A method and system for evaluating campus landscape green visual acuity and visual health recovery

Abstract

The application provides a campus landscape green view rate and visual health recovery evaluation method and system, comprising: identifying an initial imbalance degree according to a green view rate quantitative value and a sky visibility quantitative value, extracting a fixation point concentration through a fixation point spatial coordinate record, and extracting a pupil change amplitude through a pupil diameter change record; analyzing the fixation point concentration, identifying the inhibition of the pupil change amplitude when the fixation point concentration is high and the compensation of the fixation point dispersion when the fixation point concentration is low, and comprehensively obtaining an adjusted imbalance degree; determining a green view rate weight and a sky visibility weight according to the initial imbalance degree and the adjusted imbalance degree, integrating the green view rate quantitative value and the sky visibility quantitative value according to the weights to obtain a comprehensive visual recovery score; extracting a peak interval from a time record of the comprehensive visual recovery score, numerically filling the green view rate quantitative value and the sky visibility quantitative value in the peak interval, and identifying a candidate green view rate range and a candidate sky visibility range.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method and system for assessing the green visibility rate and visual health restoration of campus landscapes. Background Technology

[0002] Campus landscape design is increasingly emphasizing visual health benefits, particularly by increasing the green view ratio to alleviate students' visual fatigue. This has become an important direction in environmental psychology and landscape planning, as prolonged close-range eye use can lead to ciliary muscle tension and limited autonomic nervous system regulation, affecting learning efficiency and mental and physical health. The green view ratio, as a core indicator, reflects the proportion of green vegetation in the field of vision, and its increase is generally considered to contribute to a relaxing effect. However, in practical applications, this single indicator cannot be simply relied upon to judge the recovery effect.

[0003] Current assessment methods often emphasize increasing green view rate in isolation, neglecting the interplay of other elements in landscape design. This leads to inconsistent optimization results across different scenarios. For example, while planting tall trees is often used to achieve a higher green view rate, excessively dense canopies can obstruct the sky, reducing the sense of openness and directly impacting opportunities for distant viewing. In such an environment, observers tend to focus their gaze on nearby vegetation, resulting in a more concentrated spatial distribution and a smaller range of pupil diameter variation, reflecting limitations in physiological relaxation. Specifically, while tall trees increase green view rate, excessive canopy closure can significantly obscure the sky, leading to a lack of distant targets, decreased gaze dispersion, prolonged near-field accommodation by the ciliary muscle, and narrowed pupil dynamics. Consequently, the relaxing effect of green stimulation cannot be effectively balanced with the physiological soothing effect of distant viewing. This trade-off directly makes it difficult for assessment systems to accurately capture the true degree of restoration.

[0004] Therefore, how to dynamically balance the impact of increased green visibility and decreased sky visibility in campus landscape assessment, while comprehensively considering the spatial dispersion of fixation point and the range of pupil diameter changes, has become a key issue in achieving comprehensive optimization of visual fatigue relief effects. Summary of the Invention

[0005] On the one hand, this invention provides a method for assessing the green view rate and visual health restoration of campus landscapes, including: The panoramic image acquisition module extracts the proportion of green pixels and the ratio of sky area from the field of view elevation angle. Simultaneously, it collects the spatial coordinates of the gaze point, the dispersion of the gaze point, and the pupil diameter change. The green pixel proportion is normalized to obtain the green visibility rate quantification value, and the sky area ratio is threshold filtered to obtain the sky visibility quantification value. The initial degree of imbalance is identified based on the green visibility rate quantification value and the sky visibility quantification value; the fixation point concentration is extracted by the fixation point spatial coordinate record; and the pupil change amplitude is extracted by the pupil diameter change record. The fixation point concentration is analyzed to identify the inhibitory effect of high fixation point concentration on pupil change amplitude and the compensatory effect of low fixation point concentration on fixation point dispersion, and the degree of adjustment imbalance is obtained by combining the results. The green visibility weight and sky visibility weight are determined based on the initial imbalance degree and the adjusted imbalance degree. The green visibility quantification value and the sky visibility quantification value are integrated based on the green visibility weight and the sky visibility weight to obtain a comprehensive visual restoration score. Peak intervals are extracted from the time records of the comprehensive visual recovery score, and the green visibility rate quantification value and sky visibility quantification value within the peak interval are numerically filled to identify candidate green visibility rate ranges and candidate sky visibility ranges. K-means clustering analysis is used for the candidate green visibility range and the candidate sky visibility range to extract cluster centers and determine the optimal green visibility boundary and the optimal sky visibility boundary. The target green visibility interval and the target sky visibility interval are obtained from the optimal green visibility boundary and the optimal sky visibility boundary. Based on the target green visibility range and the target sky visibility range, the dispersion of the gaze point and the amplitude of pupil change collected subsequently are evaluated in real time to identify the matching degree. When the matching degree reaches a high matching level, the current landscape green visibility and sky visibility combination is output as the final optimization result.

[0006] Furthermore, the process of extracting the proportion of green pixels and the ratio of sky area from the field of view elevation angle range using a panoramic image acquisition module, simultaneously acquiring records of gaze point spatial coordinates, gaze point dispersion, and pupil diameter changes, normalizing the green pixel proportion to obtain a quantified value of green visibility, and threshold filtering the sky area ratio to obtain a quantified value of sky visibility, includes: The panoramic image acquisition module continuously captures environmental image data within the vertical field of view from negative to positive elevation angle. The HSV color space conversion is used to identify green vegetation areas, and the Otsu threshold segmentation method is used to extract the sky area. The ratio of the number of pixels in the green vegetation area to the total number of pixels is recorded as the basic value of the green proportion, and the ratio of the sky area to the total area of ​​the upper half of the field of view is recorded as the basic value of the sky proportion. The total length of the gaze trajectory is calculated by recording the gaze point angle coordinate sequence using an eye-tracking device, and the pupil major axis diameter value is extracted and recorded as the basic value of the pupil change amplitude using an infrared camera. The green proportion base value is normalized to obtain the green visibility rate quantification value, the sky area ratio is threshold filtered to obtain the sky visibility quantification value, and the gaze point dispersion is calculated based on the total length of the gaze trajectory.

[0007] Furthermore, the initial degree of imbalance is identified based on the quantified value of green visibility and the quantified value of sky visibility; the fixation point concentration is extracted through the fixation point spatial coordinate record; and the pupil change amplitude is extracted through the pupil diameter change record, including: The difference between the quantified value of green visibility and the quantified value of sky visibility is calculated, and the difference is divided by the upper limit of the imbalance threshold to obtain the initial imbalance degree. The spatial distance between adjacent sampling points is calculated from the spatial coordinate record of the fixation point, and the proportion of the number of fixation points in the focal area to the total number of fixation points is used as the fixation point concentration. The pupil data processing method is adjusted according to the fixation point concentration, and the diameter difference sequence is extracted from the pupil diameter change record to calculate the pupil change amplitude.

