Spatial determination device

Abstract

To easily determine a type of influence given to a person by an object space where a plant is arranged.SOLUTION: A space determination device 4 comprises: a feature amount extraction unit 7 which extracts a feature amount that changes according to the arrangement state of a plant included in an image of an object space from the image; and a classification unit 8 which classifies the object space into any of a plurality of space types. The plurality of space types include at least the first space type of reducing the stress of a person, the second space type of improving the concentration of the person and the third space type of improving the vitality of the person. The feature amount extraction unit 7 extracts at least a green coverage rate of the image of the object space, a fractal dimension of the image of the object space and a fluctuation value of the image of the object space from the image of the object space as a feature amount. The classification unit 8 classifies the object space by classifying the image of the object space into an image group in any of the plurality of space types on the basis of the extracted feature amount.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a space determination device.

Background Art

[0002] Since Edward Osborne Wilson proposed the Biophilia hypothesis in 1984, so-called biophilic design incorporating this hypothesis has been applied to various buildings and living spaces. In recent years, it has been confirmed that spaces to which biophilic design has been applied (also referred to as "biophilia spaces") have a beneficial effect on people, such as stress reduction, creativity improvement, or work efficiency improvement.

[0003] Methods for determining the influence of the naturalness of a space on people have also been proposed so far. For example, Patent Document 1 discloses a method of acquiring physiological response information when in a space within a forest and physiological response information when in a space in an urban area, and determining whether the space within the forest is a space suitable for forest bathing based on the difference between the respective physiological response information.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, the method disclosed in Patent Document 1 only determines whether a space within a natural environment is a space suitable for forest bathing. That is, with the method disclosed in Patent Document 1, even if plants are variously arranged in a target space to construct a biophilia space, it is impossible to determine what type of influence the target space can be expected to have on people.

[0006] This invention has been made in view of the above, and aims to easily determine what type of effect a space in which plants are placed has on people. [Means for solving the problem]

[0007] To solve the above problems, the spatial determination device of the present invention is a spatial determination device for determining the effect that a target space in which plants are arranged has on a person, and comprises: a feature extraction unit that extracts feature quantities from an image of the target space that change according to the arrangement of plants included in the image of the target space; and a classification unit that classifies the target space into one of a plurality of spatial types that have different effects on a person, wherein the plurality of spatial types include at least a first spatial type that reduces stress on a person, a second spatial type that improves concentration on a person, and a third spatial type that improves vitality on a person, and the feature extraction unit extracts at least the green view ratio of the image of the target space, the fractal dimension of the image of the target space, and the fluctuation value of the image of the target space as feature quantities from the image of the target space, and the classification unit classifies the target space by classifying the image of the target space into one of the plurality of spatial types based on the extracted feature quantities.

[0008] The spatial determination device can categorize the effects of a target space where plants are placed on people into at least three types: stress reduction, improved concentration, and increased vitality. Based on these effects, it can classify the target space into three types: a first spatial type, a second spatial type, and a third spatial type. In this case, simply by acquiring an image of the target space and inputting it into the spatial determination device, the device can extract each of the above-mentioned features and classify the target space based on the extracted features. Therefore, even if the target space where plants are placed is unknown, the spatial determination device can easily determine what type of effect the target space will have on people.

[0009] In a more preferred embodiment, the image of the target space is an omnidirectional image obtained by photographing the area around a predetermined position within the target space in all directions.

[0010] In this configuration, the spatial determination device can classify a target space by considering the arrangement of plants as far as a person standing at a predetermined location within the target space can see. Therefore, the spatial determination device can more accurately reflect the impact that a target space with plants has on people in its classification. Thus, the spatial determination device can easily and accurately determine what type of impact a target space has on people.

[0011] In a more preferred embodiment, the classification unit classifies the target space by performing cluster analysis or principal component analysis on the image of the target space using the extracted features.

[0012] In this embodiment, the spatial determination device can classify the target space using relatively simple methods such as cluster analysis or principal component analysis. Therefore, the spatial determination device can more easily determine what type of impact the target space has on people.

