Cloud observation system, cloud observation method, and program

Abstract

Provided are a cloud observation system, method, and program that enable observation of clouds at night by using a camera.　A cloud observation system (1) comprises: an acquisition unit (11) that acquires sky images (G1, G2, G3) which are captured by a camera (10) and which include at least the sky; a threshold value determination unit (12) that determines a threshold value on the basis of pixel values of a plurality of edges in the sky images; and a cloud determination unit (13) that determines, on the basis of the threshold value and pixel values in the sky images, pixels, from among the pixels constituting the sky images, in which clouds are reflected.

Description

Technical Field

[0001] The present disclosure relates to a cloud observation system, a cloud observation method, and a program.

[0002] As a cloud observation method, it is known to use an all-sky camera installed on the ground. For example, Patent Document 1 describes observing clouds based on an empty image captured by an all-sky camera.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] A method for observing clouds at night using a camera is required. However, the algorithm for determining clouds and the sky described in Patent Document 1 cannot be used at night.

[0005] The present disclosure provides a cloud observation system, method, and program capable of observing clouds at night with a camera.

Means for Solving the Problems

[0006] The cloud observation system of the present disclosure includes an acquisition unit that acquires an empty image including at least the sky, captured by a camera; a threshold determination unit that determines a threshold based on pixel values of a plurality of edges in the empty image; and a cloud determination unit that determines pixels in which clouds are reflected from a plurality of pixels constituting the empty image based on the pixel values in the empty image and the threshold.

Brief Description of the Drawings

[0007]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Mode for Carrying Out the Invention

[0008] [First Embodiment] Hereinafter, the cloud observation system of the first embodiment of the present disclosure will be described with reference to the drawings.

[0009] The cloud observation system 1 of the first embodiment processes an aerial image captured by at least one camera 10. The aerial image includes at least the sky. The camera 10 can be any camera as long as it can capture the sky. In this embodiment, in order to capture a wide range of the sky with one camera, an all-sky camera with a fish-eye lens is installed facing upward, but it is not limited to this. If the camera 10 is installed vertically upward and horizontally, as shown in FIG. 2, the center P1 of the aerial image G1 obtained from the camera 10 is directly above (elevation angle 90 degrees), and the elevation angle decreases as it goes from the center P1 to the edge of the image. Also, if the orientation of the camera 10 is known, the azimuth in the image is also known. That is, the pixels at any position in the aerial image are each known for both the elevation angle and the azimuth. In the figure, the azimuths are shown as north (N), south (S), west (W), and east (E).

[0010] As shown in FIG. 1, the cloud observation system 1 includes an acquisition unit 11, a threshold determination unit 12, and a cloud determination unit 13. The cloud observation system 1 may further include a moon detection unit 14 and a sky region setting unit 15. These units 11 to 15 are realized by the cooperation of software and hardware when the processor 1b, such as a CPU, executes a program stored in advance in the memory in a computer equipped with a storage 1a such as a memory and various interfaces.

[0011] <Acquisition unit 11> The acquisition unit 11 shown in FIG. 1 acquires at least one blank image. The camera 10 preferably takes pictures regularly to obtain blank images for cloud observation. The period at which the camera 10 takes pictures can be arbitrarily set, for example, every 1 minute, every 5 minutes, every 1 hour, etc. When the camera 10 takes pictures at night, it is necessary to increase the shutter speed compared to daytime. The shutter speed is also affected by the shooting environment such as the camera equipment and the aperture, but for example, it can be 2.5 seconds to 5 seconds. Since the appearance of the blank image changes depending on the shutter speed, it is preferable to take pictures with the camera 10 multiple times with different shutter speeds every time a predetermined shooting timing comes, so as to obtain a plurality of blank images. This is because an appropriate blank image with a shutter speed can be selected from the plurality of blank images.