[0008] Furthermore, the fixation concentration is analyzed to identify the inhibitory effect on pupillary variation amplitude when fixation concentration is high and the compensatory effect on fixation dispersion when fixation concentration is low, and the degree of imbalance is comprehensively adjusted, including: The inhibition state is determined by comparing the fixation point concentration with a preset concentration threshold. If the fixation point concentration exceeds the preset threshold, the ratio of the pupil change amplitude to the median of the normal amplitude range is calculated. The inhibition coefficient is obtained by subtracting the ratio from one. The inhibition correction value is obtained by multiplying the inhibition coefficient by the initial imbalance degree. When the fixation point concentration is lower than the preset concentration threshold, the geometric center of all fixation point coordinates is calculated from the fixation point spatial coordinate sequence, and the standard deviation of the distance from each fixation point to the geometric center is calculated as the fixation point dispersion. The compensation coefficient is determined according to the ratio of the fixation point dispersion to the preset standard dispersion, and the compensation correction value is obtained by multiplying the compensation coefficient by the initial imbalance degree. Based on the state of fixation concentration, the suppression correction value or the compensation correction value is selected, and the selected correction value and the initial imbalance degree are weighted and summed according to a preset weight ratio to obtain the adjusted imbalance degree.

[0009] Furthermore, based on the initial degree of imbalance and the adjusted degree of imbalance, green visibility weight and sky visibility weight are determined. A comprehensive visual restoration score is obtained by integrating the quantified values ​​of green visibility and sky visibility based on these weights, including: The weight adjustment factor is obtained by dividing the difference between the initial imbalance degree and the adjusted imbalance degree by the initial imbalance degree. The weight adjustment factor is set as the green visibility weight, and the sky visibility weight is set as one minus the green visibility weight. The comprehensive visual restoration score is obtained by multiplying the green visibility weight by the green visibility quantification value and the sky visibility weight by the sky visibility quantification value, and then adding the two product results.

[0010] Furthermore, peak intervals are extracted from the time records of the comprehensive visual restoration score, and the quantified values ​​of green visibility and sky visibility within the peak intervals are numerically filled to identify candidate green visibility ranges and candidate sky visibility ranges, including: Divide the time record of the comprehensive visual recovery score into continuous segments, calculate the average value of each segment, and record the start and end times of the average value of segments that continuously exceed a preset number threshold and the peak threshold as the peak interval. Linear interpolation is performed on the quantized green visibility rate sequence and the quantized sky visibility rate sequence within the peak interval to fill in the numerical values. The maximum and minimum values ​​are extracted from the padded sequence as candidate green visibility ranges and candidate sky visibility ranges.

[0011] Furthermore, k-means clustering analysis is applied to the candidate green visibility range and the candidate sky visibility range to extract cluster centers and determine the optimal green visibility boundary and the optimal sky visibility boundary. The target green visibility interval and the target sky visibility interval are then obtained from the optimal green visibility boundary and the optimal sky visibility boundary, including: The candidate green visibility range and the candidate sky visibility range are evenly divided into sampling points, and green visibility sample values ​​and sky visibility sample values ​​are extracted and combined into two-dimensional data points to construct a feature space dataset. The feature space dataset is processed using the k-means clustering algorithm. Initial cluster centers are randomly selected, and the Euclidean distance from each data point to each cluster center is calculated to allocate data points. The mean within each cluster is recalculated as the new cluster center. Repeat the distance calculation and center update process until the change in the center of the adjacent iteration is less than the convergence threshold, and identify the effective clusters containing more than the preset proportion threshold from the clustering results after convergence. The maximum and minimum values ​​of data points in the effective clusters are extracted as the upper and lower boundaries of the target green visibility interval in the green visibility dimension, and the maximum and minimum values ​​in the sky visibility dimension are extracted as the upper and lower boundaries of the target sky visibility interval.

[0012] Furthermore, based on the target green visibility range and the target sky visibility range, the dispersion of the subsequently collected fixation point and the amplitude of pupil change are evaluated in real time to identify the matching degree. When the matching degree reaches a high matching level, the current landscape green visibility and sky visibility combination is output as the final optimization result, including: Real-time acquisition of subsequent fixation point dispersion values ​​and pupil change amplitude values; determination of current landscape green visibility rate within the target green visibility rate range and sky visibility within the target sky visibility range; setting of environmental matching markers. When the environment matching flag is true, check that the gaze point dispersion and the pupil change amplitude both meet the threshold conditions, and calculate the average of the two ratios to obtain the matching degree value. When the matching degree value exceeds the high matching threshold, the current green visibility value and sky visibility value are extracted as the final optimization result output.

[0013] On the other hand, the present invention provides a campus landscape green view rate and visual health restoration assessment system, mainly comprising: The data acquisition and quantification module is used to extract the proportion of green pixels and the ratio of sky area from the field of view elevation angle range through the panoramic image acquisition module. It simultaneously acquires the spatial coordinates of the gaze point, the gaze point dispersion, and the pupil diameter change record. The green pixel proportion is normalized to obtain the green visibility rate quantification value, and the sky area ratio is threshold filtered to obtain the sky visibility quantification value. The imbalance identification and feature extraction module is used to identify the initial degree of imbalance based on the green visibility rate quantification value and the sky visibility quantification value, extract the fixation point concentration through the fixation point spatial coordinate record, and extract the pupil change amplitude through the pupil diameter change record. The adjustment analysis module is used to analyze the fixation point concentration, identify the inhibitory effect of pupil change amplitude when fixation point concentration is high and the compensatory effect of fixation point dispersion when fixation point concentration is low, and comprehensively obtain the degree of adjustment imbalance. The weight determination and scoring module is used to determine the green visibility weight and sky visibility weight based on the initial imbalance degree and the adjusted imbalance degree, and to integrate the green visibility quantification value and the sky visibility quantification value to obtain a comprehensive visual restoration score. The peak interval extraction module is used to extract the peak interval from the time record of the comprehensive visual recovery score, fill the green vision rate quantification value and sky visibility quantification value within the peak interval with numerical values, and identify candidate green vision rate range and candidate sky visibility range. The clustering analysis module is used to perform k-means clustering analysis on the candidate green visibility range and the candidate sky visibility range to determine the target green visibility range and the target sky visibility range; The real-time evaluation and output module is used to evaluate the dispersion of the gaze point and the amplitude of pupil change in real time based on the target green visibility range and the target sky visibility range, identify the matching degree, and determine the current landscape green visibility and sky visibility combination when the matching degree reaches a high matching level.

[0014] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a method and system for assessing the green visibility ratio and visual health recovery in campus landscapes. It addresses the unique business scenario of the impact of the proportion of green visual elements and sky visibility in natural landscapes on human psychological balance, proposing a systematic solution. The problem focuses on how to evaluate and optimize the psychological recovery effect of individuals in natural environments by quantifying visual environmental elements. This invention extracts the ratio of green pixel proportion to sky area through panoramic image analysis, combines fixation point data and pupil change records to comprehensively assess the degree of imbalance between the initial and adjusted states, and then dynamically determines the weights, integrating visual element scores. Simultaneously, cluster analysis is used to determine the optimal visual parameter boundaries, and the matching degree of subsequent data is evaluated in real time, ultimately outputting an optimized combination. This invention, through multi-dimensional data fusion and dynamic adjustment, significantly enhances the positive effect of the visual environment on psychological balance, providing a scientific basis for personalized environmental optimization. Attached Figure Description

[0015] Figure 1 This is a flowchart of a method for assessing the green visibility rate and visual health restoration of a campus landscape according to the present invention.

[0016] Figure 2 This is a schematic diagram of a method for assessing the green visibility rate and visual health restoration of a campus landscape according to the present invention.

[0017] Figure 3 This is a schematic diagram of the structure of a campus landscape green view rate and visual health restoration assessment system according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of the present invention.