[0013] In a more preferred embodiment, the spatial determination device further comprises an evaluation unit that evaluates the similarity or dissimilarity of the target space to the spatial type based on the results of the cluster analysis or principal component analysis.

[0014] In this configuration, the spatial determination device can quantitatively evaluate how closely the target space resembles each spatial type. Therefore, the spatial determination device can easily and quantitatively determine what type of impact the target space will have on people.

[0015] In a more preferred embodiment, the fluctuation value of the image in the target space is a 1 / f fluctuation value.

[0016] In this aspect, since the space determination device uses 1 / f fluctuation, which affects the comfort and relaxation effect on people, as a feature amount, the target space can be surely classified into each space type. Therefore, the space determination device can easily and surely determine what type of influence the target space has on people.

Advantages of the Invention

[0017] According to the present invention, it is possible to easily determine what type of influence the target space in which plants are arranged has on people.

Brief Description of the Drawings

[0018] [Figure 1] The figure which shows the structure of the space determination system provided with the space determination device of this embodiment. [Figure 2] The figure which shows Genki space. [Figure 3] The figure which shows the questionnaire result to the subject seated in each seat of Genki space. [Figure 4] Fig. 4(a) is a diagram showing an image of the Ten area of the Genki space, and Fig. 4(b) is a diagram showing an image of the Sen area of the Genki space. [Figure 5] Fig. 5(a) is a diagram showing an image of the Men area of the Genki space, and Fig. 5(b) is a diagram showing an image of the control area with plants. [Figure 6] The figure which shows the cluster analysis result of each image shown in Fig. 4(a)-Fig. 5(b). [Figure 7] The figure which shows the principal component analysis result of each image shown in Fig. 4(a)-Fig. 5(b).

Embodiments of the Invention

[0019] Hereinafter, embodiments of the present invention will be described with reference to the drawings. Regarding the configurations denoted by the same reference numerals in each embodiment, unless otherwise specified, they have the same functions in each embodiment, and the description thereof will be omitted.

[0020] [Configuration of Space Determination System] FIG. 1 is a diagram showing the configuration of a space determination system 1 including a space determination device 4 of the present embodiment.

[0021] The space determination system 1 is a system that determines the influence exerted by a target space in which plants are arranged on people. In particular, the space determination system 1 determines, based on an image of the target space, what kind of positive influence the target space can be expected to exert on people. The space determination system 1 is utilized in the design of the target space in order to construct the above-mentioned biophilic space. The space determination system 1 includes an imaging device 2, a display device 3, and a space determination device 4.

[0022] The imaging device 2 is composed of a camera that acquires an image of the target space. The image acquired by the imaging device 2 may be an omnidirectional image (360-degree image) acquired by photographing the surroundings of a predetermined position in all directions at the predetermined position within the target space. Alternatively, the image acquired by the imaging device 2 may be a normal image (front image) acquired by photographing in one or more directions from the predetermined position. The imaging device 2 of the present embodiment acquires an omnidirectional image. The imaging device 2 may be composed of an omnidirectional camera (360-degree camera) that acquires an omnidirectional image, or may be composed of a camera mounted on a general-purpose camera or a smartphone or the like.

[0023] The display device 3 is composed of a display that displays the image acquired by the imaging device 2 and the processing result of the space determination device 4.

[0024] The space determination device 4 is composed of a computer system that determines the influence exerted by the target space on people. The space determination device 4 includes a storage device 5 and an arithmetic processing device 6. The storage device 5 is composed of an SSD, an HDD, or the like. The storage device 5 stores various data and the like used in the processing of the arithmetic processing device 6. The arithmetic processing device 6 is composed of a CPU, a ROM, a RAM, and the like. The arithmetic processing device 6 realizes various functions of the space determination device 4 by the CPU executing a program stored in the ROM.

[0025] The arithmetic processing unit 6 includes a feature extraction unit 7, a classification unit 8, and an evaluation unit 9.

[0026] The feature extraction unit 7 extracts features from the image of the target space that change according to the arrangement of plants contained in the image. In particular, the feature extraction unit 7 extracts at least the green view ratio of the image of the target space, the fractal dimension of the image of the target space, and the fluctuation value of the image of the target space as features of the image of the target space. The feature extraction unit 7 includes a green view ratio analysis unit 71, a fractal analysis unit 72, and a fluctuation analysis unit 73.

