JP7520649B2 - Weather information processing device, weather information processing system, weather information processing method, and weather information processing program - Google Patents

Description

The present invention relates to a weather information processing device, a weather information processing system, a weather information processing method, and a weather information processing program.

Weather observations are conducted using geostationary meteorological satellites, radar, and observation stations around the country. Various types of meteorological data obtained through weather observations are made publicly available and are widely used by businesses and individuals alike.

In recent years, meteorological data has been increasingly used in a variety of fields. For example, by using the obtained meteorological data to predict future weather, it is possible to carry out weather forecasts, disaster forecasts, river management, providing information to hikers, and agricultural production management. In addition, by predicting cloud movements using meteorological data, it is possible to predict the output of solar power generation. Furthermore, it is possible to use the output predictions of solar power generation obtained from meteorological data for operational management such as adjusting the amount of power supplied from power plants. In addition, by understanding the cloud conditions, meteorological data can also be used for aircraft operation management, etc.

Cloud information in meteorological data in particular plays an important role in disaster prevention, solar power generation, and the like. The Japan Meteorological Agency provides cloud information such as cloud top altitude, cloud type, and cloud quality information obtained using observation data and numerical forecast data from meteorological satellites. There is also technology that uses meteorological satellite images to extract various weather information contained in cloud images based on meteorological interpretation. With this technology, for example, visible (VIS) images can be used to identify clouds and determine cloud thickness from their brightness. Infrared (IR) images can also be used to determine cloud top temperature and cloud top altitude, and clouds can also be identified (Patent Document 1).

As for cloud thickness, a cloud amount determination device has been proposed that uses a cloud base altimeter (ceilometer) to determine cloud amount. A cloud base altimeter emits a laser beam vertically upward, receives the reflected light of the emitted laser beam from clouds using a light receiving element, and determines the cloud base height, which is the height of the bottom of the cloud, based on the time required from emission to reception.

As a technology for handling such meteorological data, there is a conventional technology that uses reflectance information obtained from visible image data (VIS) of meteorological satellite images and brightness temperature information obtained from infrared image data to determine the type of cloud and estimate the amount of solar radiation. There is also a conventional technology that obtains the longitude and latitude of clouds from a wide-angle image centered on the zenith captured by two cameras, and determines the cloud movement speed and wind speed (Patent Document 2). There is also a conventional technology that adjusts parameters that represent the effects of clouds and aerosols in satellite image data, and estimates the amount of solar radiation from the adjusted satellite image data (Patent Document 3). There is also a conventional technology that determines the cloud type from solar radiation intensity, corrects the solar radiation intensity from the cloud type, and calculates an estimate of the solar radiation intensity (Patent Document 4).

JP 2017-90442 A JP 2019-60754 A Japanese Patent Application Publication No. 11-211560 JP 2015-137903 A

However, when using satellite image data to determine cloud appearance (cloud scale, cloud base altitude, cloud top altitude, cloud thickness), the cloud thickness is predicted from brightness, for example, and it is difficult to determine the exact thickness because it is affected by the cloud altitude, etc. Also, while it is possible to measure cloud base altitude using a cloud base altimeter, cloud base altimeters are expensive and difficult to install because they are limited to measuring the cloud base vertically. As a result, they can be difficult to install in some locations, and installing multiple devices incurs a large cost, making it difficult to measure cloud thickness over a wide area. Therefore, it is difficult to accurately identify the cloud appearance.

Furthermore, with conventional technology that determines the type of cloud from meteorological satellite images, it is difficult to determine the exact thickness of clouds. This is true for conventional technology that determines the cloud movement speed and wind speed from wide-angle images centered on the zenith, conventional technology that estimates solar radiation by adjusting parameters that represent the effects of clouds and aerosols, and conventional technology that determines cloud types from solar radiation intensity. Therefore, it is difficult to accurately identify the state of clouds, and it is difficult to improve the performance of systems that use meteorological data.

The disclosed technology has been developed in consideration of the above, and aims to provide a weather information processing device, a weather information processing system, a weather information processing method, and a weather information processing program that improve the performance of a system that utilizes weather data.

In one aspect of the weather information processing device, weather information processing system, weather information processing method, and weather information processing program disclosed in the present application, the image matching unit performs image matching by identifying the cloud distribution captured in the celestial image and the cloud distribution captured in the satellite image based on the celestial image and the satellite image received from a ground sensor having a camera that captures a planar celestial image centered in the zenith direction and a weather data distribution system that distributes satellite images captured by a meteorological satellite. The range acquisition unit acquires horizontal distance information in the celestial image from horizontal distance information in the satellite image based on the result of the image matching. The cloud base altitude calculation unit calculates the cloud base altitude of the cloud based on the horizontal distance information in the celestial image and the zenith angle of the cloud in the celestial image.

In one aspect, the present invention can improve the performance of systems that utilize weather data.

FIG. 1 is a block diagram of a weather information providing system according to a first embodiment. FIG. 2 is a block diagram showing the details of the celestial camera and the weather information processing device. FIG. 3 is a diagram showing an example of converting a celestial spherical image into a planar image. FIG. 4 is a diagram for explaining an outline of image acquisition by the celestial camera. FIG. 5 is a diagram showing an example of a meteorological satellite image. FIG. 6 is a diagram showing an example of a cloud type determination table. FIG. 7 is a diagram showing an example of the image matching and analysis process. FIG. 8 is a diagram for explaining the cloud matching process. FIG. 9 is a diagram for explaining the zenith angle and the azimuth angle. FIG. 10 is a diagram showing an example of calculation of the cloud base altitude. FIG. 11 is a diagram showing the relationship between cloud thickness and the cloud top and cloud base altitudes. FIG. 12 is a flow chart for providing weather information including cloud appearance. FIG. 13 is a flowchart showing the process of identifying cloud appearances by the weather information processing device. FIG. 14 is a diagram for explaining the positional relationship in a planar image of clouds at different altitudes. FIG. 15 is a diagram for explaining the calculation of the cloud base altitudes of clouds with different heights by the weather information processing device according to the modified example of the first embodiment. FIG. 16 is a block diagram of a weather information processing device according to the second embodiment. FIG. 17 is a diagram for explaining an example of calculation of the movement direction and the movement speed. FIG. 18 is a flowchart of a calculation process of the moving direction and moving speed of clouds and cloud areas by the weather information processing device according to the second embodiment. FIG. 19 is a hardware configuration diagram of the weather information processing device.

Below, examples of the weather information processing device, weather information processing system, weather information processing method, and weather information processing program disclosed in the present application will be described in detail with reference to the drawings. Note that the weather information processing device, weather information processing system, weather information processing method, and weather information processing program disclosed in the present application are not limited to the following examples.

Figure 1 is a block diagram of a weather information providing system according to the first embodiment. Also, Figure 2 is a block diagram showing the details of the celestial camera and the weather information processing device. As shown in Figure 1, the weather information providing system 100 has a weather information processing device 1, a ground sensor 2, a weather information database 3, and an information utilization platform 4.

The ground sensor 2 is connected to the weather information processing device 1. As shown in FIG. 2, the ground sensor 2 has a celestial camera 21 and a weather data acquisition unit 22. The ground sensor 2 also has a measurement function for measuring the latitude and longitude of the location where the ground sensor 2 is installed and the time.