[0019] like Figures 1-3 This embodiment of a method and system for assessing the green visibility rate and visual health restoration of a campus landscape may specifically include: S101. The panoramic image acquisition module extracts the proportion of green pixels and the ratio of sky area from the field of view elevation angle range. Simultaneously, it collects the spatial coordinates of the gaze point, the dispersion of the gaze point, and the pupil diameter change record. The green pixel proportion is normalized to obtain the green visibility rate quantification value, and the sky area ratio is threshold filtered to obtain the sky visibility quantification value.

[0020] The panoramic image acquisition module continuously captures images within the vertical field of view, ranging from negative to positive elevation angles, to obtain horizontal full-view environmental image data. HSV color space conversion is used to identify pixels with H values ​​in the green range and S values ​​exceeding a preset saturation threshold as green vegetation areas. The Otsu thresholding method is used to extract connected components with brightness values ​​exceeding an adaptive threshold from the upper region of the image as the sky region. The ratio of the number of pixels in the green vegetation area to the total number of pixels is recorded as the basic value of the green proportion, and the ratio of the sky region area to the total area of ​​the upper half of the field of view is recorded as the basic value of the sky proportion. Simultaneously with image acquisition, an eye-tracking device records the horizontal and vertical angle values ​​of the observer's gaze point in the panoramic coordinate system at a fixed sampling frequency. The angular distance between adjacent gaze points is calculated based on the gaze point angle coordinate sequence, and the total length of the gaze trajectory is obtained by accumulating these angular distances. An infrared camera captures the pupil edge contour in real time, and an ellipse fitting method is used to extract the pupil's major axis diameter. The difference between the maximum and minimum pupil diameter values ​​within the observation period is recorded as the basic value of the pupil change amplitude. The green proportion baseline value is normalized, and a benchmark green visibility threshold is set according to the scene type. When the green proportion baseline value is lower than a preset proportion of the benchmark threshold, a non-linear mapping function is used to map the baseline value to the low segment of the zero-to-one interval. When it is higher than the benchmark threshold, a compression mapping function is used to map the baseline value to the high segment of the zero-to-one interval, thus obtaining the quantified green visibility value. For the sky proportion baseline value, a minimum sky visibility threshold is set. When the sky proportion baseline value is lower than this threshold, the sky visibility quantified value is output as zero. When it is higher than this threshold, the sky proportion baseline value is mapped to the zero-to-one interval according to a linear proportion, thus obtaining the sky visibility quantified value. The gaze point movement rate is calculated based on the total length of the gaze trajectory and the observation duration. The gaze point coordinate sequence is spatially grouped using a clustering method to obtain the gaze point dispersion. The baseline value of the pupil change amplitude is recorded as the pupil diameter change record.

[0021] Specifically, in one implementation, the panoramic image acquisition module uses a fisheye lens in conjunction with a CMOS sensor to achieve omnidirectional acquisition of environmental images. The lens focal length and field of view are set to ensure that the entire field of view is covered when the observer is standing or sitting normally. During image acquisition, the sensor continuously captures images at an appropriate frame rate, and the image resolution setting ensures accurate detail capture. The acquired RGB image data is first subjected to distortion correction to eliminate edge distortion caused by the fisheye lens. Then, the corrected image is converted from the RGB color space to the HSV color space. During the conversion, the H component represents hue, with a value range of 0 to 360 degrees; the S component represents saturation, with a value range of 0 to 1; and the V component represents brightness, also with a value range of 0 to 1.

[0022] Specifically, the identification of green vegetation areas is achieved by setting a judgment range for the H component. When the H value is within the angular range corresponding to the green hue, the pixel is initially determined to be within the green range. Simultaneously, combined with the condition that the S component is greater than a preset saturation threshold, interference from non-vegetation colors such as gray-green or light green is eliminated. Pixels filtered under these dual conditions are marked as green vegetation areas. The total number of these pixels is counted and divided by the total number of pixels in the image to obtain the basic value of the green proportion. This basic value directly reflects the original coverage of green vegetation in the observer's field of view, providing basic data for subsequent normalization processing.

[0023] It should be noted that the sky region extraction employs the Otsu thresholding method, which automatically determines the optimal segmentation threshold by calculating the inter-class variance of the image's grayscale histogram. In actual processing, the upper half of the panoramic image is first converted to grayscale, and the pixel distribution probability for each grayscale level is calculated. By traversing all possible thresholds, the threshold point that maximizes the inter-class variance between foreground and background pixels is found. When the average brightness value of a connected component exceeds this adaptive threshold, that region is identified as the sky region. The ratio of the sky region's area to the total area of ​​the upper half of the field of view is the baseline value for the sky proportion, reflecting the degree to which the observer can see an open sky when looking upwards.

[0024] In one possible implementation, the eye-tracking device employs infrared tracking technology based on corneal reflection. It calculates the gaze direction by emitting invisible near-infrared light into the eye and capturing changes in the position of the reflected light spot on the corneal surface. The device's sampling frequency is set to a fixed value to ensure the capture of subtle movements such as rapid eye movements and microsaccades. The gaze point position recorded at each sampling moment includes data in both horizontal and vertical angles, forming a gaze trajectory in a spherical coordinate system. The angular distance between two adjacent sampling points is obtained using a spherical distance calculation formula, which calculates the shortest arc length on a sphere based on the latitude and longitude coordinates of the two points. The total length of the gaze trajectory is obtained by summing the distances between all adjacent points; a larger length indicates more frequent gaze movements and a higher degree of attentional distraction.

[0025] For example, pupil diameter is detected by capturing real-time images of the eye using an infrared camera equipped with an 850nm wavelength infrared LED light source, providing stable illumination without affecting the observer's vision. After binarization, the captured eye images show the pupil area as a distinct dark elliptical region. The ellipse fitting method, using the least squares principle, fits the ellipse equation to the pupil edge points, obtaining the major and minor axis parameters of the ellipse. The major axis diameter is recorded as a representative value of the pupil diameter. Throughout the observation period, the pupil diameter changes with variations in light and psychological state. The difference between its maximum and minimum values ​​is recorded as the baseline value of the pupil change amplitude. This value reflects the activity level of the autonomic nervous system; a larger amplitude generally indicates more active physiological regulation.

[0026] Preferably, the normalization of green visibility rate adopts a segmented mapping strategy. Corresponding benchmark green visibility rate thresholds are set according to different campus scenarios, and benchmark values ​​for different functional areas are set separately based on their usage characteristics. When the actual measured base value of green visibility rate is lower than the benchmark value, a sigmoid function is used for mapping, amplifying small differences in the low-value range. When the base value of green visibility rate is higher than the benchmark value, a logarithmic function is used for compression mapping. When the actual measured base value of green visibility rate is lower than a preset proportion of the benchmark value, a sigmoid function is used for mapping. In the expression of the sigmoid function, the input value is the base value of green visibility rate, and the steepness parameter of the function is adjusted according to the scenario type, amplifying small differences in the low-value range and enhancing distinguishability. When the base value of green visibility rate is higher than the benchmark value, a logarithmic function is used for compression mapping to avoid numerical saturation problems in high green visibility rate scenarios, ensuring that the normalized green visibility rate quantification value always remains within the standard range of 0 to 1, facilitating subsequent comprehensive evaluation.

[0027] In one embodiment, the quantification of sky visibility employs a clear threshold judgment mechanism, with the minimum sky visibility threshold dynamically adjusted based on the environment type. When the actual measured baseline sky proportion is lower than the corresponding threshold, it indicates that the sky is severely obscured, and opportunities for distant views are extremely rare; in this case, the quantified sky visibility value is directly output as zero. When the baseline sky proportion is higher than the threshold, a linear mapping is used to transform it to the 0-1 range, with the mapping slope adjusted according to scene characteristics to ensure that different degrees of sky openness can be reasonably quantified.