[0027] The green view ratio analysis unit 71 analyzes the green view ratio (also called "green visual encirclement ratio") of the image of the target space. The green view ratio is an index that indicates the proportion of green (plants) in the field of view. Specifically, the green view ratio analysis unit 71 converts the image of the target space (e.g., an RGB image) acquired by the imaging device 2 into the HSV color space. Then, the green view ratio analysis unit 71 counts pixels from the entire image that fall within the range of hue 40 / 360 to 130 / 360, saturation 15 / 100 to 100 / 100, and brightness 10 / 100 to 100 / 100. The green view ratio analysis unit 71 uses the counted number of pixels as the number of green pixels. The green view ratio analysis unit 71 calculates the proportion of the number of green pixels to the total number of pixels in the image (4096 × 2048). In this way, the green view ratio analysis unit 71 can calculate the green view ratio of the image of the target space.

[0028] The fractal analysis unit 72 analyzes the fractal dimension of the image in the target space. A fractal figure is a figure that possesses self-similarity, where the shape of the details and the shape of the whole are geometrically similar. Fractal figures exist in large numbers in the natural environment, such as tree branches or snowflakes, and can have an effect on people such as a sense of nature and healing. The fractal dimension is an index that represents the complexity of a fractal figure. When the entire figure is composed of m miniatures of a similar shape that are reduced to 1 / n, the fractal dimension of this figure is D = log n m is defined as m.

[0029] Specifically, the fractal analysis unit 72 converts the image of the target space acquired by the imaging device 2 into grayscale. For example, the fractal analysis unit 72 converts the image of the target space into grayscale using the luminance of YCbCr, the lightness obtained from the HSV color space, the saturation obtained from the HSV color space, the lightness obtained from the HLS color space, the saturation obtained from the HLS color space, etc. The fractal analysis unit 72 analyzes the fractal dimension from the grayscale image using the stereo method or the FBM (Fractal Brown Motion) method, etc. Alternatively, the fractal analysis unit 72 may analyze the fractal dimension by performing scale dimension analysis or power spectrum analysis without converting the image of the target space into grayscale. Or, the fractal analysis unit 72 may further binarize the grayscale image and analyze the fractal dimension from the binarized image using the box count method. The binarization method is not particularly limited. For example, the fractal analysis unit 72 may perform binarization by fixing a threshold, binarization using various filters, or binarization after extracting image edges. The fractal analysis unit 72 may also combine these methods to analyze the fractal dimension. In this embodiment, the fractal analysis unit 72 analyzes the fractal dimension using the FBM method.

[0030] The fluctuation analysis unit 73 analyzes the fluctuation values ​​of the image in the target space. Image fluctuations include DFA (detrended fluctuation analysis) and 1 / f fluctuation. In particular, 1 / f fluctuation is a type of fluctuation in which the power spectrum is inversely proportional to the frequency, and it is a type of fluctuation in which regularity and irregularity are appropriately combined. 1 / f fluctuation can have an effect on people such as making them feel comfortable and relaxed. In this embodiment, the fluctuation analysis unit 73 analyzes the 1 / f fluctuation values ​​of the image in the target space. Specifically, the fluctuation analysis unit 73 converts the image in the target space to grayscale. The fluctuation analysis unit 73 treats the changes in density (changes in brightness) in the vertical and horizontal directions of the grayscale image as waves (fluctuations) and performs a Fourier transform on each. Then, the fluctuation analysis unit 73 calculates the logarithm of the power spectrum and the logarithm of the frequency from the Fourier transformed waves. The fluctuation analysis unit 73 calculates the slope of the line plotted on a two-dimensional map, which represents the logarithm of the power spectrum and the logarithm of the frequency, as the 1 / f fluctuation value.