The ground sensor 2 is provided with a portion through which wind passes. The weather data acquisition unit 22 uses the wind passing through the ground sensor 2 to measure ground weather data including ground temperature, air pressure, and relative humidity.

The ground sensor 2 transmits the celestial image 201 captured by the celestial camera 21 and the measurement results by the meteorological data acquisition unit 22 to the weather information processing device 1. At this time, the ground sensor 2 also transmits the latitude and longitude of the location where the ground sensor 2 is installed, and the time of capture to the weather information processing device 1.

The celestial camera 21 captures the celestial image 201 shown in FIG. 3, which is a hemispherical image centered in the zenith direction, using visible light and infrared light at preset time intervals. FIG. 3 is a diagram showing an example of converting the celestial image into a planar image. Here, the celestial image 201 will be described in detail.

The celestial camera 21 has an image sensor 23 and a fisheye lens 24 as shown in FIG. 4. FIG. 4 is a diagram for explaining an outline of image acquisition in the celestial camera. The fisheye lens 24 is hemispherical in shape and captures light with a field angle of, for example, 180 degrees all around the circumference centered on the zenith. The light captured by the fisheye lens 24 is received by the image sensor 23 via an incident route where it is refracted. As a result, the image sensor 23 forms an image with a field angle of 180 degrees all around the circumference centered on the zenith.

Here, the celestial image 201 is an image taken from the ground, so the image is inverted east-west as shown in FIG. 3. The celestial image 201 shows the cloud base.

Returning to Figure 1, we will continue the explanation. The weather data distribution system 5 acquires meteorological satellite images taken by meteorological satellites 6 such as Himawari-8. The meteorological data distribution system 5 then stores the acquired meteorological satellite images. The meteorological satellite images include visible images, infrared images, and near-infrared images.

The weather information processing device 1 uses the celestial image 201 and ground weather data acquired from the ground sensor 2, and the satellite image acquired from the weather data distribution system 5 to identify clouds captured in the celestial image 201 and clouds captured in the satellite image. The weather information processing device 1 then acquires information on the appearance of clouds from the identified celestial image. The weather information processing device 1 then stores the acquired information on the appearance of clouds in the weather information database 3. The weather information processing device 1 is connected to the ground sensor 2, the weather information database 3, and the weather data distribution system 5. The weather information processing device 1 is described in detail below with reference to FIG. 2.

As shown in FIG. 2, the weather information processing device 1 according to this embodiment has an image analysis unit 101, a conversion unit 102, a cloud top height calculation unit 103, an image matching unit 104, a range acquisition unit 105, a cloud thickness calculation unit 106, a cloud base height calculation unit 107, and an appearance identification unit 108.

The image analysis unit 101 receives the celestial image 201 captured by the celestial camera 21 from the ground sensor 2. At this time, the image analysis unit 101 also receives the latitude and longitude of the installation location of the ground sensor 2, the capture time, and ground meteorological data from the ground sensor 2. The image analysis unit 101 also receives notification of the capture time of the celestial image 203 from the ground sensor 2. The image analysis unit 101 then acquires the meteorological satellite image of the notified capture time from the meteorological data distribution system 5. The meteorological satellite image acquired by the image analysis unit 101 includes a visible image 301 and an infrared image 302 as shown in FIG. 5. FIG. 5 is a diagram showing an example of a meteorological satellite image.

Now, referring to FIG. 5, we will explain meteorological satellite images. For example, we will explain the case where the meteorological satellite Himawari-8 is used as the meteorological satellite 6. In this case, the aerial resolution at the satellite's nadir when the meteorological satellite 6 captures a visible image 301 is 0.5 to 1.0 km. When the meteorological satellite 6 captures an infrared image 302, the aerial resolution is 1 to 2 km. When the meteorological satellite 6 captures a near-infrared image, the aerial resolution is also 2 km. The meteorological satellite 6 has three observation bands that generate the visible image 301, three bands that generate near-infrared images, and ten observation bands that generate the infrared image 302. The meteorological satellite 6 acquires the visible image 301 and the infrared image 302 at an observation time interval of 2.5 minutes.

Weather satellite images show cloud tops. For any point on the image, the satellite image provides a location on the ground expressed as latitude and longitude. Therefore, by using the satellite image, it is possible to determine the horizontal distance of clouds in the satellite image.

Visible image 301 is an image that observes the albedo, which is the reflection of sunlight. In visible image 301, areas with clouds appear white, and areas without clouds appear black. Visible image 301 does not show differences in the height of each cloud. Also, visible image 301 does not show differences in the thickness of the clouds.

On the other hand, the infrared image 302 is an image that represents the radiance temperature of the cloud. Here, the highest point of the cloud is called the cloud top. In the infrared image 302, clouds with high tops appear white, and parts with low tops appear black. In other words, the height of the clouds can be grasped by using the infrared image 302.

The brightness temperature difference image 303 on the far right side of the page in Figure 5 is an image of the difference between infrared images 302 taken using different observation bands. The brightness temperature difference image 303 appears white when the clouds are thick, and appears black when the clouds are thin. In other words, by obtaining the brightness temperature difference image 303 using the infrared image 302, the approximate thickness of the clouds can be grasped.

The image analysis unit 101 performs image analysis using meteorological satellite images, such as visible images 301 and infrared images 302, to determine whether there are clouds in each area and to identify individual clouds. The image analysis unit 101 also accumulates satellite images taken at each hour to grasp the time series changes of individual clouds.

The image analysis unit 101 also performs image analysis using the celestial image 201 to determine whether or not there are clouds in each area, identify each cloud, and specify the type of cloud. The image analysis unit 101 also accumulates the celestial image 201 for each hour to grasp the time series changes of each cloud. Furthermore, the image analysis unit 101 calculates the azimuth angle and zenith angle of each cloud from the ground sensor 2 using several points on the edge of each cloud in the identified celestial image 201. The zenith angle here refers to the angle from the zenith.

Furthermore, the image analysis unit 101 performs image analysis using either or both of the celestial image and the meteorological satellite image to identify the type of clouds shown in the celestial image and the meteorological satellite image. There are various methods for the image analysis unit 101 to determine the type of cloud. Some examples of the cloud type determination methods are described below.

For example, the image analysis unit 101 has a cloud type determination table 304 shown in FIG. 6. FIG. 6 is a diagram showing an example of the cloud type determination table. As shown in the determination table 304, the type of cloud can be identified from a combination of the brightness temperature and the brightness temperature difference when a predetermined band is used. B13 in the brightness temperature difference in FIG. 6 represents an infrared image in the 10.4 μm wavelength band, and B15 represents an infrared image in the 12.3 μm wavelength band. That is, the brightness temperature difference B13-B15 is the brightness temperature difference obtained from the difference between the infrared image in the 10.4 μm wavelength band and the infrared image in the 12.3 μm wavelength band. Then, the image analysis unit 101 obtains the brightness temperature and the brightness temperature difference from the infrared image 302. Then, the image analysis unit 101 identifies the type of cloud from the obtained brightness temperature and brightness temperature difference.

Additionally, the image analysis unit 101 performs deep learning using existing celestial images and satellite images and information on the types and conditions of clouds in each image to learn in advance the relationship between the celestial images and satellite images and the types and conditions of clouds. The image analysis unit 101 then uses the learning results with the celestial images and satellite images acquired when performing image analysis as input to identify the types of clouds shown in the celestial images and satellite images. In this method, the image analysis unit 101 may use either the celestial images or the satellite images, or both.