[0012] <Blank area setting unit 15> As shown in Fig. 1, the empty area setting unit 15 sets an empty area Ar10 in the empty image as shown in Fig. 6. Fig. 6 is a diagram showing the empty image G2 obtained from the all-day camera 10. As shown in Fig. 6, in the peripheral part of the empty image G2 obtained from the all-day camera 10, there may be objects P2 other than empty, such as buildings in the street around the all-day camera 10, street lights, and obstacles. Therefore, the empty area setting unit 15 removes the peripheral part of the empty image G1 from the empty area Ar10 where empty is shown. The peripheral part means an area outside the circle with a predetermined radius φ1 from directly above P1 of the empty image. Also, when an obstacle P3 that is not in the peripheral part is captured, the obstacle P3 is also removed from the empty area Ar10. In Fig. 6, the empty area Ar10 is shown as the range surrounded by the dashed-dotted line. As a method for the empty area setting unit 15 to set the empty area Ar10, it may be a method of specifying based on coordinates individually specified by the user in advance, or it may be set to recognize automatically. As a method of automatically recognizing, for example, a plurality of empty images are continuously captured during the day, and the method of utilizing the change in the pixel values of the area where the clouds flow is mentioned. Specifically, in the inner area of the circle with a predetermined radius φ1 in the empty image G2, an area where the change in pixel values is relatively large among a plurality of empty images is recognized as the empty area Ar10, and conversely, an area where the change in pixel values is relatively small is recognized as a non-empty area. The predetermined radius φ1 can be appropriately changed according to the performance of the camera. In the case of a fish-eye lens, the larger the predetermined radius φ1, the greater the distortion of the image, so it can be appropriately changed according to the performance of the camera.

[0013] <Threshold determination unit 12> As shown in Fig. 1, the threshold determination unit 12 determines a threshold based on the pixel values of a plurality of edges in the empty image. The threshold is used for the cloud determination unit 13 for cloud determination. As shown in Fig. 2, when clouds (C1, C2, C3) are captured in the empty image G1, the boundary between the empty and the clouds can be detected as an edge. An edge is a pixel whose pixel value changes abruptly compared to the surrounding pixels in the empty image. Various known algorithms can be used for the edge detection method. For example, using the differential value of the pixel value can be mentioned.

[0014] During the day, clouds are white and the sky appears blue due to the refraction of sunlight. Therefore, clouds can be identified by comparing their blueness. However, at night, since there is no sunlight, clouds cannot be detected by comparing blueness. At night, the light of the moon and streetlights is reflected by clouds and reaches camera 10, so clouds are brighter than the sky. Thus, in this embodiment, it is determined whether a pixel is a cloud or not using the brightness difference based on pixel values. The brightness (L) used in this embodiment is calculated as (the red value (R) of the pixel + the green value (G) of the pixel + the blue value (B) of the pixel) / 3, but the calculation formula for brightness can be variously changed.

[0015] Since the threshold value of the pixel value (brightness) for discriminating clouds and the sky by the threshold determination unit 12 is likely to be different for each sky image, the threshold value is determined for each sky image. The threshold determination unit 12 determines the threshold value based on the pixel values (brightness) of a plurality of edges in the sky image. Specifically, as shown in FIG. 2, the pixel values (brightness) of a plurality of edges that are often the boundaries between clouds and the sky are calculated respectively, and statistical processing is performed on the plurality of pixel values (brightness) to determine the threshold value. In this embodiment, the median value of the plurality of pixel values (brightness) is used as the threshold value, but it is not limited to the median value as the statistical processing. For example, the average value or the mode value may be used. By using the median value, an appropriate threshold value excluding noise such as outliers can be determined. In the example shown in FIG. 2, at the edge of cloud C1, there are a portion with brightness = 0.27 and a portion with brightness = 0.30, at the edge of cloud C2, there are a portion with brightness = 0.28 and a portion with brightness = 0.31, and at the edge of cloud C3, there are a portion with brightness = 0.27 and a portion with brightness = 0.28. In this case, the median value of each brightness value is 0.28, and the threshold value is determined to be 0.28. In the drawing, since the brightness of all pixels at the edges cannot be shown, some brightness values are shown for easy understanding.

[0016] Figures 3 to 5 are diagrams showing a method for determining a threshold value based on an actual empty image. Figure 3 shows the empty image acquired by the acquisition unit 11 in grayscale. Figure 4 is a diagram showing the edges extracted from the empty image shown in Figure 3 in white. Figure 5 is a diagram showing the brightness of the extracted edges by numbers in the figure and indicating that the median value of the brightness is 0.28. As shown in Figure 5, it can be seen that the pixel values (brightness) serving as the threshold value differ depending on the location in the empty image.