[0031] The classification unit 8 classifies the target space into one of several spatial types that have different effects on people. The classification unit 8 classifies the target space by classifying images of the target space into one of several spatial type image groups based on the features extracted by the feature extraction unit 7. The multiple spatial types include at least a first spatial type that reduces stress on people, a second spatial type that improves concentration on people, and a third spatial type that improves vitality on people. Specifically, the classification unit 8 classifies the target space by performing cluster analysis or principal component analysis on images of the target space using the extracted features.

[0032] The classification unit 8 uses a feature vector for the image of the target space that includes at least the green view ratio of the image of the target space, the fractal dimension of the image of the target space, and the fluctuation value of the image of the target space. The classification unit 8 then classifies the images of the target space into one of several spatial types of image groups (clusters) by performing hierarchical cluster analysis on the multiple feature vectors. In this embodiment, the cluster analysis of the classification unit 8 uses the squared Euclidean distance as the distance between images used for classification, and the group average method is used as the method for measuring the distance between clusters. However, the cluster analysis of the classification unit 8 may use distances other than the squared Euclidean distance and measurement methods other than the group average method. For example, the mean Euclidean distance, Mahalanobis distance, correlation coefficient, cosine coefficient, Brecurtis coefficient, or Canberra distance coefficient may be used as the distance between images used for classification. The Ward method, variable method, shortest distance method, or longest distance method may be used as the method for measuring the distance between clusters. Furthermore, the principal component analysis of the classification unit 8 in this embodiment is performed using a correlation coefficient matrix or a variance-covariance matrix, etc. Note that the classification unit 8 may also perform classification using methods other than cluster analysis and principal component analysis.

[0033] The evaluation unit 9 evaluates the similarity or dissimilarity of the target space to each spatial type based on the results of the cluster analysis or principal component analysis performed by the classification unit 8. For example, the evaluation unit 9 evaluates the similarity or dissimilarity of the target space to each spatial type using the squared Euclidean distance described above.

[0034] [Specific examples of classification of target spaces] Figures 2 to 7 illustrate specific examples of classifying the spatial type of the target space. In this embodiment, the Genki space (registered trademark), designed to construct a biophilic space, is used as an example.

[0035] Figure 2 shows the Genki space.

[0036] Genki Space is an office space designed for long-term stays while coexisting with plants, located within a test room (9.40m wide, 11.45m deep, 2.8m high) in Toyota City, Aichi Prefecture. Genki Space incorporates multiple plants classified by impression evaluations by a plant coordinator (a specialist in plant space design). In addition to these plants, Genki Space also features interior design and furnishings. The test room before the introduction of plants serves as a control group without any plants.

[0037] The Genki Space has three different types of areas, each with a different arrangement of plants. Specifically, the Genki Space has a Ten area with plants that reduce stress, a Sen area with plants that improve concentration, and a Men area with plants that improve vitality. In each of the Ten, Men, and Sen areas, a plant coordinator has selected and placed plants that contribute to the positive effects on people. The selection of these plants can be done using, for example, the selection method disclosed in Japanese Patent Publication No. 2022-019372. Each of the Ten, Men, and Sen areas has six seats. Blinds are installed on the windows to eliminate the influence of distant views.

[0038] Figure 3 shows the results of a questionnaire given to subjects seated in each seat of the Genki Space.

[0039] The following seats were selected to prevent visual information from different areas from entering the subject's field of view when they are seated facing forward: Ten2, Ten6, Sen2, Sen4, Men1, and Men2. Eight subjects sat in these seats for several minutes and then completed a questionnaire about their impressions. As shown in Figure 3, the impressions received from these three types of seats were different. In the Ten area, many respondents reported feeling "somewhat relaxed." In the Sen area, many respondents reported feeling "mentally focused." In the Men area, many respondents reported feeling "full of energy," "easier to discuss and make decisions," and "somewhat relaxed."

[0040] In the Ten area, there were more responses indicating "feeling somewhat relaxed" than responses indicating "feeling energized" or "finding it easier to discuss and make decisions." Therefore, it can be seen that the Ten area induces a static feeling of relaxation and stress reduction when alone. In other words, the Ten area can be expected to be the primary spatial type that reduces a person's stress.

[0041] Many respondents in the Sen area reported feeling "mentally focused." Therefore, it can be seen that the Sen area is a promising type of secondary space that can improve a person's concentration.

