Snow dropping damage range estimation device

Abstract

To provide a snow dropping damage range estimation device for accurately estimating a snow dropping damage range where damage caused by snow dropping from power transmission equipment occurs in a structure.SOLUTION: A snow dropping damage range estimation device 1 comprises: a model generation part 7 that performs machine learning using teacher data in which pre-snowfall images obtained by photographing, by photographing means 16a mounted on a flying body 16, a region on the ground including a structure before snowfall, meteorological conditions around the structure during snowfall, and characteristic elements of power transmission lines during snowfall are defined as input data and a snow dropping damage range is defined as output data, thereby generating a model; and an estimation part 8 that estimates the snow dropping damage range corresponding to estimation input data including post-snowmelt images obtained by photographing, by the photographing means 16a, the range after snow melting, meteorological conditions during a predetermined snowfall, and characteristic elements during the predetermined snowfall, based on the model.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a snow damage range estimation device that estimates the range of damage to a structure caused by falling snow, and more particularly to a snow damage range estimation device that estimates the range of damage caused by snow falling from power transmission equipment. [Background technology]

[0002] Traditionally, snow and ice that has adhered to power transmission equipment such as power lines and towers has fallen and damaged structures on the ground, such as greenhouses. In such cases, the power transmission and distribution company has compensated the owners of the structures. However, in reality, after the snow melts, it is difficult to determine whether the damage was caused by snow falling from the transmission equipment or by other factors. As a result, transmission and distribution companies are unable to accurately determine the need for compensation and the extent of such compensation, and they have difficulty determining the details of compensation when negotiating with owners. To determine whether damage to structures is due to snow falling from power transmission equipment, it is desirable to grasp the area where the snow falls on a map or image. The area where the snow falls depends on the amount of snow that has accumulated on the power transmission lines, as well as weather and topographical conditions. Therefore, in recent years, technology has been developed to predict snow accretion potential, which is proportional to the amount of snow accretion, based on weather conditions, although this is not for the purpose of compensation negotiations, and an invention related to this has already been disclosed.

[0003] Patent Document 1, entitled "Snow Accumulation Prediction Method and Snow Accumulation Prediction Program," discloses an invention relating to a snow accumulation prediction method that improves the accuracy of predicting the amount of snow that will accumulate on overhead wires. The invention disclosed in Patent Document 1 divides the target area for predicting snow accretion on overhead wires into meshes of any size, and predicts the precipitation, wind speed, wind direction, and temperature for each mesh up to any n hours later at any output time interval using a numerical weather forecast model based on weather data at the time of prediction.It calculates the snow accretion rate for each mesh using the predicted temperature for each mesh and a snow accretion rate function that shows the relationship between temperature and snow accretion rate that has been constructed in advance, and calculates the snow accretion potential for each mesh, which is proportional to the amount of snow accretion, based on the precipitation, wind speed, wind direction, and snow accretion rate, by wind direction. With this type of invention, it is possible to accurately predict the amount of snow accumulation for each mesh according to wind direction, thereby making it possible to predict in advance the areas and time periods where snow accumulation is likely to cause accidents. In addition, the snow accretion potential distribution, which displays statistical values ​​of snow accretion potential every few hours at corresponding locations on a map of the target area, allows users to determine at a glance which areas and at what time periods will experience the most snow accretion. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2008-157745 Summary of the Invention [Problem to be solved by the invention]

[0005] The invention disclosed in Patent Document 1 makes it possible to determine at a glance which areas and at what time periods will experience heavy snow accumulation. However, as mentioned above, the range in which snow will fall is thought to depend on the amount of snow that has accumulated on the power lines, weather conditions, etc., so the invention disclosed in Patent Document 1 does not go so far as to predict the range in which snow will fall.