[0028] For example, in practical applications, the rest area in front of a teaching building is measured and calculated to obtain corresponding quantitative values ​​of green visibility rate, sky visibility rate, and gaze point movement rate, etc. These quantitative indicators together constitute the basic data for visual environment assessment of the observation point.

[0029] Understandably, clustering methods play a crucial role in handling spatial grouping of gaze points. By performing spatial clustering analysis on the collected gaze point coordinate sequences, the distribution patterns of the observer's main gaze regions can be identified. During the clustering process, each gaze point is assigned to the nearest cluster center based on its spatial location. The average distance between cluster centers reflects the spatial dispersion of the gaze points. Higher dispersion indicates a more dispersed distribution of the observer's attention, which is beneficial for eye muscle relaxation and accommodation.

[0030] S102. Identify the initial degree of imbalance based on the quantified values ​​of green visibility rate and sky visibility, extract the fixation point concentration by recording the fixation point spatial coordinates, and extract the pupil change amplitude by recording the pupil diameter change.

[0031] The difference between the quantified green visibility rate and the quantified sky visibility rate is calculated. When the difference exceeds the upper limit of a preset imbalance threshold, it is marked as high imbalance; when the difference is within the preset imbalance threshold range, it is marked as slight imbalance. The difference is divided by the upper limit of the imbalance threshold to obtain a normalized initial imbalance degree value. Horizontal and vertical angle data for each sampling moment are extracted from the fixation point spatial coordinate record. The spatial distance between adjacent sampling points is calculated. When the spatial distance between more than 10 consecutive sampling points is less than 5 degrees, these sampling points are determined to belong to the same focusing area. The proportion of fixation points within the focusing area to the total number of fixation points is counted as the fixation point concentration. The pupil data processing method is adjusted according to the product of the fixation point concentration and the initial imbalance degree value. When the product exceeds a preset threshold, the diameter difference sequence between adjacent moments is extracted from the pupil diameter change record. The difference between the maximum and minimum values ​​of the difference sequence is calculated as the pupil change amplitude.

[0032] In one implementation, the initial degree of imbalance is identified based on the numerical relationship between the quantified value of green visibility and the quantified value of sky visibility.

[0033] Specifically, the arithmetic difference between the two values ​​reflects the equilibrium state of the landscape configuration. When the quantified value of green view rate is 0.8 and the quantified value of sky visibility is only 0.2, the difference reaches 0.6, indicating that the vegetation is excessively dense and the sky is severely obstructed, resulting in a highly unbalanced environment. The preset upper limit of the imbalance threshold is 0.5. When the difference exceeds this upper limit, the system automatically marks it as highly unbalanced; when the difference is within the preset range of 0.2 to 0.5, it is marked as slightly unbalanced. By dividing the actual difference by the upper limit of the imbalance threshold of 0.5, a normalized initial imbalance value is obtained, ranging from 0 to 1, to facilitate subsequent quantification.

[0034] It's important to note that the extraction of fixation concentration relies on spatial analysis of the gaze trajectory. From the time-series data recorded by the eye-tracking device, the horizontal and vertical angular coordinates of each sampling moment are extracted, and the distance between adjacent sampling points in angular space is calculated using the Euclidean distance formula. When the spatial distance between more than 10 consecutive sampling points is less than a preset distance threshold of 5 degrees, these sampling points are determined to belong to the same focal region, indicating that the observer's gaze remains continuously within that region. The total number of fixations within all focal regions is counted and divided by the total number of fixations throughout the entire observation period to obtain the fixation concentration value. A higher concentration indicates that the observer's gaze is more easily drawn to a specific area, lacking free movement of the gaze.

[0035] Preferably, the processing method for pupil data is dynamically adjusted according to the environmental imbalance state. By calculating the product of fixation concentration and the initial imbalance level, when this product exceeds a preset threshold of 0.4, it indicates that environmental imbalance and visual behavioral concentration coexist, and at this time, it is necessary to focus on subtle changes in the pupil. The diameter difference between every two adjacent moments is extracted from the pupil diameter change records to form a difference sequence, which reflects the dynamic adjustment process of the pupil. The difference between the maximum and minimum values ​​in the difference sequence is calculated to obtain the pupil change amplitude. This amplitude value can reflect the adjustment range of the observer's autonomic nervous system under a specific environmental configuration, providing a physiological basis for subsequent visual health assessment.

[0036] For example, in the rest area outside the campus library, when the green visibility quantification value was 0.7 and the sky visibility quantification value was 0.25, the difference was 0.45, and the initial imbalance after normalization was 0.9. At the same time, the fixation point concentration reached 0.6, and the product of the two was 0.54, exceeding the threshold and triggering fine extraction of pupil change amplitude. The final measured pupil change amplitude was 0.8 mm. These data collectively reflect the actual impact of the landscape configuration in this area on visual restoration.

[0037] S103. Analyze fixation concentration, identify the inhibitory effect of pupil change amplitude when fixation concentration is high and the compensatory effect of fixation dispersion when fixation concentration is low, and comprehensively obtain the degree of adjustment imbalance.

[0038] The suppression state is determined by comparing the fixation concentration value with a preset concentration threshold. If the fixation concentration exceeds the preset threshold, the ratio of the pupil change amplitude to the median of a preset normal amplitude range is calculated. When the ratio is lower than the preset suppression threshold, the suppression coefficient is obtained by subtracting the ratio from the preset suppression threshold. The suppression correction value is obtained by multiplying the suppression coefficient by the initial imbalance degree. When the fixation concentration is lower than the preset concentration threshold, the geometric center of all fixation coordinates is calculated from the fixation spatial coordinate sequence. The standard deviation of the distance from each fixation point to the geometric center is calculated as the fixation dispersion. The compensation coefficient is determined based on the ratio of the fixation dispersion to the preset standard dispersion. The compensation correction value is obtained by multiplying the compensation coefficient by the initial imbalance degree. The corresponding correction value is selected according to the fixation concentration level. When the fixation concentration is higher than the threshold, the suppression correction value is used; when the fixation concentration is lower than the threshold, the compensation correction value is used. The selected correction value and the initial imbalance degree are weighted and summed according to a preset weight ratio to obtain the adjustment imbalance degree.

[0039] In one implementation, the inhibition effect is identified based on the negative correlation between fixation concentration and pupillary variation. When the observer's fixation is highly concentrated in a specific area, the eye muscles maintain a fixed accommodative state, limiting the range of spontaneous pupillary accommodation.

[0040] Specifically, the preset concentration threshold is set to 0.6. When the fixation concentration exceeds this threshold, the system determines that an inhibition state exists. At this time, the pupil change amplitude value is extracted and compared with the preset normal amplitude range median of 1.5 mm. If the actual pupil change amplitude is only 0.9 mm, the ratio is 0.6. When this ratio is lower than the preset inhibition threshold of 0.7, 0.6 is subtracted from 1 to obtain an inhibition coefficient of 0.4. This coefficient is multiplied by the initial imbalance level of 0.8 to obtain an inhibition correction value of 0.32.