[0017] [Determination method in the case of all clouds or all clear] When almost all clouds are captured in the empty area Ar10 of the empty image and when almost all clear sky is captured, no edges are detected or the number of detected edges decreases. In this case, the threshold value cannot be determined based on the pixel values of a plurality of edges. Therefore, in the first embodiment, the threshold value determination unit 12 is configured as follows.

[0018] That is, the threshold value determination unit 12 extracts the edges in the first empty image captured at the first time point, and when the number of the extracted edges is equal to or less than a predetermined value, adopts the threshold value determined based on the second empty image captured at the second time point as the threshold value of the first empty image. The second empty image is an image captured in the same time zone on a day different from the shooting date of the first empty image. If it is a different day, there is a possibility that clouds and the sky are mixed and edges exceeding a predetermined value can be detected. If it is the same time zone, it is considered that the conditions of the threshold value are the same or almost the same. Here, the same time zone is preferably the same time in the same camera, for example, but there may be a time difference. Although the time difference between the first time point and the second time point is preferably zero or small, for example, even if there is a time difference of about 10 minutes, it can exhibit an effect. More preferably, it is within 2 minutes, preferably within 1 minute, and more preferably within 10 seconds. The difference in the number of days between the first empty image and the second empty image is preferably within 7 days, more preferably within 3 days, and even more preferably 1 day. This is because it is considered that there is almost no difference in the threshold value with a one-day difference. When there are multiple second empty images, it is preferable to select, from the multiple second empty images, the second empty image that has the same month age as or the closest month age to the first empty image. This is because the degree of the moon's waxing and waning will be the same or close, and the brightness of the clouds due to the moon will be the same or almost the same.

[0019] <Cloud determination unit 13> The cloud determination unit 13 shown in FIG. 1 determines, based on the pixel values and threshold values in the empty image, the pixels in which clouds are reflected (hereinafter also referred to as cloud pixels) from among the multiple pixels constituting the empty image. In the present embodiment, the threshold determination unit 12 determines the threshold based on the brightness of the pixels, and the cloud determination unit 13 determines the cloud pixels based on the brightness of the pixels and the threshold in the empty image. In the examples of FIGS. 2 and 5, since the threshold value is 0.28, the cloud determination unit 13 determines that the pixels with a brightness of 0.28 or more, or pixels with a brightness greater than 0.28, are cloud pixels. In the present embodiment, one threshold value is used for the target range of the empty region Ar10 in the empty image.

[0020] <Moon detection unit 14> The moon detection unit 14 shown in FIG. 1 detects the pixels in which the moon is reflected in the empty image. The pixels in which the moon is reflected are relatively the brightest compared to the other pixels in the empty region. In the example of FIG. 6, the brightness of the pixels in the region where the moon (M1) is present becomes high. The moon detection unit 14 calculates the brightness of all the pixels in the empty region Ar10 and determines that the pixel with the highest brightness is the pixel in which the moon is reflected. When the moon detection unit 14 is implemented, it is preferable that the target range for the threshold determination unit 12 to extract the edges and the target range for the cloud determination unit 13 to determine the clouds are the ranges obtained by removing the range of the moon from the empty region Ar10. This is to avoid the moon pixels from affecting the threshold value.

[0021] [Cloud observation method of the first embodiment] The cloud observation method executed by the above system 1 will be described with reference to FIG. 7.

[0022] First, in step ST100, the acquisition unit 11 acquires the first empty image captured by the camera 10. In the next step ST101, the empty region setting unit 15 sets an empty region Ar10 in the first empty image. In the next step ST102, the moon detection unit 14 detects the moon in the empty region Ar10 of the empty image. In the next step ST103, the moon region is removed from the empty region Ar10 to obtain a target region for threshold determination and cloud determination. Next, in step ST104, the threshold determination unit 12 extracts the edges of the target region. If the number of edges of the target region is less than or equal to a predetermined value (ST105: YES), the threshold determination unit 12, in step ST106, adopts the threshold determined based on the second empty image captured in the same time zone on a day different from the shooting date and time of the first empty image as the threshold of the first empty image, and proceeds to the next step ST108. On the other hand, if the number of edges of the target region is not less than or equal to the predetermined value (ST105: NO), in step ST107, the threshold determination unit 12 determines the threshold based on the pixel values of the plurality of edges, and proceeds to the next step ST108. In step ST108, the cloud determination unit 13 determines the pixels in which clouds are reflected from the plurality of pixels constituting the empty image based on the pixel values and the threshold in the empty image.