In the following, the image information including the cloud type determination results associated with each cloud obtained by the image analysis unit 101 through image analysis of the visible image 301 and the infrared image 302 is referred to as "satellite image analysis information". Also, the information obtained by the image analysis unit 101 through image analysis of the celestial image 201 is referred to as "celestial image analysis information". In this embodiment, both the satellite image analysis information and the celestial image analysis information are described as including the cloud type determination results.

The image analysis unit 101 outputs the celestial image 201 to the conversion unit 102. The image analysis unit 101 also outputs meteorological satellite images including a visible image 301 and an infrared image 302, satellite image analysis information, celestial image analysis information, and ground meteorological data to the image matching unit 104. At this time, the image analysis unit 101 also outputs the latitude and longitude of the ground sensor 2 and the shooting time of the celestial image 203 that was the basis of the planar image 202 to the image matching unit 104. Furthermore, the image analysis unit 101 outputs the ground meteorological data and the infrared image 302 to the cloud top height calculation unit 103.

The conversion unit 102 receives an input of a celestial image 201 from the image analysis unit 101. As shown in FIG. 4, the angle of view of the fisheye lens 24 approaches 180 degrees, i.e., the further away from the zenith, the greater the refractive index of the incident route. For this reason, the celestial image 201 captured using the fisheye lens 24 becomes more distorted the further away from the zenith. In contrast, the meteorological satellite image captured by the meteorological satellite 6 is an image captured as a plane. For this reason, in order to compare and match the celestial image with the meteorological satellite image, it is preferable to convert the celestial image 201 into an image captured as a plane. Therefore, the conversion unit 102 converts the celestial image 201 into a plane image 202 by expanding it into a plane, as shown in FIG. 4.

Here, an example of a method of planar development by the conversion unit 102 will be described. The conversion unit 102 accumulates in advance a plurality of images for adjustment obtained by photographing a plane for adjustment with the celestial camera 21. The conversion unit 102 then performs calibration using the plurality of images for adjustment as art book data, and performs planar development of the image photographed by the fisheye lens 24 based on the obtained curvature value, thereby obtaining optimal values of parameters for planar development of the celestial image 201. When the conversion unit 102 performs planar development of the celestial image 201, it cuts off the portions near the ends of the celestial image 201 that have large distortion. As a result, the range included in the planar image 202 generated by the conversion unit 102 is narrower than that of the celestial image 201. For example, the conversion unit 102 generates a planar image 202 with a zenith angle of 45 degrees centered on the zenith, with the zenith at 0 degrees. For example, if the zenith angle is 45 degrees and the altitude at which clouds exist in the sky is between 1 km and 10 km, then a circular image that falls within the zenith angle will be an image of clouds with a radius of 0.5 km to 5.0 km in horizontal distance on the ground.

However, here, as an example, the planar image 202 with a zenith angle of 45 degrees is used for explanation, but the zenith angle can be widened by further increasing the parameters and performing learning through high-dimensional calculations. For example, the conversion unit 102 can set the zenith angle of the planar image 202 to 60 degrees.

Then, the conversion unit 102 uses the parameters obtained by learning to expand the celestial sphere image 201 captured by the celestial sphere camera 21 onto a plane to generate a planar image 202. As shown in FIG. 4, the conversion unit 102 removes the image near the edge of the celestial sphere image 201 to generate a planar image 202 with a zenith angle of 45 degrees. Here, the planar image 202 generated by converting the celestial sphere image 201 is also an image with east-west inversion, just like the celestial sphere image 201. The conversion unit 102 outputs the generated planar image 202 to the image matching unit 104.

The image matching unit 104 receives inputs of meteorological satellite images including visible images 301 and infrared images 302, satellite image analysis information, celestial image analysis information, and ground meteorological data from the image analysis unit 101. The image matching unit 104 also receives inputs of the latitude and longitude of the ground sensor 2 and the capture time of the celestial image 203 that is the basis of the planar image 202 from the image analysis unit 101. The image matching unit 104 also receives inputs of the planar image 202 from the conversion unit 102.

Here, since the planar image 202 is an image of a range according to the zenith angle, the area captured varies depending on the height of the clouds. That is, if the clouds are at a high altitude, a wide area of the clouds will be captured in the planar image 202, and if the clouds are at a low altitude, a narrow area of the clouds will be captured in the planar image 202. Therefore, by estimating the cloud height to a certain extent, it is possible to grasp the range of the clouds captured in the planar image 202, and it becomes possible to compare the clouds captured in the visible image 301 with the clouds captured in the planar image 202.

The image matching unit 104 estimates the cloud altitude from the type of clouds shown in the visible image 301 and the celestial image 201 included in the satellite image analysis information and the celestial image analysis information. Next, since the planar image 202 is an image in which east and west are reversed, the image matching unit 104 reverses the east and west of the planar image 202 to align the orientations of the visible image 301 and the planar image 202. Next, the image matching unit 104 specifies in which area of the visible image 301 the planar image 202 is located, as shown in FIG. 7, from the latitude and longitude information of the ground sensor 2. FIG. 7 is a diagram showing an example of an image matching analysis process. Next, the image matching unit 104 uses the estimated cloud altitude to extract clouds with similar shapes, etc., within the same range in the area of the visible image 301 that includes the clouds shown in the planar image 202, as clouds shown in the planar image 202.

Then, as shown in FIG. 8, the image matching unit 104 enlarges, rotates, and translates the planar image 202 to match the clouds captured in the visible image 301 with the clouds captured in the planar image 202. FIG. 8 is a diagram for explaining the cloud matching process. Each frame represented by a dotted line in FIG. 8 represents the planar image 202 to which enlargement and translation have been applied. For convenience of explanation, FIG. 8 shows the planar image 202 without rotation, but in actual matching, the planar image 202 may be rotated. The image matching unit 104 is able to perform this matching because the distance of the clouds in the planar image 202 has been estimated.

In this way, the image matching unit 104 performs an image matching analysis process to identify the clouds captured in the planar image 202 by matching the clouds captured in the visible image 301. After that, the image matching unit 104 outputs the planar image 202 and the visible image 301 that have been subjected to the image matching analysis process to the range acquisition unit 105. The image matching unit 104 also outputs satellite image analysis information, celestial image analysis information, and ground meteorological data to the range acquisition unit 105. The image matching unit 104 also identifies the area in the infrared image 302 that corresponds to the planar image 202 from the correspondence between the visible image 301 and the planar image 202 that have been subjected to the image matching analysis process. The image matching unit 104 then outputs information on the area in the infrared image 302 that corresponds to the planar image 202 to the cloud top height calculation unit 103. Furthermore, the image matching unit 104 outputs the satellite image analysis information and the celestial image analysis information to the cloud thickness calculation unit 106.