[0041] It should be noted that the compensation mechanism operates under conditions of gaze dispersion. When the gaze concentration is below the threshold of 0.6, the observer's gaze moves freely at different distances and in different directions. This diversity of visual behavior can partially compensate for the imbalance caused by environmental configuration. The geometric center is determined by calculating the arithmetic mean of the coordinates of all gaze points. The Euclidean distance from each gaze point to this center forms a distance sequence, and the standard deviation of this sequence is calculated to obtain the gaze dispersion. The preset standard dispersion is 15 degrees. When the actual dispersion is 22.5 degrees, the ratio of the two is 1.5. This ratio is used as the compensation coefficient, which is multiplied by the initial imbalance level of 0.8 to obtain a compensation correction value of 1.2.

[0042] Preferably, the correction value is dynamically determined based on the actual state of fixation concentration. In a high concentration state, the suppression correction value reflects the negative impact of limited physiological accommodation on visual recovery; in a low concentration state, the compensation correction value reflects the positive effect of visual behavioral diversity. The preset weight ratio is 0.3 to 0.7, meaning the selected correction value accounts for 30% of the weight, and the initial imbalance degree accounts for 70%. Specifically, the adjustment imbalance degree A equals 0.3 multiplied by the selected correction value B plus 0.7 multiplied by the initial imbalance degree C, i.e., A = 0.3 × B + 0.7 × C.

[0043] For example, in the corridor area between teaching buildings, the fixation concentration was measured at 0.75, exceeding the threshold. The ratio of pupillary change amplitude to the median of the normal range was 0.5, below the inhibition threshold. The inhibition coefficient was calculated to be 0.5, and the inhibition correction value was 0.45, which is the initial imbalance level of 0.9 multiplied by 0.5. Adding 0.45 and 0.9 with weights of 0.3 and 0.7 respectively, we obtain an adjusted imbalance level of 0.765. This value more accurately reflects the actual imbalance state under the combined effects of environmental configuration and physiological response.

[0044] In one possible implementation, the system monitors changes in gaze concentration in real time and dynamically switches between suppression and compensation mechanisms to achieve accurate assessment of the degree of imbalance.

[0045] S104. Determine the weights of green visibility and sky visibility based on the initial degree of imbalance and the adjusted degree of imbalance. Integrate the quantitative values ​​of green visibility and sky visibility based on the weights to obtain a comprehensive visual restoration score.

[0046] A weight adjustment factor is obtained by dividing the difference between the initial imbalance degree and the adjusted imbalance degree by the initial imbalance degree. When the adjustment factor exceeds a preset threshold, the green visibility weight is set to the adjustment factor, and the sky visibility weight is set to one minus the green visibility weight. The green visibility weight is multiplied by the quantified green visibility value, and the sky visibility weight is multiplied by the quantified sky visibility value. The two products are then added together to obtain the comprehensive visual restoration score.

[0047] In one implementation, the weighting adjustment factor is determined by calculating the relative change in the degree of imbalance.

[0048] Specifically, the difference between the initial degree of imbalance (denoted as I) and the adjusted degree of imbalance (denoted as A) is divided by the initial degree of imbalance, i.e., (IA) / I. The resulting ratio reflects the degree of improvement or deterioration of the imbalance state after physiological response regulation. A positive value indicates improvement, and a negative value indicates deterioration (e.g., when A is greater than I). Negative values ​​can be used to reduce the weight to emphasize deterioration. When I is 0.8 and A is 0.6, the difference is 0.2, and dividing by 0.8 yields a regulation factor of 0.25.

[0049] It should be noted that the weight allocation adopts a complementary mechanism. When the adjustment factor exceeds the preset threshold of 0.2, it indicates that the environmental configuration needs to be rebalanced. At this time, the green visibility rate weight is directly set to the adjustment factor value of 0.25, and the sky visibility weight is 1 minus 0.25, which is 0.75. This complementary allocation ensures that the sum of the weights of the two indicators is always 1, maintaining the normalization characteristics of the assessment.

[0050] Preferably, the comprehensive visual restoration score is obtained through weighted summation. Assuming a campus landscape assessment, the quantified value of green view rate is 0.65, and the quantified value of sky visibility is 0.52. Multiplying the green view rate weight of 0.25 by 0.65 yields 0.1625, and multiplying the sky visibility weight of 0.75 by 0.52 yields 0.39. Adding these two together gives a comprehensive score of 0.5525. This score falls within the standard range of 0 to 1, directly reflecting the overall support of the campus landscape environment for visual health.

[0051] S105. Extract the peak interval from the time record of the comprehensive visual recovery score, fill the green visibility rate quantification value and sky visibility quantification value within the peak interval with numerical values, and identify the candidate green visibility rate range and candidate sky visibility range.

[0052] The time series of the comprehensive visual recovery score is divided into continuous segments according to a preset time interval. The average score within each segment is calculated. When the average score of segments continuously exceeding a preset number threshold all exceed a preset peak threshold, the start and end time points are recorded to obtain the time boundary of the peak interval. For the green visibility rate quantification value sequence and sky visibility quantification value sequence within the peak interval, the time position of missing data points is detected. For each missing point, linear interpolation is performed according to the time interval ratio based on the value of the preceding and following valid data points to obtain the interpolated value of the missing point, thus completing the numerical filling of the sequence. The maximum and minimum values ​​are extracted from the filled green visibility rate quantification value sequence, and this interval is used as the candidate green visibility rate range. Similarly, the maximum and minimum values ​​are extracted from the filled sky visibility quantification value sequence, and this interval is used as the candidate sky visibility range.

[0053] In one implementation, the peak interval is extracted based on the time-series features of the comprehensive visual recovery score.

[0054] Specifically, the scoring is collected in time-series format, with one score value generated for each sampling moment. The observation time of one day is divided into 48 continuous segments of 30 minutes each, with each segment containing score data from multiple sampling moments. The arithmetic mean of all scores within each segment is calculated to form a segment average sequence. When the average of five or more consecutive segments exceeds a preset peak threshold of 0.7, the system determines that the continuous interval is a peak interval. The start time of the first segment exceeding the threshold is recorded as the start time of the peak interval, and the end time of the last segment exceeding the threshold is recorded as the end time of the peak interval, thus defining the complete time boundary.

[0055] It should be noted that during actual data collection, due to equipment failure, network interruption, or the observer's temporary absence, the green visibility rate and sky visibility data sequences within the peak range may be missing. Missing data is detected through timestamp continuity checks; when the time interval between two adjacent data points exceeds 1.5 times the normal sampling interval, a data gap is identified. For each detected missing point, the system extracts the value and timestamp of the preceding and following valid data points, and calculates the interpolated value of the missing point using a linear interpolation formula. In the interpolation calculation, the missing point value equals the preceding value plus the product of the difference between the preceding and following values ​​and the time interval ratio. This method assumes that the data exhibits a linear trend over a short period.

[0056] Preferably, the candidate range is identified by performing extreme value analysis on the padded complete data sequence. Within the peak interval, the numerically padded green visibility quantification value sequence contains all valid observations and interpolated values ​​within that time period. By traversing this sequence to extract the maximum and minimum values, these two extreme values ​​constitute the upper and lower boundaries of the candidate green visibility range. Similarly, the maximum and minimum values ​​of the sky visibility quantification value sequence determine the candidate sky visibility range.

[0057] For example, within the peak period from 10:00 AM to 11:30 AM, the green visibility quantification value sequence, after padding, contains 180 data points, with a maximum value of 0.82 and a minimum value of 0.58. Therefore, the candidate green visibility range is 0.58 to 0.82. The maximum value of the sky visibility quantification value sequence is 0.65, and the minimum value is 0.35. The candidate sky visibility range is 0.35 to 0.65.