[0023] [Second Embodiment] The cloud observation system and method according to the second embodiment will be described. The same parts as those in the first embodiment are denoted by the same reference numerals, and the description thereof will be omitted. As shown in FIG. 8, the cloud observation system 1 according to the second embodiment further includes a region setting unit 16. As shown in FIG. 9, the region setting unit 16 sets a plurality of regions (A1, A2... An) for the empty image G2 (preferably, the empty region Ar10 of the empty image G2). n is a natural number of 2 or more. In the example of FIG. 9, n = 9, but the number of regions can be changed as appropriate. It is also possible to arbitrarily change how the empty region Ar10 is divided into a plurality of regions in terms of shape. The threshold determination unit 12 determines the threshold for each of the plurality of regions A1 to A9. The cloud determination unit 13 determines the pixels in which clouds are reflected in the region using the threshold for each of the regions A1 to A9.

[0024] That is, in the first embodiment, only one threshold is determined for the empty image, and clouds are determined using one threshold. In contrast, in the second embodiment, the empty image is divided into a plurality of regions A1 to A9, a threshold is determined for each region, and cloud pixels are determined for each region.

[0025] For example, as shown in FIG. 9, in the empty image G2, the region A1 close to the moon (M1) is likely to become bright under the influence of the moon. Conversely, the part of the empty image G2 far from the moon (M1) (for example, region A7, etc.) is not affected by the moon and becomes relatively dark. Also, the empty image G2 includes directions close to streetlights (for example, N, W, S) and directions without streetlights such as the sea (E). Directions without streetlights such as the sea are relatively darker than directions close to streetlights. Therefore, in the empty image G2, the brightness of clouds varies depending on the location. Thus, by dividing the empty image G2 into a plurality of regions A1 to A9, determining the threshold for each region, and determining clouds for each region, it is possible to improve the cloud determination accuracy compared to the case of using one threshold for the entire empty image.

[0026] Specific methods for determining the threshold include the following methods. (1) For each of the regions A1 to A9, the threshold may be determined based on the pixel values (brightness) of the edges within the region. (2) A first threshold may be determined based on the pixel values (brightness) of all the edges within the target range of the empty region Ar10 (within the plurality of regions A1 to A9), and the thresholds for each of the regions A1 to A9 may be determined by correcting the first threshold according to the direction corresponding to the region. For example, when the east (E) is dark and the northwest (N, W) is bright, the value obtained by subtracting the value corresponding to the east (E) from the first threshold may be used as the threshold for region A2, and the value obtained by adding the value corresponding to the northwest (N, W) to the first threshold may be used as the thresholds for regions A6 to A9. In this case, the correction values corresponding to the directions are preset. (3) Determine the first threshold based on the pixel values (brightness) of all edges within the target range of the empty region Ar10, and the thresholds for each of the regions A1 to A9 may be determined by correcting the first threshold with a correction value corresponding to the pixel value (brightness) of the peripheral part corresponding to the region. For example, the threshold for region A7 may be determined by correcting the first threshold with a correction value corresponding to the pixel value of the outer peripheral part of region A6 (the pixel value of the northern peripheral part).

[0027] That is, summarizing the above (2) and (3), the threshold for each region is determined based on the pixel values of the edges within the plurality of regions A1 to A9 and the pixel values of the peripheral part of the empty image G2.

[0028] (4) As shown in FIG. 10, the plurality of regions include a region A1 (first region) including the moon (M1) and regions A2 to A9 (second regions) that are further away from the moon (M1) than the region A1 (first region). A threshold of 0.28 is determined based on the pixel values of the edges within the plurality of regions (A1 to A9) including the first region and the second region, and the threshold for the first region (A1) and the threshold for the second region (A2 to A9) may be made different. As an example of a specific calculation method, the threshold of 0.28 is calculated from the pixel values of the edges of all regions, and the threshold for the first region (A1) may be corrected in the brighter direction to, for example, 0.29. Also, as shown in FIG. 11, the threshold of 0.28 calculated based on the plurality of regions (A1 to A9) may be used as the threshold for the first region (A1), and the threshold for the second region (A2 to A9) may be corrected in the darker direction to, for example, 0.27.