The range acquisition unit 105 receives inputs of the planar image 202 and the visible image 301 that have been subjected to image matching analysis processing from the image matching unit 104. The range acquisition unit 105 also receives inputs of satellite image analysis information, celestial image analysis information, and ground meteorological data from the image matching unit 104. The range acquisition unit 105 can calculate accurate horizontal distances by using the visible image 301, which is a meteorological satellite image. Therefore, as shown in FIG. 8, the range acquisition unit 105 uses the visible image 301 to obtain the horizontal distance of clouds reflected in the planar image 202 that has been subjected to image matching analysis processing. By identifying clouds in the satellite image and clouds in the planar image of the celestial sphere, distance information is added to the planar image. Since it can be assumed that the altitude of the cloud base is the same for the same type of cloud, the range acquisition unit 105 can obtain information on the distance between any two points required to measure the scale (size) of the cloud. The distance LH in FIG. 8 corresponds to the horizontal distance of the cloud reflected in the planar image 202.

Then, the range acquisition unit 105 outputs the satellite image analysis information, the celestial image analysis information, and the ground meteorological data, together with information on the horizontal distance of the clouds shown in the planar image 202, to the cloud base height calculation unit 107.

The cloud base altitude calculation unit 107 obtains the azimuth angle and zenith angle of the edge of the cloud captured in the planar image 202 from the zenith of the ground sensor 2 from the zenith, from the celestial image analysis information. The cloud base altitude calculation unit 107 then calculates the cloud base altitude, which is the altitude of the bottom of the cloud, using the horizontal distance of the cloud captured in the planar image 202, and the azimuth angle and zenith angle of the edge of the cloud captured in the planar image 202 from the zenith of the ground sensor 2.

The cloud base altitude calculation unit 107 determines the zenith angle of the cloud from the azimuth angle. Figure 9 is a diagram for explaining the zenith angle and azimuth angle. α is the azimuth angle, which is the angle between a direction and due south, and is an angle measured around the west. In other words, due south is the direction with an azimuth angle of 0 degrees. Then, as shown in Figure 9, the cloud base altitude calculation unit 107 calculates the zenith angle θ from the three-dimensional coordinates representing the position of the cloud, which are expressed as (R sin θ cos α, R sin θ sin α, R cos θ).

Figure 10 is a diagram showing an example of calculating the cloud base height. Figure 10 shows a case where the cloud base height of the cloud directly above the ground sensor 2 is calculated. Here, a case where the zenith angle is θ degrees is explained. In this case, the cloud base height calculation unit 107 calculates the cloud base height using the following formula (1). H represents the cloud base height.

For example, if the horizontal cloud range in the planar image 202 is 18 km and the zenith angle is 45 degrees, the cloud base altitude calculation unit 107 calculates the cloud base altitude to be approximately 9 km.

Furthermore, the cloud base height calculation unit 107 calculates the lifting condensation height using the air temperature, air pressure, and relative humidity on the ground contained in the ground weather data acquired by the ground sensor 2. The cloud base height calculation unit 107 then calculates an estimate of the cloud base height from the calculated lifting condensation height. Here, the estimate of the cloud base height is an approximate cloud base height calculated from the calculated lifting condensation height. The cloud base height calculation unit 107 then uses the estimate of the cloud base height to verify and correct the cloud base height calculated from the visible image 301. The cloud base height calculation unit 107 then outputs the verified and corrected cloud base height to the cloud thickness calculation unit 106.

Returning to FIG. 2, the explanation will be continued. The cloud top height calculation unit 103 receives an input of the infrared image 302 from the image analysis unit 101. The cloud top height calculation unit 103 also receives an input of information on the area in the infrared image 302 that corresponds to the planar image 202 from the image matching unit 104. Next, the cloud top height calculation unit 103 calculates the temperature of the cloud top from the brightness of the clouds included in the area of the infrared image 302 that corresponds to the planar image 202. The cloud top height calculation unit 103 also acquires the ground temperature from the meteorological data acquisition unit 22.

Next, the cloud top height calculation unit 103 calculates the temperature difference between the temperature on the ground and the temperature at the cloud top. The cloud top height calculation unit 103 then calculates the cloud top height using the temperature lapse rate for the calculated temperature difference. For example, the temperature lapse rate of standard atmosphere is 6.5 K/km. The cloud top height calculation unit 103 then outputs the calculated cloud top height to the cloud thickness calculation unit 106.

For example, let us assume that the ground temperature measured by the ground sensor 2 is 20°C, and the cloud top temperature obtained from the infrared image 302 is 220K. In this case, the cloud top height calculation unit 103 calculates the cloud top height to be approximately 11km from 220-(273.15+20)=-73.15K/-6.5K=11.25.

The cloud thickness calculation unit 106 receives an input of the cloud base height from the cloud base height calculation unit 107. The cloud thickness calculation unit 106 also receives an input of the cloud top height from the cloud top height calculation unit 103. FIG. 11 is a diagram showing the relationship between the cloud thickness and the cloud top height and the cloud base height. In FIG. 11, distance H represents the cloud base height, distance H' represents the cloud top height, distance LT represents the cloud thickness, and distance LW represents the horizontal distance of the cloud, which corresponds to the scale of the cloud. As shown in FIG. 11, distance LT = distance H' - distance H. Therefore, the cloud thickness calculation unit 106 subtracts the cloud base height from the cloud top height to calculate the cloud thickness. For example, if the cloud top height is 11 km and the cloud base height is 9 km, the cloud thickness calculation unit 106 calculates the cloud thickness to be 2 km. The cloud thickness calculation unit 106 then outputs the calculated cloud thickness to the aspect identification unit 108.

Here, the cloud thickness calculation unit 106 may verify and correct the calculated cloud thickness using cloud thickness information that can be obtained from the satellite image analysis information and celestial image analysis information. For example, the cloud thickness calculation unit 106 obtains the satellite image analysis information and celestial image analysis information from the image matching unit 104. Next, the cloud thickness calculation unit 106 obtains an estimate of the cloud thickness from the cloud type information contained in the satellite image analysis information and celestial image analysis information. Then, the cloud thickness calculation unit 106 uses the estimate of the cloud thickness to verify and correct the cloud thickness calculated from the cloud base altitude and cloud top altitude.

The appearance identification unit 108 receives input of cloud thickness information from the cloud thickness calculation unit 106. Then, based on the cloud thickness information, the appearance identification unit 108 identifies the cloud appearance, such as the cloud scale LW, using satellite images, etc. Then, the appearance identification unit 108 transmits the cloud appearance photographed by the celestial camera 21 of the ground sensor 2 to the meteorological information database 3 for storage.

Returning to FIG. 1, we will continue the explanation. The weather information database 3 is a database that holds and manages various weather information. The weather information database 3 acquires and stores the cloud appearance photographed by the celestial camera 21 of the ground sensor 2 that is identified by the appearance identification unit 108 of the weather information processing device 1. The weather information database 3 also stores cloud base altitude, cloud top altitude, and cloud thickness. In addition, the weather information database 3 stores various weather information such as radar rainfall obtained from the Japan Meteorological Agency's high-resolution precipitation nowcast, lightning information obtained from the Japan Meteorological Agency's LIDEN (Lightning Detection Network system), and wind gust information obtained from the Japan Meteorological Agency's top database of tornadoes, etc.

By storing cloud type, cloud base altitude, cloud top altitude, cloud thickness, and cloud scale in the meteorological information database 3 and using this information to perform machine learning, even if one piece of information cannot be obtained, it is possible to supplement the missing information from other information. This information supplementation may be performed by the information utilization platform 4 or by another computer.