[0058] In one possible implementation, these candidate ranges provide a data foundation for subsequent cluster analysis, enabling the system to find the optimal configuration combination within the range of actually observed values.

[0059] S106. K-means clustering analysis is used for the candidate green visibility range and the candidate sky visibility range to extract the cluster centers and determine the optimal green visibility boundary and the optimal sky visibility boundary. The target green visibility range and sky visibility range are obtained by limiting the boundaries.

[0060] The candidate green visibility range and candidate sky visibility range are each uniformly divided into a predetermined number of points. Green visibility sample values ​​are extracted from the candidate green visibility range, and sky visibility sample values ​​are extracted from the candidate sky visibility range. The green visibility sample values ​​and sky visibility sample values ​​at corresponding locations are combined into two-dimensional data points to construct a feature space dataset. The k-means clustering algorithm is used to process the feature space dataset. The number of clusters is set according to the data distribution characteristics. Initial cluster centers are randomly selected, and the Euclidean distance from each data point to each cluster center is calculated. Data points are assigned to the nearest cluster, and the mean of the coordinates of all data points in each cluster is recalculated based on the assignment results as the new cluster center. The distance calculation and center update process is repeated. Iteration stops when the change in the cluster center position between two adjacent iterations is less than a predetermined convergence threshold. Clusters containing more than a predetermined proportion of data points are identified from the converged clustering results as valid clusters. Extract the maximum and minimum values ​​of all data points in the effective cluster in the green visibility dimension as the upper and lower boundaries of the target green visibility interval, and extract the maximum and minimum values ​​of all data points in the effective cluster in the sky visibility dimension as the upper and lower boundaries of the target sky visibility interval.

[0061] In one implementation, the feature space is constructed through systematic data sampling and combination.

[0062] Specifically, the candidate green visibility range and the candidate sky visibility range are each uniformly divided into 100 sampling points, forming a dense sampling grid. For the candidate green visibility range of 0.58 to 0.82, sample values ​​are extracted at intervals of 0.0024, resulting in a green visibility sample sequence containing values ​​from 0.58, 0.5824, 0.5848 up to 0.82. Similarly, the candidate sky visibility range of 0.35 to 0.65 is sampled at intervals of 0.003, forming a sky visibility sample sequence. The i-th green visibility sample value is combined with the i-th sky visibility sample value to form a two-dimensional coordinate point. All coordinate points together constitute the feature space dataset, which reflects the possible combined distribution of the two key visual parameters.

[0063] It should be noted that the k-means clustering algorithm finds the natural grouping of data through iterative optimization. The algorithm sets the number of clusters k to 5 to capture the main distribution patterns and avoid over-segmentation. Initial centers are selected using the k-means++ method to ensure a large distance between centers, promoting fast convergence. Each data point is assigned to the cluster center with the smallest Euclidean distance, where the distance d is calculated as sqrt((x1-x2)). 2 +(y1-y2) 2 ), where x is the green visibility rate and y is the sky visibility.

[0064] Preferably, the cluster center update process reflects the adaptive nature of the algorithm. In each iteration, after all data points have been clustered, the center position of each cluster is recalculated. The new cluster center is the arithmetic mean of the coordinates of all data points within that cluster. In the green visibility dimension, the green visibility value of the new center is equal to the sum of the green visibility values ​​of all points in that cluster divided by the number of points; the same calculation method is used in the sky visibility dimension. This mean-based update mechanism allows the cluster centers to gradually move towards data-dense areas, eventually stabilizing at positions that represent the characteristics of the cluster.

[0065] For example, the algorithm's convergence is determined using the change in cluster center location as the criterion. In two consecutive iterations, the distance each cluster center moves in the feature space is calculated. The algorithm is considered converged when the distances of all cluster centers are less than a convergence threshold of 0.001. Typically, the algorithm reaches convergence within 10 to 20 iterations. The converged clustering results exhibit clear data grouping characteristics, with each cluster representing a typical combination of green visibility and sky visibility.

[0066] In one possible implementation, the identification of effective clusters is based on data point density analysis. The number of data points contained in each cluster is counted, and its proportion of the total number of data points is calculated. When a cluster contains more than 20% of the total number of data points, it is identified as an effective cluster. This threshold is determined by analyzing data distribution characteristics using the elbow method, indicating that this parameter combination region occurs frequently in actual observations and is representative. Typically, one or two effective clusters are identified, reflecting the most frequently occurring visual environment configurations during peak periods. Furthermore, the boundaries of the target interval are determined by analyzing the centers of the effective clusters. For each effective cluster, the value of its cluster center in the green visibility dimension is extracted as a representative value for the target green visibility interval. The same method is applied to the sky visibility dimension, extracting the corresponding value of the cluster center to form the target sky visibility interval. This boundary determination method based on actual data distribution ensures that the target interval covers the main high-quality configuration range.

[0067] For example, cluster analysis of a campus rest area showed that the largest effective cluster contained 45% of the data points. Within this cluster, green view rate ranged from 0.62 to 0.75, and sky visibility ranged from 0.40 to 0.55. Therefore, the target green view rate range was determined to be 0.62 to 0.75, and the target sky visibility range was determined to be 0.40 to 0.55. These ranges represent the typical environmental configuration range for this area during periods of good visual recovery.

[0068] Understandably, the target interval obtained through k-means clustering analysis is not a simple statistical range, but a high-quality configuration concentration area identified after data mining, providing a quantitative basis for the precise optimization of the campus landscape.

[0069] S107. Based on the target green visibility range and the target sky visibility range, the dispersion of the gaze point and the amplitude of pupil change collected subsequently are evaluated in real time to identify the matching degree. When the matching degree reaches a high matching level, the current landscape green visibility and sky visibility combination is output as the final optimization result.

[0070] The system acquires the fixation point dispersion and pupil change amplitude values ​​in real time to determine whether the current landscape green visibility rate falls within the target green visibility rate range and whether the sky visibility falls within the target sky visibility range. If both are within their respective ranges, the environment matching flag is set to true; otherwise, it is set to false. When the environment matching flag is true, the system checks whether the fixation point dispersion exceeds a preset dispersion threshold and whether the pupil change amplitude exceeds a preset amplitude threshold. If both physiological indicators meet the threshold conditions, the ratio of the fixation point dispersion to the preset standard dispersion value and the ratio of the pupil change amplitude to the preset standard amplitude value are calculated. The two ratios are added together and divided by two to obtain the matching degree value. When the matching degree value exceeds the preset high matching threshold, the current landscape green visibility rate and sky visibility values ​​are extracted, and the combination of the green visibility rate and sky visibility values ​​is output as the final optimization result.

[0071] In one implementation, the real-time evaluation process is achieved through continuous monitoring and dynamic judgment. The system integrates an eye-tracking device to track eye movements and a pupil detection device to monitor pupil diameter. Every 5 seconds, it collects the current environment's green visibility rate and sky visibility values, while simultaneously obtaining the fixation point dispersion and pupil change amplitude from these devices. The current landscape green visibility rate value is compared with the target green visibility rate interval; if the value falls within the interval, the green visibility rate condition is met. Similarly, the sky visibility value is compared with the target sky visibility interval to determine if the sky visibility condition is met. Only when both conditions are simultaneously met is the environment matching flag set to true, indicating that the current environment configuration meets the optimization target range.