[0029] (5) As shown in FIG. 12, when setting a plurality of regions (A10 to A13) excluding the moon (M1), the regions (A10, A11, A12) adjacent to the moon may be used as the first region, and the region (A13) that is further away from the moon (M1) than the first region (A10, A11, A12) may be used as the second region.

[0030] [Third Embodiment] The third embodiment, similar to the first embodiment shown in FIG. 1, does not have the region setting unit 16 that sets a plurality of regions. As shown in FIG. 13, the threshold determination unit 12 determines the threshold for each of the plurality of pixels constituting the empty image G1 based on the distance from the pixel to the edge and the pixel value of the edge. The cloud determination unit 13 makes a determination using the threshold for each pixel constituting the empty image. For example, as shown in FIG. 13, when calculating the threshold for a certain pixel (U1), the threshold is calculated based on the distances (d1, d2, d3, d4) from the pixel (U1) to the edges and the pixel values (v1, v2, v3, V4) of the edges. The influence of the pixel value (brightness) of the edge given to a certain pixel (U1) is greater as the distance is closer and smaller as the distance is farther. Therefore, when obtaining the threshold for a certain pixel (U1), it is possible to weight the pixel values of all the edges by the distance. Thereby, the influence of the pixel value of the edge with a short distance from a certain pixel (U1) becomes large, and the influence of the pixel value of the edge with a long distance from a certain pixel (U1) can be reduced. According to this method, although the calculation cost increases, the accuracy of the threshold can be significantly improved.

[0031] [Modification Example] In the above embodiment, the threshold used for cloud determination is brightness, but it is not limited to this. For example, if it is a pixel value, it may be a single-color pixel value or an appropriately combined pixel value.

[0032] As described above, the cloud observation system 1 according to the first, second, and third embodiments includes an acquisition unit 11 that acquires an empty image (G1; G2; G3) including at least the sky photographed by the camera 10, a threshold determination unit 12 that determines a threshold based on the pixel values of a plurality of edges in the empty image, and a cloud determination unit 13 that determines, based on the pixel values and the threshold in the empty image, the pixels in which clouds are reflected from the plurality of pixels constituting the empty image.

[0033] The cloud observation methods according to the first, second, and third embodiments include obtaining an empty image including at least the sky, captured by the camera 10 (ST100), determining a threshold value based on the pixel values of a plurality of edges in the empty image (ST107; ST106), and determining pixels in which clouds are reflected from among a plurality of pixels constituting the empty image based on the pixel values and the threshold value in the empty image (ST108).

[0034] In this way, since many of the edges in the empty image are the boundaries between clouds and the sky, the threshold value is determined based on the pixel values of the plurality of edges in the empty image. Therefore, even at night when the color of the sky due to sunlight during the day is not reflected, clouds can be determined and clouds can be observed.

[0035] Although not particularly limited, as in the first, second, and third embodiments, it is preferable that the threshold determination unit 12 determines the threshold value based on the brightness of the pixels of the plurality of edges, and the cloud determination unit 13 determines the pixels in which clouds are reflected based on the brightness of the pixels and the threshold value in the empty image.

[0036] According to this configuration, for example, it is possible to improve the determination accuracy of clouds as compared with the case of determining clouds based on pixel values of a single color.

[0037] Although not particularly limited, as in the second embodiment, it is provided with a region setting unit 16 that sets a plurality of regions (A1 to A9; A10 to A13) for the empty image, the threshold determination unit 12 determines the threshold value for each of the plurality of regions (A1 to A9; A10 to A13), and the cloud determination unit 13 preferably determines the pixels in which clouds in the region are reflected using the threshold value for each region.

[0038] For example, in the blank image, the area closer to the moon tends to become brighter under the influence of the moon. Conversely, the part of the blank image farther from the moon is not affected by the moon and becomes relatively darker. Also, the blank image includes the direction closer to streetlights and the direction without streetlights such as the sea, and the direction without streetlights such as the sea is relatively darker compared to the direction closer to streetlights. Therefore, by setting multiple regions in the blank image and using a threshold value for each region, it is possible to improve the cloud determination accuracy compared to the case of using a single threshold value for the entire blank image.

[0039] Although not particularly limited, as in the second embodiment shown in FIG. 9, the blank image G2 is an image captured by an all-sky camera, and it is preferable that the threshold determination unit 12 determines the threshold value for each of the regions (A1 to A9) based on the pixel values of the edges in the plurality of regions and the pixel values of the peripheral portion of the blank image G2.