The information utilization platform 4 acquires various weather information, including cloud appearances, stored in the weather information database 3, and processes the information into a form that is easy to use in the various weather information utilization systems 7. For example, the information utilization platform 4 associates the type, size, and shape of clouds with information such as radar rainfall, lightning information, and wind gust information. This allows the information utilization platform 4 to process the various weather information into data that can be used to operate and manage various facilities and equipment. The information utilization platform 4 provides the processed data to the various weather information utilization systems 7 that utilize the weather information.

The various weather information utilization systems 7 include, for example, a solar power generation system 71, a power plant operation management system 72, a river management system 73, an air traffic control system 74, a mountain information provision system 75, a weather forecast system 76, and a weather information visualization system 77. However, the system shown in FIG. 1 is only an example, and other systems may be used as long as they are capable of utilizing weather information.

For example, the photovoltaic power generation system 71 can use the cloud aspects provided by the information utilization platform 4 to more accurately predict the output of photovoltaic power generation. The power plant operation management system 72 can use the cloud aspects provided by the information utilization platform 4 to more appropriately adjust the amount of power supply. The river management system 73 can use the cloud aspects provided by the information utilization platform 4 to more accurately predict rain, and can accurately predict river flooding and perform appropriate river management. The air traffic management system 74 can use the cloud aspects provided by the information utilization platform 4 to accurately grasp the cloud conditions on the route, and can more appropriately manage air traffic. The mountain information provision system 75 and the weather forecast system 76 can use the cloud aspects provided by the information utilization platform 4 to more accurately predict future weather and provide appropriate information. The weather information visualization system 77 can use the cloud aspects provided by the information utilization platform 4 to more accurately provide cloud images.

Next, the flow of providing weather information including cloud appearance will be described with reference to FIG. 12. FIG. 12 is a flowchart of providing weather information including cloud appearance.

The ground sensor 2 captures a celestial image 201 using the celestial camera 21. The ground sensor 2 also acquires ground meteorological data including ground temperature, air pressure, and relative humidity using the meteorological data acquisition unit 22 (step S1). The ground sensor 2 then transmits the celestial image 201 and the ground meteorological data to the weather information processing device 1.

The weather information processing device 1 receives a celestial image 201 and ground meteorological data from the ground sensor 2. The weather information processing device 1 also acquires meteorological satellite images including a visible image 301 and an infrared image 302 from the meteorological data distribution system 5. The weather information processing device 1 then uses the celestial image 201, the ground meteorological data, and the meteorological satellite images including the visible image 301 and the infrared image 302 to determine the thickness of clouds shown in the celestial image 201 and identify the state of the clouds (step S2).

The weather information processing device 1 stores the identified cloud appearance information in the weather information database 3 (step S3).

The information utilization platform 4 processes the various weather information, including information on cloud appearance stored in the weather information database 3, so that it can be easily used by the various weather information utilization systems 7 (step S4).

Then, the information utilization platform 4 provides the processed weather information to various weather information utilization systems 7 (step S5).

The various weather information utilization system 7 acquires the processed weather information from the information utilization base 4. The various weather information utilization system 7 then utilizes the processed weather information for weather forecasts, solar power generation output predictions, etc. (step S6).

Next, the flow of the cloud appearance identification process by the weather information processing device 1 will be described with reference to FIG. 13. FIG. 13 is a flowchart of the cloud appearance identification process by the weather information processing device.

The image analysis unit 101 acquires a celestial image 201 and meteorological data from the ground sensor 2 (step S101).

The image analysis unit 101 also acquires meteorological satellite images, including a visible image 301 and an infrared image 302, taken by a meteorological satellite 6 from the meteorological data distribution system 5 (step S102).

Next, the image analysis unit 101 performs image analysis of the celestial image 201, and obtains celestial image analysis information including a determination of the presence or absence of clouds in each area, identification of each cloud, identification of the type of cloud, time series changes of each cloud, and the azimuth angle and zenith angle of each cloud from the ground sensor 2. The image analysis unit 101 also performs image analysis of the meteorological satellite image, and obtains satellite image analysis information including a determination of the presence or absence of clouds in each area, identification of each cloud, identification of the type of cloud, and time series changes of each cloud (step S103). The image analysis unit 101 then outputs the celestial image 201 to the conversion unit 102. The image analysis unit 101 also outputs the meteorological satellite image including the visible image 301 and the infrared image 302, the satellite image analysis information, the celestial image analysis information, and the ground meteorological data to the image matching unit 104. At this time, the image analysis unit 101 also outputs the latitude and longitude of the ground sensor 2 and the capture time of the celestial image 203 that was the basis of the planar image 202 to the image matching unit 104. Furthermore, the image analysis unit 101 outputs the ground meteorological data and the infrared image 302 to the cloud top height calculation unit 103.

The conversion unit 102 receives an input of the celestial sphere image 201 from the image analysis unit 101. The conversion unit 102 then expands the celestial sphere image 201 onto a plane to convert it into a planar image 202 (step S104). The conversion unit 102 then outputs the generated planar image 202 to the image matching unit 104.

The image matching unit 104 receives inputs of meteorological satellite images including the visible image 301 and the infrared image 302, satellite image analysis information, celestial image analysis information, ground meteorological data, the latitude and longitude of the ground sensor 2, and the shooting time of the celestial image 201 that was the basis of the planar image 202 from the image analysis unit 101. Then, the image matching unit 104 acquires the type of cloud from the satellite image analysis information and the celestial image analysis information. Next, the image matching unit 104 estimates the cloud height from the acquired cloud type. Next, using the estimated cloud height, an estimate of the horizontal distance of the cloud contained in the planar image 202 is obtained from the estimated cloud height. Furthermore, the image matching unit 104 inverts the east-west of the planar image 202. Next, the image matching unit 104 identifies in which area of the visible image 301 the planar image 202 is located from the latitude and longitude information of the ground sensor 2, and extracts the clouds reflected in the planar image 202 from the visible image 301. Next, the image matching unit 104 performs parallel translation, rotation, zooming, etc. on the planar image 202, and matches the planar image 202 with the visible image 301. By performing such image matching analysis processing, the image matching unit 104 identifies clouds captured in the planar image 202 from among clouds captured in the visible image 301 (step S105). After that, the image matching unit 104 outputs the planar image 202 and the visible image 301 that have been subjected to the image matching analysis processing to the range acquisition unit 105.

The range acquisition unit 105 acquires the planar image 202 and the visible image 301 that have been subjected to the image matching analysis process from the image matching unit 104. Next, the range acquisition unit 105 uses the visible image 301 to determine the horizontal distance of the clouds reflected in the planar image 202 (step S106). After that, the range acquisition unit 105 outputs the horizontal distance of the clouds reflected in the planar image 202 to the cloud base altitude calculation unit 107.

The cloud base height calculation unit 107 receives input of the horizontal distance of the cloud captured in the planar image 202 from the range acquisition unit 105. The cloud base height calculation unit 107 then calculates the cloud base height of the cloud captured in the planar image 202 from the horizontal distance of the cloud using the azimuth angle and zenith angle of the edge of the cloud captured in the planar image 202 relative to the ground sensor 2 (step S107). At this time, the cloud base height calculation unit 107 calculates an estimate of the cloud base height by determining the lift condensation height from the ground temperature, air pressure, and relative temperature. The cloud base height calculation unit 107 then verifies and corrects the calculated cloud base height using the estimated cloud base height. The cloud base height calculation unit 107 then outputs the calculated cloud base height to the cloud thickness calculation unit 106.