[0042] In the Men area, there were more responses indicating "feeling energized" and "finding it easier to discuss and make decisions" than responses indicating "feeling somewhat relaxed." Therefore, it can be seen that the Men area makes people feel energized and dynamic, replenishing them with the power to discuss with others. In other words, the Men area can be expected to be a third spatial type that enhances people's vitality.

[0043] Thus, it can be seen that the Ten, Sen, and Men areas within the Genki space function as different spatial types that have different effects on people depending on the arrangement of the plants.

[0044] Figure 4(a) shows an image of the Ten area in the Genki space. Figure 4(b) shows an image of the Sen area in the Genki space. Figure 5(a) shows an image of the Men area in the Genki space. Figure 5(b) shows an image of the control area with plants.

[0045] The images shown in Figures 4(a) to 5(a) were acquired as follows: An Insta360 Pro2 was used as the imaging device 2, and it was mounted on a tripod at 120 cm above the floor, which is the eye level of each seat in the Genki space, to acquire omnidirectional images. The acquired omnidirectional images were processed as follows: The acquired omnidirectional images were converted from equirectangular projection to Lambert equal-area cylindrical projection using the image stitching software Hugin, and then corrected so that the front of the seat was aligned with the center of the image. This image (7680 x 3840 pixels, aspect ratio 1:2) was reduced in size (4096 x 2048 pixels, aspect ratio 1:2) using the image editing software Adobe Photoshop, and the color tone was adjusted using the Camera Raw filter.

[0046] The images of Ten1 to Ten6 shown in Figure 4(a) were taken at each of the Ten1 to Ten6 seats in the Ten area of ​​the Genki space and processed as described above. The images of nTen1 to nTen6 shown in Figure 4(a) are images of the control area without plants, taken at each of the Ten1 to Ten6 seats before plants were introduced and processed as described above. The same applies to the images of Sen1 to Sen6 shown in Figure 4(b), the images of nSen1 to nSen6 shown in Figure 4(b), the images of Men1 to Men6 shown in Figure 5(a), and the images of nMen1 to nMen6 shown in Figure 5(a).

[0047] The images of Forest1 to Forest6 shown in Figure 5(b) are images of the control area with plants. The images of Forest1 to Forest6 shown in Figure 5(b) were taken using imaging device 2 on the forest therapy road in Aya Forest, Aya-cho, Higashimorokata-gun, Miyazaki Prefecture, and processed as described above. Aya Forest is a place where stress reduction effects have been observed based on the results of measuring physiological responses such as heart rate and pulse wave of subjects who have taken a forest bath.

[0048] The green view ratio in each image shown in Figures 4(a) to 5(b) is calculated by the green view ratio analysis unit 71 as shown in Table 1. The green view ratio in each area of ​​the Genki space was highest in the Men area, followed by the Ten area and then the Sen area. The green view ratio in each area of ​​the Genki space was significantly lower than that of the Aya Forest, a control area with plants, and the values ​​were close to each other. When actually staying in the Genki space, the amount of plants in the Men area felt significantly larger than the others and was clearly distinguishable subjectively, but there was no significant difference in green view ratio. It is difficult to classify each area using only the green view ratio as a feature. This means that indicators other than the green view ratio are necessary when considering the relationship between human impression and feature.

[0049] [Table 1]

[0050] Therefore, the inventors investigated whether it was possible to classify the Ten, Sen, and Men areas in the Genki space into three spatial types—first spatial type, second spatial type, and third spatial type—by extracting various features from the images shown in Figures 4(a) to 5(b). As a result, the inventors found that by extracting at least the green view ratio, fractal dimension, and fluctuation value of each image, it was possible to classify each area in the Genki space into its respective spatial type.

[0051] Figure 6 shows the cluster analysis results for each image shown in Figures 4(a) to 5(b).

[0052] The feature extraction unit 7 of the spatial determination device 4 extracted the green view ratio calculated by the green view ratio analysis unit 71, the fractal dimension analyzed using the FBM method by the fractal analysis unit 72, and the 1 / f fluctuation value calculated by the fluctuation analysis unit 73 as features from each image shown in Figures 4(a) to 5(b). Then, the classification unit 8 of the spatial determination device 4 performed hierarchical cluster analysis using the squared Euclidean distance and the group average method. The classification unit 8 can perform cluster analysis and principal component analysis using, for example, statistical analysis software (JUSE-StatWorks V5).