[0040] In this way, since the threshold value for each region is determined using not only the pixel values of the edges in the plurality of regions but also the pixel values of the peripheral portion of the blank image where streetlights or the like may be reflected, the threshold value can be determined in consideration of streetlights, and it is possible to improve the cloud determination accuracy.

[0041] Although not particularly limited, as in the second embodiment shown in FIGS. 10 to 12, the blank image G2 is provided with a moon detection unit 14 that detects pixels in which the moon is reflected, and the plurality of regions (A1 to A9; A10 to A13) include a first region (A1; A10 to A12) that includes or is adjacent to the moon (M1) and a second region (A2 to A9; A13) that is farther from the moon than the first region. It is preferable that the threshold determination unit 12 determines the threshold value for each region based on the pixel values of the edges in the plurality of regions including the first region and the second region, and sets the threshold value of the first region and the threshold value of the second region to different values.

[0042] The edges of the first region including or adjacent to the moon are easily affected by the moonlight and tend to be bright. Therefore, while determining a threshold value based on the pixel values of the edges in a plurality of regions, the threshold values of the first region and the second region are made different. As a result, it is possible to appropriately set the threshold value of the first region that is easily affected by the moonlight and the threshold value of the second region that is not affected by the moonlight, and it becomes possible to improve the cloud determination accuracy.

[0043] Although not particularly limited, as in the third embodiment, the threshold determination unit 12 extracts a plurality of edges in the empty image G1, and determines the threshold value for each of the plurality of pixels constituting the empty image G1 based on the distances (d1, d2, d3, d4) from the pixel to the edge and the pixel values (v1, v2, v3, v4) of the edge. It is preferable that the cloud determination unit 13 makes a determination using the threshold value for each pixel constituting the empty image G1.

[0044] An edge closer to the pixel has a greater influence on cloud determination than an edge farther from the pixel. Therefore, since the threshold value is determined based on the distance from the pixel to the edge and the pixel value of the edge, it becomes possible to improve the cloud determination accuracy as compared with the case where the threshold value is determined uniformly.

[0045] Although not particularly limited, as in the first, second, and third embodiments, a moon detection unit 14 that detects pixels in which the moon appears in the empty image is provided, and it is preferable that the threshold determination unit 12 removes the range of the moon from the target range for extracting the edge and then extracts the edge.

[0046] According to this configuration, since the range of the moon that affects the determination of the threshold value is removed and the edge is extracted, it becomes possible to improve the cloud determination accuracy.

[0047] Although not particularly limited, as in the first, second, and third embodiments, the acquisition unit 11 acquires a first empty image, and when the number of edges extracted based on the acquired first empty image is equal to or less than a predetermined value, the threshold value determined based on a second empty image taken in the same time zone on a day different from the shooting date and time of the first empty image is adopted as the threshold value of the first empty image.

[0048] The fact that the number of edges is equal to or less than a predetermined value means that the empty area shown in the first empty image is either completely cloudy or completely clear, and thus the threshold value cannot be determined. Therefore, the threshold value of the second empty image in the same time zone on another day, which is considered to have substantially the same threshold conditions, is used, so that clouds in the first empty image can be appropriately determined.

[0049] Although not particularly limited, as in the first, second, and third embodiments, the empty image is an image captured by an all-sky camera, and the threshold determination unit 12 preferably removes the peripheral portion of the empty image from the target range for edge extraction and then extracts the edges.

[0050] According to this configuration, the peripheral portion of the empty image obtained from an all-sky camera using a fish-eye lens, which is likely to show the streetlights themselves, is removed from the target range for edge extraction, so that an appropriate threshold value can be determined and the cloud determination accuracy can be improved.

[0051] The program according to this embodiment is a program that causes a computer (one or more processors) to execute the above method. Further, the computer-readable temporary recording medium according to this embodiment stores the above program.

[0052] As described above, the embodiments of the present disclosure have been described with reference to the drawings. However, the specific configuration should be considered not to be limited to these embodiments. The scope of the present disclosure is indicated not only by the description of the above embodiments but also by the scope of the claims, and further includes all modifications within the meaning and scope equivalent to the scope of the claims.