The cloud top height calculation unit 103 receives the infrared image 302 and ground meteorological data from the image analysis unit 101. Next, the cloud top height calculation unit 103 obtains the ground temperature and cloud top temperature from the ground meteorological data and the infrared image 302 (step S108).

Next, the cloud top height calculation unit 103 calculates the cloud top height using the temperature lapse rate for the temperature difference between the ground temperature and the cloud top temperature (step S109). After that, the cloud top height calculation unit 103 outputs the calculated cloud top height to the cloud thickness calculation unit 106.

The cloud thickness calculation unit 106 receives an input of the cloud base altitude from the cloud base altitude calculation unit 107. The cloud thickness calculation unit 106 also receives an input of the cloud top altitude from the cloud top altitude calculation unit 103. The cloud thickness calculation unit 106 then subtracts the cloud base altitude from the cloud top altitude to calculate the cloud thickness (step S110). At this time, the cloud thickness calculation unit 106 may obtain an estimate of the cloud thickness using information on the cloud type contained in the satellite image analysis information and the celestial image analysis information, and verify and correct the calculated cloud thickness. The cloud thickness calculation unit 106 then outputs the calculated cloud thickness to the appearance identification unit 108.

The appearance identification unit 108 receives information on the cloud thickness from the cloud thickness calculation unit 106. Then, based on the cloud thickness, the appearance identification unit 108 identifies the cloud appearance reflected in the planar image 202, such as the type, size, and shape of the cloud, using the visible image 301, the infrared image 302, and the planar image 202 (step S111). After that, the appearance identification unit 108 stores the information on the identified cloud appearance in the weather information database 3.

As described above, the weather information processing device according to this embodiment converts a celestial image taken from the ground into a planar image, and identifies clouds that appear in the planar image from among the clouds that appear in the visible image. Furthermore, the weather information processing device uses the identification results to determine the horizontal distance of the clouds that appear in the planar image, and determines the cloud base height based on the horizontal distance. The weather information processing device also determines the cloud top height from the infrared image and ground temperature. The weather information processing device then calculates the cloud thickness from the determined cloud base height and cloud top height, and identifies the cloud appearance using the calculated cloud thickness.

This allows the weather information processing device to accurately identify the state of clouds, improving the performance of systems that utilize weather data. In addition, by accumulating cloud type, cloud base altitude, cloud top altitude, and cloud thickness in a weather information database and using this information to perform machine learning, even if one piece of information is not available, it is possible to compensate for the missing information from the other information. Therefore, even if sufficient information is not available, it is possible to provide appropriate weather information to various weather information utilization systems.

(Modification)
Next, a modified example of the first embodiment will be described. FIG. 14 is a diagram for explaining the positional relationship in a planar image of clouds at different altitudes. For example, as shown in FIG. 14, a case where a cloud C1 at a low altitude and a cloud C2 at a high altitude exist will be described. In the planar image 202, the difference in altitude is not expressed, and the clouds C1 and C2 at different altitudes are both displayed on the same plane. Therefore, it is also considered that the clouds C1 and C2 exist at the same altitude, and in that case, the apparent position of the cloud C2 is the same altitude as the cloud C1, and it appears to exist at the position of the cloud C21. At this time, if the cloud C2 is considered to be at the same altitude as the cloud C1, the distance from the zenith of the cloud C2 is calculated as the distance L1. However, in reality, the distance from the zenith of the cloud C2 is the distance L2, which is significantly different from the calculated distance L1. For example, if the altitude of cloud C1 is 1000 m, the altitude of cloud C2 is 10000 m, and the angle φ between the zenith and cloud C2 is 30 degrees, then distance L2 is 5800 m, while distance L1 is 580 m. In this way, when the cloud structure is complex, if the altitudes of the clouds shown in the planar image 202 are not taken into consideration and the clouds are calculated as being at the same altitude, an error will occur in the horizontal distance of the clouds.

Therefore, when the cloud structure is complex, the weather information processing device 1 according to this modification classifies clouds of different heights and calculates the horizontal distance for each. Below, we will explain how the weather information processing device 1 according to the modification calculates the horizontal distance of clouds. Figure 15 is a diagram for explaining the calculation of the cloud base altitude of each cloud of different heights by the weather information processing device according to the modification of Example 1.

The range acquisition unit 105 acquires information on each identified cloud type from the satellite image analysis information and the celestial image analysis information. The range acquisition unit 105 then classifies the clouds shown in the planar image 202 according to the cloud type, thereby classifying the clouds according to their altitude. Here, the range acquisition unit 105 may assign the same classification to clouds that exist at the same altitude. The range acquisition unit 105 then determines the horizontal distance of the cloud for each classification using the azimuth angle and zenith angle of the cloud edge from the zenith. The range acquisition unit 105 then outputs the horizontal distance of the cloud for each classification according to the cloud altitude to the cloud base altitude calculation unit 107.

For example, the range acquisition unit 105 classifies clouds into a cloud group to which cloud C1 belongs and a cloud group to which cloud C2 belongs, as shown in FIG. 15. The range acquisition unit 105 then determines the horizontal distance of the clouds in the group to which cloud C1 belongs as distance LH1. The range acquisition unit 105 also determines the horizontal distance of the clouds in the group to which cloud C2 belongs as distance LH2. In this way, the range acquisition unit 105 determines the horizontal distance for each of clouds C1 and C2, which are at different heights.

The cloud base height calculation unit 107 acquires the horizontal distance of the cloud for each classification according to the cloud height from the range acquisition unit 105. Then, the cloud base height calculation unit 107 calculates the cloud base height for each classification according to the cloud height. For example, as shown in FIG. 15, the cloud base height calculation unit 107 calculates the cloud base height of cloud C1 as distance H1. The cloud base height calculation unit 107 also calculates the cloud base height of cloud C2 as distance H2. Then, the cloud base height calculation unit 107 outputs the cloud base height for each classification according to the cloud height to the cloud thickness calculation unit 106.

Then, the cloud thickness calculation unit 106 receives an input of the cloud base altitude for each classification according to the cloud altitude from the cloud base altitude calculation unit 107. Then, the cloud thickness calculation unit 106 calculates the cloud thickness for each classification according to the cloud altitude using the cloud base altitude for each classification according to the cloud altitude.

In this embodiment, the case of identifying the appearance of each cloud at different altitudes has been described, but when focusing on a specific cloud, the weather information processing device 1 can acquire information on the specific cloud. For example, the range acquisition unit 105 may classify clouds according to their altitude, exclude clouds other than the specific cloud of interest, and perform horizontal calculations limited to classifications that include the specific cloud.

As described above, the weather information processing device according to this modified example classifies clouds according to altitude by classifying them using the cloud type. The weather information processing device then calculates the cloud thickness by determining the horizontal distance of the cloud for each classification according to the cloud altitude. This makes it possible to reduce errors in the horizontal distance of clouds at different altitudes captured in a planar image. In other words, even when the cloud structure is complex, it is possible to properly identify the appearance of each cloud, thereby improving the performance of a system that utilizes weather data.

Figure 16 is a block diagram of a weather information processing device according to Example 2. The weather information processing device 1 according to this example determines the direction and speed of movement of clouds and cloud areas in addition to identifying the state of clouds. Therefore, the following mainly describes the calculation of the direction and speed of movement of clouds and cloud areas by the weather information processing device 1. The weather information processing device 1 according to this example has a speed calculation unit 109 in addition to the components of Example 1. In the following description, the operation of the components similar to those of Example 1 will be omitted.

