duration of each group of fixations was superimposed. The fixation points with a total fixation duration of less than 500 ms were grouped and removed as a whole.

3. The method for analyzing architectural space design by using VR to collect gaze points according to claim 2, characterized in that: In step S6, the specific operation of determining the center point of the stable gaze point is as follows: using 300mm as the threshold, group the stable gaze points of all the experimenters into F groups, and superimpose the x, y, and z coordinates of all the stable gaze points in the group and calculate the average value. The average value of the x, y, and z coordinates is the center point P of each group of stable gaze points. f The x, y, and z coordinates of When calculating the total duration of stable fixation, the stable fixation duration corresponding to each group of stable fixation points is grouped in the same way, and all stable fixation durations in the group are superimposed to obtain the total duration of stable fixation in each group, t f .

4. The method for analyzing architectural space design by using VR to collect gaze points according to claim 3 is characterized by: In step S7, the radius R of the visualized sphere is f The calculation formula is: ； K is the coefficient, which is determined according to the proportion of the building space, and the total duration of each group of stable gaze time is t f The color value mapped by the numerical value is the color of the visualization sphere.

Images