[0053] Figure 6 shows a dendrogram representing the results of the cluster analysis. As shown in Figure 6, the image group of the control area with plants (Forest) and the image group of the Genki space are classified into different clusters. Furthermore, the image group of the areas with plants in the Genki space (Ten, Sen, Men) and the image group of the control area without plants (nTen, nSen, nMen) are classified into different clusters. Moreover, in the Genki space, the image group of the Ten area, the image group of the Sen area, and the image group of the Men area are classified into different clusters. In other words, it can be seen that the spatial determination device 4 of this embodiment is capable of classifying the image group of the Ten area, which represents a first spatial type that reduces human stress, the image group of the Sen area, which represents a second spatial type that improves human concentration, and the image group of the Men area, which represents a third spatial type that improves human vitality, into different clusters. Note that the reason why the image of Men6 and the images of Ten3 and Ten4 were classified into the same cluster is thought to be because plants from the Ten area are visible on both sides of the image of Men6.

[0054] Furthermore, the vertical axis in Figure 6 shows the dissimilarity between clusters; a higher value indicates less similarity between clusters, while a lower value indicates greater similarity. Similarity or dissimilarity is evaluated according to the squared Euclidean distance between clusters. In other words, by evaluating similarity or dissimilarity, it is possible to quantitatively assess how similar the spatial representations of the images belonging to a cluster are.

[0055] Here, when the arrangement of plants in a specific area of ​​the Genki space is changed from the arrangement shown in Figures 4(a) to 5(a), an image is acquired and input to the spatial determination device 4, the spatial determination device 4 extracts the above-mentioned features from the image and performs cluster analysis. As a result of this analysis, the image of the area where the plant arrangement has been changed will belong to one of the clusters of areas with plants in the Genki space shown in Figure 6, or it will belong to a new cluster. In this case, the spatial determination device 4 can evaluate the similarity or dissimilarity between the cluster to which the image of the area where the plant arrangement has been changed belongs and the cluster representing the desired spatial type. This allows the spatial determination device 4 to quantitatively evaluate how close the area where the plant arrangement has been changed is to the desired spatial type.

[0056] Figure 7 shows the principal component analysis results for each image shown in Figures 4(a) to 5(b).

[0057] As shown in Figure 7, the image group of the control area with plants (Forest) and the image group of the Genki space are distributed in different ranges. Furthermore, the image group of the areas with plants in the Genki space (Ten, Sen, Men) and the image group of the control area without plants (nTen, nSen, nMen) are distributed in different ranges. Moreover, in the Genki space, the image group of the Ten area, the image group of the Sen area, and the image group of the Men area are distributed in different ranges. In other words, the spatial determination device 4 of this embodiment can classify the images shown in Figures 4(a) to 5(b) as different spatial types of image groups by using the green view ratio, fractal dimension, and 1 / f fluctuation value as feature quantities for each image shown in Figures 4(a) to 5(b). Note that the image group of the control area with plants (Forest) is distributed at a location far from the image group of the Genki space. This is because the average green view ratio of the image group in the control area with plants (Forest) is 2 to 3 times higher than the average green view ratio of each image group in the Genki space areas with plants (Ten, Sen, Men).

[0058] As described above, the spatial determination device 4 of this embodiment is a device for determining the impact that a target space has on a person. The spatial determination device 4 comprises a feature extraction unit 7 that extracts feature quantities from an image of the target space that change according to the arrangement of plants contained in the image, and a classification unit 8 that classifies the target space into one of a plurality of spatial types that have different impacts on people. The plurality of spatial types include at least a first spatial type that reduces stress on people, a second spatial type that improves concentration on people, and a third spatial type that improves vitality on people. The feature extraction unit 7 extracts at least the green view ratio of the image of the target space, the fractal dimension of the image of the target space, and the fluctuation value of the image of the target space as feature quantities from the image of the target space. The classification unit 8 classifies the target space by classifying the image of the target space into one of the plurality of spatial type image groups based on the extracted feature quantities.