The image analysis unit 101 outputs the satellite image analysis information and the celestial image analysis information to the moving speed calculation unit 109. In addition, the image matching unit 104 outputs the planar image 202 and the visible image 301 that have been subjected to the image matching analysis process to the moving speed calculation unit 109.

The moving speed calculation unit 109 receives inputs of satellite image analysis information and celestial image analysis information from the image analysis unit 101. The moving speed calculation unit 109 also receives inputs of the planar image 202 and the visible image 301 that have been subjected to image matching analysis processing from the image matching unit 104.

Then, the moving speed calculation unit 109 uses the satellite image analysis information, the celestial image analysis information, and the planar image 202 and the visible image 301 that have been subjected to image matching analysis processing to trace the time series changes of individual clouds and cloud areas.Then, the moving speed calculation unit 109 uses the traced time series changes to calculate the moving direction and moving speed of individual clouds and cloud areas.

Figure 17 is a diagram for explaining an example of calculation of the moving direction and moving speed. For example, the moving speed calculation unit 109 extracts the cloud at the position 411 identified in the visible image 401 captured at 12:00. The image analysis unit 101 acquires the next satellite image in about 2.5 to 10 minutes. Then, the moving speed calculation unit 109 extracts the cloud at the position 412 identified in the visible image 402 captured at 12:10. Then, the moving speed calculation unit 109 determines that the cloud that existed at the position 411 has moved to the position 412 based on the type and shape of the cloud. Then, the moving speed calculation unit 109 calculates the moving speed and moving direction of the cloud represented by the arrow P1 based on the fact that the cloud has moved from the position 411 to the position 412 in 10 minutes.

Next, the movement speed calculation unit 109 calculates the future movement speed and position of each individual cloud or cloud area from the history of the movement direction and movement speed of each individual cloud or cloud area up to the present, and predicts the future position of each individual cloud or cloud area. For example, as shown in FIG. 17, the movement speed calculation unit 109 calculates the future movement direction and movement speed of the cloud from the history of the movement direction and movement speed of the cloud up to that point, including the movement direction and movement speed represented by the arrow P1. Then, the movement speed calculation unit 109 predicts positions 413 to 415, which are the future movement positions of the cloud.

Then, the movement speed calculation unit 109 transmits the history of the movement direction and movement speed of each cloud or cloud area to the weather information database 3 for storage. In addition, the movement speed calculation unit 109 transmits the predicted values of the movement speed and movement direction of each cloud or cloud area, as well as information on the predicted position of each cloud or cloud area, to the weather information database 3 for storage.

The information utilization platform 4 processes data using the direction and speed of movement of individual clouds and cloud areas, as well as future positions, etc. Then, the information utilization platform 4 provides the weather information processed using the direction and speed of movement of individual clouds and cloud areas, as well as future positions, etc. to the various weather information utilization systems 7.

Next, the flow of the calculation process of the moving direction and moving speed of clouds and cloud areas by the weather information processing device 1 will be described with reference to FIG. 18. FIG. 18 is a flowchart of the calculation process of the moving direction and moving speed of clouds and cloud areas by the weather information processing device according to the second embodiment.

The image analysis unit 101 acquires a celestial image 201 and meteorological data from the ground sensor 2 (step S201).

The image analysis unit 101 also acquires meteorological satellite images, including a visible image 301 and an infrared image 302, taken by a meteorological satellite 6 from the meteorological data distribution system 5 (step S202).

Next, the image analysis unit 101 performs image analysis of the celestial image 201, and obtains celestial image analysis information including a determination of the presence or absence of clouds in each area, identification of each cloud, identification of the type of cloud, time series changes of each cloud, and the azimuth angle and zenith angle of each cloud from the ground sensor 2. The image analysis unit 101 also performs image analysis of the meteorological satellite image, and obtains satellite image analysis information including a determination of the presence or absence of clouds in each area, identification of each cloud, identification of the type of cloud, and time series changes of each cloud (step S203). The image analysis unit 101 then outputs the celestial image 201 to the conversion unit 102. The image analysis unit 101 also outputs the meteorological satellite image including the visible image 301 and the infrared image 302, the satellite image analysis information, the celestial image analysis information, and the ground meteorological data to the image matching unit 104. At this time, the image analysis unit 101 also outputs the latitude and longitude of the ground sensor 2 and the capture time of the celestial image 203 that was the basis of the planar image 202 to the image matching unit 104. Furthermore, the image analysis unit 101 outputs the satellite image analysis information and the celestial image analysis information to the movement speed calculation unit 109.

The conversion unit 102 receives the input of the celestial sphere image 201 from the image analysis unit 101. The conversion unit 102 then expands the celestial sphere image 201 onto a plane to convert it into a planar image 202 (step S204). The conversion unit 102 then outputs the generated planar image 202 to the image matching unit 104.

The image matching unit 104 receives inputs from the image analysis unit 101 of meteorological satellite images including the visible image 301 and the infrared image 302, satellite image analysis information, celestial image analysis information, ground meteorological data, the latitude and longitude of the ground sensor 2, and the capture time of the celestial image 203 that is the basis of the planar image 202. The image matching unit 104 then performs image matching analysis processing to identify clouds captured in the planar image 202 from among the clouds captured in the visible image 301 (step S205). The image matching unit 104 then outputs the planar image 202 and the visible image 301 that have been subjected to image matching analysis processing to the movement speed calculation unit 109.

The moving speed calculation unit 109 receives input of satellite image analysis information and celestial image analysis information from the image analysis unit 101. The moving speed calculation unit 109 also receives input of the planar image 202 and visible image 301 that have been subjected to image matching and analysis processing from the image matching unit 104. The moving speed calculation unit 109 then uses the satellite image analysis information, celestial image analysis information, and the planar image 202 and visible image 301 that have been subjected to image matching and analysis processing to trace the time-series changes of individual clouds and cloud areas to determine the history of the moving direction and moving speed (step S206).

Next, the movement speed calculation unit 109 predicts the future movement of individual clouds and cloud areas from the history of the movement direction and movement speed (step S207). After that, the movement speed calculation unit 109 stores information on the predicted future movement of individual clouds and cloud areas in the weather information database 3.

As described above, the weather information processing device according to this embodiment uses satellite image analysis information, celestial image analysis information, and planar and visible images with cloud identification to determine the history of the movement direction and speed of individual clouds and cloud areas. In addition, the weather information processing device predicts the future movement of individual clouds and cloud areas from the history of the movement direction and speed of individual clouds and cloud areas. By adding this information and providing processed weather information to various weather information utilization systems, it is possible to provide more detailed and useful information, thereby improving the performance of systems that utilize weather data.

(Hardware configuration)
19 is a hardware configuration diagram of the weather information processing device. The weather information processing device 1 has a CPU (Central Processing Unit) 91, a memory 92, a hard disk 93, and a network interface 94. The CPU 91 is connected to the memory 92, the hard disk 93, and the network interface 94 via a bus.

The network interface 94 is an interface for communicating with, for example, the ground sensor 2, the weather information database 3, and the weather data distribution system 5.