[0072] It should be noted that the thresholds for physiological indicators are determined based on statistical analysis of a large amount of observational data. The preset minimum threshold for dispersion is typically set at 20 degrees, indicating sufficient spatial dispersion of the fixation point, considered the minimum acceptable threshold; the preset minimum threshold for amplitude is set at 1.0 mm, reflecting active pupillary accommodation, also considered the minimum acceptable threshold. When the environmental match flag is true and both physiological indicators exceed their respective thresholds, the system proceeds to the matching degree calculation stage. The preset ideal dispersion value is 25 degrees, representing the ideal degree of fixation point dispersion; the preset ideal amplitude value is 1.5 mm, representing the ideal pupillary accommodation range. Normalization is performed using preset standard values ​​to ensure that the ratio of fixation point dispersion and the ratio of pupillary amplitude changes are dimensionless values, allowing for arithmetic operations.

[0073] Preferably, the matching degree value is obtained by calculating a normalized ratio. The fixation point dispersion ratio is obtained by dividing the actual measured value by a preset standard dispersion value to obtain a dimensionless ratio, and the pupil change amplitude ratio is obtained by dividing the actual measured value by a preset standard amplitude value to obtain a dimensionless ratio. Both dimensionless ratios are normalized to the zero-to-one interval and then arithmetically averaged. Fixation point dispersion ratio r1 = min(D actual / D standard ,1) Reflects the degree to which the current fixation behavior is close to the ideal state; the ratio of pupil change amplitude r2 = min(P actual / P standard 1) This reflects the level of activity in physiological regulation. Adding the two ratios and dividing by 2 yields a matching degree value between 0 and 1. This averaging method balances the contributions of both visual behavior and physiological response, avoiding the excessive influence of a single indicator.

[0074] For example, when the matching degree reaches a high matching threshold of 0.8, it indicates that the current landscape configuration not only conforms to the target range in terms of parameters but also elicits a good physiological response. At this point, the system extracts the current green view rate value of 0.68 and the sky visibility value of 0.48, and outputs this combination as the final optimization result. This result represents an ideal landscape configuration validated in real-world application scenarios, which can effectively alleviate visual fatigue.

[0075] In one possible implementation, the final optimization results can guide the actual transformation of the campus landscape by adjusting vegetation density and pruning strategies to achieve the target green visibility rate and by controlling the height of tree canopies to achieve the target sky visibility.

[0076] This invention provides a campus landscape green view rate and visual health restoration assessment system, mainly comprising: The data acquisition and quantification module is used to extract the proportion of green pixels and the ratio of sky area from the field of view elevation angle range through the panoramic image acquisition module. It simultaneously acquires the spatial coordinates of the gaze point, the gaze point dispersion, and the pupil diameter change record. The green pixel proportion is normalized to obtain the green visibility rate quantification value, and the sky area ratio is threshold filtered to obtain the sky visibility quantification value. The imbalance identification and feature extraction module is used to identify the initial degree of imbalance based on the green visibility rate quantification value and the sky visibility quantification value, extract the fixation point concentration through the fixation point spatial coordinate record, and extract the pupil change amplitude through the pupil diameter change record. The adjustment analysis module is used to analyze the fixation point concentration, identify the inhibitory effect of pupil change amplitude when fixation point concentration is high and the compensatory effect of fixation point dispersion when fixation point concentration is low, and comprehensively obtain the degree of adjustment imbalance. The weight determination and scoring module is used to determine the green visibility weight and sky visibility weight based on the initial imbalance degree and the adjusted imbalance degree, and to integrate the green visibility quantification value and the sky visibility quantification value to obtain a comprehensive visual restoration score. The peak interval extraction module is used to extract the peak interval from the time record of the comprehensive visual recovery score, fill the green vision rate quantification value and sky visibility quantification value within the peak interval with numerical values, and identify candidate green vision rate range and candidate sky visibility range. The clustering analysis module is used to perform k-means clustering analysis on the candidate green visibility range and the candidate sky visibility range to determine the target green visibility range and the target sky visibility range; The real-time evaluation and output module is used to evaluate the dispersion of the gaze point and the amplitude of pupil change in real time based on the target green visibility range and the target sky visibility range, identify the matching degree, and determine the current landscape green visibility and sky visibility combination when the matching degree reaches a high matching level.

[0077] The above description is merely a preferred embodiment of one or more embodiments of this specification and is not intended to limit the scope of one or more embodiments of this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments of this specification should be included within the protection scope of one or more embodiments of this specification.

Claims

1. A method for assessing the green view rate and visual health restoration of campus landscapes, characterized in that, include: The panoramic image acquisition module extracts the proportion of green pixels and the ratio of sky area from the field of view elevation angle. Simultaneously, it collects the spatial coordinates of the gaze point, the dispersion of the gaze point, and the pupil diameter change. The green pixel proportion is normalized to obtain the green visibility rate quantification value, and the sky area ratio is threshold filtered to obtain the sky visibility quantification value. The initial degree of imbalance is identified based on the green visibility rate quantification value and the sky visibility quantification value; the fixation point concentration is extracted by the fixation point spatial coordinate record; and the pupil change amplitude is extracted by the pupil diameter change record. The fixation point concentration is analyzed to identify the inhibitory effect of high fixation point concentration on pupil change amplitude and the compensatory effect of low fixation point concentration on fixation point dispersion, and the degree of adjustment imbalance is obtained by combining the results. The green visibility weight and sky visibility weight are determined based on the initial imbalance degree and the adjusted imbalance degree, and the comprehensive visual restoration score is obtained by integrating the quantified value of green visibility and the quantified value of sky visibility. Peak intervals are extracted from the time records of the comprehensive visual recovery score, and the green visibility rate quantification value and sky visibility quantification value within the peak interval are numerically filled to identify candidate green visibility rate ranges and candidate sky visibility ranges. The candidate green visibility range and the candidate sky visibility range are subjected to k-means clustering analysis to determine the target green visibility range and the target sky visibility range; Based on the target green visibility range and the target sky visibility range, the dispersion of the gaze point and the amplitude of pupil change collected subsequently are evaluated in real time to identify the matching degree. When the matching degree reaches a high matching level, the combination of the target current landscape green visibility and sky visibility is determined.

2. The method for assessing the green view rate and visual health restoration of campus landscapes according to claim 1, characterized in that, The process involves extracting the proportion of green pixels and the ratio of sky area from the elevation angle range of the field of view using a panoramic image acquisition module. Simultaneously, it acquires records of the spatial coordinates of the gaze point, the gaze point dispersion, and pupil diameter changes. The green pixel proportion is normalized to obtain a quantified value of green visibility, and the sky area ratio is threshold-filtered to obtain a quantified value of sky visibility. This includes: The panoramic image acquisition module continuously captures environmental image data within the vertical field of view from negative to positive elevation angle. The HSV color space conversion is used to identify green vegetation areas, and the Otsu threshold segmentation method is used to extract the sky area. The ratio of the number of pixels in the green vegetation area to the total number of pixels is recorded as the basic value of the green proportion, and the ratio of the sky area to the total area of ​​the upper half of the field of view is recorded as the basic value of the sky proportion. The total length of the gaze trajectory is calculated by recording the gaze point angle coordinate sequence using an eye-tracking device, and the pupil major axis diameter value is extracted and recorded as the basic value of the pupil change amplitude using an infrared camera. The green proportion base value is normalized to obtain the green visibility rate quantification value, the sky area ratio is threshold filtered to obtain the sky visibility quantification value, and the gaze point dispersion is calculated based on the total length of the gaze trajectory.