[0053] For example, in the claims, the specification, and the drawings, the execution order of each process such as the operations, procedures, steps, and stages in the apparatus, system, program, and method shown can be realized in any order, unless the output of the previous process is used in the subsequent process. Regarding the flow in the claims, the specification, and the drawings, even if terms such as "first" and "next" are used for convenience in the description, it does not mean that it is essential to execute in this order.

[0054] Each of the parts 11 to 15 shown in FIG. 1 is realized by executing a predetermined program with one or more processors, but each part may be configured with a dedicated memory or a dedicated circuit.

[0055] In the system 1 of the above embodiment, each of the parts 11 to 15 is implemented in the processor 1b of one computer, but each of the parts 11 to 15 may be distributed and implemented in a plurality of computers or in the cloud. That is, the above method may be executed by one or more processors.

[0056] It is possible to adopt the structure employed in each of the above embodiments in any other embodiment. In FIG. 1, for convenience of explanation, each of the parts 11 to 15 is implemented, but some of them can be arbitrarily omitted.

[0057] The specific configuration of each part is not limited to only the above-described embodiments, and various modifications are possible without departing from the spirit of the present disclosure.

Explanation of Reference Numerals

[0058] 10 Camera 11 Acquisition Unit 12 Threshold Determination Unit 13 Cloud Determination Unit 14 Moon Detection Unit 16 Region Setting Unit

Claims

1. An acquisition unit that acquires an empty image including at least the sky, captured by a camera; Among the plurality of pixels constituting the empty image, a plurality of pixels that become edges are extracted based on the brightness difference between adjacent pixels in the surroundings, and statistical processing is executed on the brightness values of each of the extracted plurality of edges to determine the statistical value of the brightness obtained by the statistical processing as a threshold value; a threshold value determination unit; Among the plurality of pixels constituting the empty image, a cloud determination unit that determines that a pixel in which the brightness is equal to or greater than the threshold value or the brightness is greater than the threshold value is a pixel in which a cloud is reflected; A cloud observation system comprising:

2. The cloud observation system according to claim 1, comprising a region setting unit that sets a plurality of regions for the empty image, The threshold value determination unit executes a process of extracting a plurality of edges from within the region and a process of determining the threshold value based on the plurality of extracted edges for each of the plurality of regions, The cloud determination unit determines pixels in which clouds are reflected in the region using the threshold value determined for each region; a cloud observation system.

3. The cloud observation system according to claim 2, The empty image is an image captured by an all-sky camera, The threshold value determination unit determines the threshold value determined for each region based on the pixel values of the edges in the region and the pixel values of the peripheral portion of the empty image; a cloud observation system.

4. The cloud observation system according to claim 2 or 3, comprising a moon detection unit that detects pixels in which the moon is reflected in the empty image, The plurality of regions include a first region including the moon or adjacent to the moon, and a second region further away from the moon than the first region, The threshold value determination unit sets the threshold value determined in the first region and the threshold value determined in the second region to different values; a cloud observation system.

5. The cloud observation system according to any one of claims 1 to 4, comprising a moon detection unit that detects pixels in which the moon appears in the sky image, The cloud observation system, wherein the threshold determination unit removes the range of the moon from the target range for extracting the edge and extracts the edge.

6. The cloud observation system according to any one of claims 1 to 5, wherein the acquisition unit acquires a first sky image, The cloud observation system, wherein when the number of edges extracted based on the acquired first sky image is equal to or less than a predetermined value, the threshold determined based on a second sky image taken in the same time zone on a day different from the shooting date and time of the first sky image is adopted as the threshold for the first sky image.

7. The cloud observation system according to any one of claims 1 to 6, wherein the sky image is an image taken by an all-sky camera, The cloud observation system, wherein the threshold determination unit removes the peripheral portion of the sky image from the target range for extracting the edge and extracts the edge.

8. obtaining a sky image including at least the sky, taken by a camera; among a plurality of pixels constituting the sky image, extracting a plurality of pixels that become edges based on the brightness difference between adjacent surrounding pixels, performing statistical processing on the brightness values of each of the extracted plurality of edge pixels, and determining the statistical value of the brightness obtained by the statistical processing as the threshold; determining that a pixel having a brightness equal to or higher than the threshold or a pixel having a brightness greater than the threshold among the plurality of pixels constituting the sky image is a pixel in which a cloud appears; A cloud observation method comprising:

9. A program for causing one or more processors to execute the method according to claim 8.

Images