The hard disk 93 stores various programs including programs for realizing the functions of the image analysis unit 101, conversion unit 102, cloud top height calculation unit 103, image matching unit 104, range acquisition unit 105, cloud thickness calculation unit 106, cloud base height calculation unit 107, appearance identification unit 108, and movement speed calculation unit 109 illustrated in Figures 2 and 16.

The CPU 91 reads out the various programs exemplified in Figs. 2 and 16 from the hard disk 93, and deploys and executes them on the memory 92. As a result, the CPU 91 and memory 92 realize the functions of the image analysis unit 101, conversion unit 102, cloud top height calculation unit 103, image matching unit 104, range acquisition unit 105, cloud thickness calculation unit 106, cloud base height calculation unit 107, appearance identification unit 108, and movement speed calculation unit 109 exemplified in Figs. 2 and 16.

REFERENCE SIGNS LIST 1 Weather information processing device 2 Ground sensor 3 Weather information database 4 Information utilization platform 5 Weather data distribution system 6 Meteorological satellite 7 Various weather information utilization systems 101 Image analysis unit 102 Conversion unit 103 Cloud top height calculation unit 104 Image matching unit 105 Range acquisition unit 106 Cloud thickness calculation unit 107 Cloud base height calculation unit 108 Aspect identification unit 109 Travel speed calculation unit 21 Celestial camera 22 Weather data acquisition unit

Claims (13)

an image matching unit that performs image matching by identifying the cloud distribution captured in the celestial image and the cloud distribution captured in the satellite image based on the celestial image and the satellite image received from a ground sensor having a camera that captures a planar celestial image centered in the zenith direction and a meteorological data distribution system that distributes satellite images captured by a meteorological satellite;
a range acquisition unit that acquires horizontal distance information in the celestial image from horizontal distance information in the satellite image based on a result of the image matching;
and a cloud base altitude calculation unit that calculates a cloud base altitude of the cloud based on the horizontal distance information in the celestial image and the zenith angle of the cloud in the celestial image. an image analysis unit that performs image analysis using either or both of the celestial image and the satellite image and identifies cloud types including cumulus, cumulonimbus, and cirrus;
a conversion unit for converting the celestial image into a planar image; and a cloud-top height calculation unit for calculating the cloud-top height based on the satellite image and the ground temperature.
A cloud thickness calculation unit that calculates the thickness of the cloud from the difference between the cloud base altitude and the cloud top altitude;
A cloud aspect identification unit that identifies a cloud aspect including a cloud scale based on the cloud thickness,
The weather information processing apparatus according to claim 1 , wherein the image matching unit executes the image matching based on the type of cloud identified by the image analysis unit. The weather information processing device according to claim 2, further comprising a movement information acquisition unit that acquires cloud movement information based on the type of cloud identified by the image analysis unit, the result of the image matching by the image matching unit, and the satellite image. The weather information processing device according to claim 2, characterized in that the cloud thickness calculation unit estimates the cloud thickness from the type of cloud identified by the image analysis unit, and verifies and corrects the calculated cloud thickness using the estimation result. The weather information processing device according to any one of claims 2 to 4, characterized in that the range acquisition unit classifies the clouds into altitudes corresponding to the individual cloud types based on the cloud types identified by the image analysis unit, and determines the horizontal distance of the clouds for each classification. The ground sensor further includes a meteorological data acquisition unit that acquires ground meteorological data including a ground temperature, an air pressure, and a relative humidity,
The weather information processing device according to any one of claims 1 to 5, characterized in that the cloud base altitude calculation unit calculates a lifting condensation altitude from the ground meteorological data acquired by the meteorological data acquisition unit, estimates the cloud base altitude based on the calculated lifting condensation altitude, and verifies and corrects the cloud base altitude using the estimated result. A weather information processing system having a ground sensor, a weather data distribution system, and a weather information processing device,
The ground sensor includes:
A camera is provided for capturing a flat celestial image centered on the zenith direction,
The meteorological data distribution system distributes satellite images taken by a meteorological satellite,
The weather information processing device includes:
an image matching unit that performs image matching by identifying a cloud distribution shown in the celestial image and a cloud distribution shown in the satellite image based on the celestial image and the satellite image received from the ground sensor and the weather data distribution system;
a range acquisition unit that acquires horizontal distance information in the celestial image from horizontal distance information in the satellite image based on a result of the image matching;
a cloud base altitude calculation unit that calculates a cloud base altitude of the cloud based on the horizontal distance information in the celestial image and the zenith angle of the cloud in the celestial image. The weather information processing device includes:
an image analysis unit that performs image analysis using either or both of the celestial image and the satellite image and identifies cloud types including cumulus, cumulonimbus, and cirrus;
a conversion unit for converting the celestial image into a planar image; and a cloud-top height calculation unit for calculating the cloud-top height based on the satellite image and the ground temperature.
A cloud thickness calculation unit that calculates the thickness of the cloud from the difference between the cloud base altitude and the cloud top altitude;
A cloud aspect identification unit that identifies a cloud aspect including a cloud scale based on the cloud thickness,
The weather information processing system according to claim 7 , wherein the image matching unit executes the image matching based on the type of cloud identified by the image analysis unit. The ground sensor further includes a meteorological data acquisition unit that acquires ground meteorological data including a ground temperature, an air pressure, and a relative humidity,
the cloud top altitude calculation unit calculates a cloud top altitude of the cloud based on the ground temperature acquired by the meteorological data acquisition unit;
The weather information processing system according to claim 8, characterized in that the cloud base altitude calculation unit calculates a lifting condensation altitude from the ground meteorological data acquired by the meteorological data acquisition unit, estimates the cloud base altitude based on the calculated lifting condensation altitude, and verifies and corrects the cloud base altitude using the estimated result. The weather information processing system according to claim 8 or 9, further comprising a weather information infrastructure that performs machine learning using information on the cloud type, the cloud base altitude, the cloud top altitude, the cloud thickness, and the cloud scale that has been previously acquired, and estimates remaining information using some of the information contained in the information on the cloud type, the cloud base altitude, the cloud top altitude, the cloud thickness, and the cloud scale based on the results of the machine learning. The weather information processing system according to claim 10, characterized in that the weather information life infrastructure provides the cloud appearance by associating it with other weather information and processing it. Based on the celestial image and the satellite image received from a ground sensor having a camera for capturing a planar celestial image centered in the zenith direction and a meteorological data distribution system for distributing satellite images captured by a meteorological satellite, the cloud distribution captured in the celestial image is identified with the cloud distribution captured in the satellite image, and an image matching is performed;
acquiring horizontal distance information in the celestial image from horizontal distance information in the satellite image based on a result of the image matching;
13. A weather information processing method comprising: calculating a cloud base altitude of the cloud based on the horizontal distance information in the celestial image and a zenith angle of the cloud in the celestial image. Based on the celestial image and the satellite image received from a ground sensor having a camera for capturing a planar celestial image centered in the zenith direction and a meteorological data distribution system for distributing satellite images captured by a meteorological satellite, the cloud distribution captured in the celestial image is identified with the cloud distribution captured in the satellite image, and an image matching is performed;
acquiring horizontal distance information in the celestial image from horizontal distance information in the satellite image based on a result of the image matching;
Calculating the cloud base altitude of the cloud based on the horizontal distance information in the celestial image and the zenith angle of the cloud in the celestial image.

Images