3. The method for assessing the green view rate and visual health restoration of campus landscapes according to claim 1, characterized in that, The initial degree of imbalance is identified based on the quantified green visibility rate and the quantified sky visibility rate. Fixation point concentration is extracted using the fixation point spatial coordinate record, and pupil change amplitude is extracted using the pupil diameter change record, including: The difference between the quantified value of green visibility and the quantified value of sky visibility is calculated, and the difference is divided by the upper limit of the imbalance threshold to obtain the initial imbalance degree. The spatial distance between adjacent sampling points is calculated from the spatial coordinate record of the fixation point, and the proportion of the number of fixation points in the focal area to the total number of fixation points is used as the fixation point concentration. The pupil data processing method is adjusted according to the fixation point concentration, and the diameter difference sequence is extracted from the pupil diameter change record to calculate the pupil change amplitude.

4. The method for assessing the green view rate and visual health restoration of campus landscapes according to claim 1, characterized in that, The analysis of fixation concentration identifies the inhibitory effect on pupillary variation when fixation concentration is high and the compensatory effect on fixation dispersion when fixation concentration is low, and comprehensively determines the degree of adjustment imbalance, including: The inhibition state is determined by comparing the fixation point concentration with a preset concentration threshold. If the fixation point concentration exceeds the preset threshold, the ratio of the pupil change amplitude to the median of the normal amplitude range is calculated. The inhibition coefficient is obtained by subtracting the ratio from one. The inhibition correction value is obtained by multiplying the inhibition coefficient by the initial imbalance degree. When the fixation point concentration is lower than the preset concentration threshold, the geometric center of all fixation point coordinates is calculated from the fixation point spatial coordinate sequence, and the standard deviation of the distance from each fixation point to the geometric center is calculated as the fixation point dispersion. The compensation coefficient is determined according to the ratio of the fixation point dispersion to the preset standard dispersion, and the compensation correction value is obtained by multiplying the compensation coefficient by the initial imbalance degree. Based on the state of fixation concentration, the suppression correction value or the compensation correction value is selected, and the selected correction value and the initial imbalance degree are weighted and summed according to a preset weight ratio to obtain the adjusted imbalance degree.

5. The method for assessing the green view rate and visual health restoration of campus landscapes according to claim 1, characterized in that, The green visibility weight and sky visibility weight are determined based on the initial imbalance degree and the adjusted imbalance degree. The quantified values ​​of green visibility and sky visibility are then integrated to obtain a comprehensive visual restoration score, including: The weight adjustment factor is obtained by dividing the difference between the initial imbalance degree and the adjusted imbalance degree by the initial imbalance degree. The weight adjustment factor is set as the green visibility weight, and the sky visibility weight is set as one minus the green visibility weight. The comprehensive visual restoration score is obtained by multiplying the green visibility weight by the green visibility quantification value and the sky visibility weight by the sky visibility quantification value, and then adding the two product results.

6. The method for assessing the green view rate and visual health restoration of campus landscapes according to claim 1, characterized in that, Peak intervals are extracted from the time records of the comprehensive visual restoration score. The quantified values ​​of green visibility and sky visibility within these peak intervals are numerically filled to identify candidate green visibility ranges and candidate sky visibility ranges, including: Divide the time record of the comprehensive visual recovery score into continuous segments, calculate the average value of each segment, and record the start and end times of the average value of segments that continuously exceed a preset number threshold and the peak threshold as the peak interval. Linear interpolation is performed on the quantized green visibility rate sequence and the quantized sky visibility rate sequence within the peak interval to fill in the numerical values. The maximum and minimum values ​​are extracted from the padded sequence as candidate green visibility ranges and candidate sky visibility ranges.

7. The method for assessing the green view rate and visual health restoration of campus landscapes according to claim 1, characterized in that, K-means clustering analysis was performed on the candidate green visibility range and the candidate sky visibility range to determine the target green visibility interval and the target sky visibility interval, including: The candidate green visibility range and the candidate sky visibility range are evenly divided into sampling points, and green visibility sample values ​​and sky visibility sample values ​​are extracted and combined into two-dimensional data points to construct a feature space dataset. The feature space dataset is processed using the k-means clustering algorithm. Initial cluster centers are randomly selected, and the Euclidean distance from each data point to each cluster center is calculated to allocate data points. The mean within each cluster is recalculated as the new cluster center. Repeat the distance calculation and center update process until the change in the center of the adjacent iteration is less than the convergence threshold, and identify the effective clusters containing more than the preset proportion threshold from the clustering results after convergence. The maximum and minimum values ​​of data points in the effective clusters are extracted as the upper and lower boundaries of the target green visibility interval in the green visibility dimension, and the maximum and minimum values ​​in the sky visibility dimension are extracted as the upper and lower boundaries of the target sky visibility interval.

8. The method for assessing the green view rate and visual health restoration of campus landscapes according to claim 1, characterized in that, Based on the target green visibility range and the target sky visibility range, the dispersion of the gaze point and the amplitude of pupil change collected subsequently are evaluated in real time to identify the matching degree. When the matching degree reaches a high matching level, the current landscape green visibility and sky visibility combination is determined, including: Real-time acquisition of subsequent fixation point dispersion values ​​and pupil change amplitude values; determination of current landscape green visibility rate within the target green visibility rate range and sky visibility within the target sky visibility range; setting of environmental matching markers. When the environment matching flag is true, check that the gaze point dispersion and the pupil change amplitude both meet the threshold conditions, and calculate the average of the two ratios to obtain the matching degree value. When the matching degree value exceeds the high matching threshold, the current green visibility value and sky visibility value are extracted as the final optimization result output.

9. A campus landscape green view rate and visual health restoration assessment system, characterized in that, The system includes: The data acquisition and quantification module is used to extract the proportion of green pixels and the ratio of sky area from the field of view elevation angle range through the panoramic image acquisition module. It simultaneously acquires the spatial coordinates of the gaze point, the gaze point dispersion, and the pupil diameter change record. The green pixel proportion is normalized to obtain the green visibility rate quantification value, and the sky area ratio is threshold filtered to obtain the sky visibility quantification value. The imbalance identification and feature extraction module is used to identify the initial degree of imbalance based on the green visibility rate quantification value and the sky visibility quantification value, extract the fixation point concentration through the fixation point spatial coordinate record, and extract the pupil change amplitude through the pupil diameter change record. The adjustment analysis module is used to analyze the fixation point concentration, identify the inhibitory effect of pupil change amplitude when fixation point concentration is high and the compensatory effect of fixation point dispersion when fixation point concentration is low, and comprehensively obtain the degree of adjustment imbalance. The weight determination and scoring module is used to determine the green visibility weight and sky visibility weight based on the initial imbalance degree and the adjusted imbalance degree, and to integrate the green visibility quantification value and the sky visibility quantification value to obtain a comprehensive visual restoration score. The peak interval extraction module is used to extract the peak interval from the time record of the comprehensive visual recovery score, fill the green vision rate quantification value and sky visibility quantification value within the peak interval with numerical values, and identify candidate green vision rate range and candidate sky visibility range. The clustering analysis module is used to perform k-means clustering analysis on the candidate green visibility range and the candidate sky visibility range to determine the target green visibility range and the target sky visibility range; The real-time evaluation and output module is used to evaluate the dispersion of the gaze point and the amplitude of pupil change in real time based on the target green visibility range and the target sky visibility range, identify the matching degree, and determine the current landscape green visibility and sky visibility combination when the matching degree reaches a high matching level.

Images