[0059] The spatial determination device 4 can classify the effects of a target space where plants are placed on people into at least three categories: stress reduction, improved concentration, and increased vitality. According to these effects, it can then classify the target space into at least three spatial types: a first spatial type, a second spatial type, and a third spatial type. In this case, simply by acquiring an image of the target space and inputting it into the spatial determination device 4, the device can extract the above-mentioned features and classify the target space based on the extracted features. Therefore, the spatial determination device 4 of this embodiment can easily determine what type of effect the target space will have on people, even if the target space where plants are placed is an unknown space.

[0060] Furthermore, in the spatial determination device 4 of this embodiment, the image of the target space is an omnidirectional image obtained by capturing images in all directions around a predetermined position within the target space.

[0061] As a result, the spatial determination device 4 of this embodiment can classify a target space by considering the arrangement of plants as far as a person standing at a predetermined position within the target space can see. Therefore, the spatial determination device 4 of this embodiment can more accurately reflect the impact that a target space with plants has on people in its classification of the target space. Thus, the spatial determination device 4 of this embodiment can easily and accurately determine what type of impact a target space has on people.

[0062] Furthermore, in the spatial determination device 4 of this embodiment, the classification unit 8 classifies the target space by performing cluster analysis or principal component analysis on the image of the target space using the extracted features.

[0063] As a result, the spatial determination device 4 of this embodiment can classify the target space using relatively simple methods such as cluster analysis or principal component analysis. Therefore, the spatial determination device 4 of this embodiment can more easily determine what type of impact the target space has on people.

[0064] Furthermore, the spatial determination device 4 of this embodiment further includes an evaluation unit 9 that evaluates the similarity or dissimilarity of the target space to the spatial type based on the results of cluster analysis or principal component analysis.

[0065] As a result, the spatial determination device 4 of this embodiment can quantitatively evaluate how closely the target space resembles each spatial type. Therefore, the spatial determination device 4 of this embodiment can easily and quantitatively determine what type of influence the target space has on people.

[0066] Furthermore, in the spatial determination device 4 of this embodiment, the fluctuation value of the image of the target space is the 1 / f fluctuation value.

[0067] As a result, the spatial determination device 4 of this embodiment uses 1 / f fluctuation, which affects people in terms of comfort and relaxation, as a feature quantity, so that it can reliably classify the target space into each spatial type. Therefore, the spatial determination device 4 of this embodiment can easily and reliably determine what type of effect the target space has on people.

[0068] Although embodiments of the present invention have been described in detail above, the present invention is not limited to the embodiments described above, and various modifications can be made without departing from the spirit of the invention as described in the claims. The present invention can be modified by adding the configuration of one embodiment to the configuration of another embodiment, replacing the configuration of one embodiment with that of another embodiment, or deleting a part of the configuration of one embodiment. [Explanation of Symbols]

[0069] 4...Spatial determination device, 7...Feature extraction unit, 8...Classification unit, 9...Evaluation unit

Claims

1. A spatial determination device that determines the impact on people of a target space in which plants are arranged, A feature extraction unit extracts feature quantities from an image that change according to the arrangement of plants contained in the image of the target space, The system includes a classification unit that classifies the aforementioned target space into one of several spatial types that have different effects on people, The aforementioned multiple spatial types include at least a first spatial type that reduces stress, a second spatial type that improves concentration, and a third spatial type that improves vitality. The feature extraction unit extracts at least the green view ratio of the image of the target space, the fractal dimension of the image of the target space, and the fluctuation value of the image of the target space from the image of the target space as features, The classification unit classifies the target space by classifying the images of the target space into one of the multiple spatial types of image groups based on the extracted feature quantities. A spatial determination device characterized by the following features.

2. The image of the target space is an omnidirectional image obtained by taking photographs in all directions around a predetermined position within the target space. The spatial determination device according to feature 1.

3. The fluctuation value of the image in the aforementioned target space is the 1 / f fluctuation value. The spatial determination device according to feature 1.

Images