[0006] The present invention has been made in response to the above-mentioned conventional circumstances, and aims to provide a snow fall damage range estimation device that can accurately estimate the snow fall damage range that causes damage to structures due to snow falling from power transmission equipment. [Means for solving the problem]

[0007] In order to achieve the above-mentioned object, the first invention is a snow damage range estimation device that estimates the range of snow damage to a structure on the ground caused by snow falling from power transmission equipment, and is characterized by having a model generation unit that performs machine learning to generate a model from training data that uses as input data pre-snow accumulation images obtained by an imaging means mounted on an aircraft to photograph an area on the ground including the structure before snow accumulation, the weather conditions at the time of snow accumulation around the structure, and characteristic elements of the electric wires that make up the power transmission equipment at the time of snow accumulation, and the snow damage range as output data, and an estimation unit that estimates the snow damage range based on the model, corresponding to estimation input data that includes post-snow melt images obtained by the imaging means to photograph the area after snow melts, the weather conditions at the time of a specified snow accumulation, and the characteristic elements at the time of the specified snow accumulation.

[0008] In the invention configured as above, the pre-snowfall images included in the input data of the training data are multiple images taken each time all of the past snowfalls have melted except for the past snowfalls (hereinafter referred to as the most recent snowfalls) immediately prior to the time when the snowfall damage range is to be estimated. The time of snowfall refers to the past snowfalls other than the most recent snowfalls. The post-snowmelt image included in the estimation input data is an image of the snow melted from the most recent snowfall. Furthermore, the predetermined snowfall is the most recent snowfall.

[0009] Furthermore, the snow damage range, which is output data of the training data, is a range that includes only damage to structures caused by snow falling from electric wires, and does not include ranges where damage occurs only due to factors other than snow falling from electric wires. Such snow damage range can be obtained, for example, by identifying the maximum range that includes multiple traces of fallen snow from an image taken when snow has accumulated on a structure. Note that an image taken when all the snow in the snow-covered image used to identify the maximum range has melted becomes the pre-snow image used as input data.

[0010] In addition, the extent of snow damage is thought to depend on the weather conditions at the time of snow accumulation and the characteristics of the power lines at the time of snow accumulation. Therefore, the model is designed to learn what kind of snow damage area is formed on a structure captured in a pre-snowfall image under the weather conditions during snowfall and the characteristic elements of the power lines during snowfall. Therefore, in the invention having the above configuration, the estimation unit can estimate, based on the model, what kind of snow damage range will be formed on a structure captured in an image after snow melting under the weather conditions at the time of a specified snowfall and the characteristic elements at the time of the specified snowfall.

[0011] The second invention is characterized in that, in the first invention, the weather conditions are outside temperature, humidity, amount of snowfall, first wind direction, and wind speed, and the characteristic elements are at least the wire thickness, the wire height above ground, the wire current, and the wire temperature. In the invention configured as described above, the weather conditions around the structure and the characteristics of the electric wire are factors that determine the amount of snow that adheres to the electric wire, the amount of snow that falls from the electric wire, and the direction and speed at which the snow falls from the electric wire. Therefore, the weather conditions and the characteristics of the electric wire are factors that determine the size and location of the area of ​​snow damage.

[0012] Of these, for example, observation data from the Japan Meteorological Agency is used for the weather conditions. Furthermore, the wire thickness and the wire height above ground are known. The wire temperature may also be included as a feature element. In this case, the wire temperature is calculated as a value that depends on, for example, the outside air temperature, the wire thickness, the wire current, etc. Alternatively, the wire temperature and wire current may be measured using a temperature sensor and a current sensor that are directly attached to the wire.

[0013] In the invention having the above configuration, in addition to the function of the first invention, the weather conditions that affect the size and location of the snow damage area and the types of characteristic elements of the electric wire are input in detail when generating the model and estimating the snow damage area, so that the estimated snow damage area reflects the actual weather conditions.

[0014] The third invention is characterized in that, in the first or second invention, the input data includes the topographical conditions at the time of snowfall, and the topographical conditions are the elevation of the area, a second wind direction of wind generated depending on the features surrounding the structure, and the direction in which the electric wires are stretched, and the estimation unit estimates the snowfall damage range corresponding to the estimation input data including the post-snowmelt image, the weather conditions at the specified time of snowfall, characteristic elements at the specified time of snowfall, and the topographical conditions at the specified time of snowfall.

[0015] In an invention of this configuration, the reason why terrain conditions are included when generating a model or estimating the extent of snow damage is that these terrain conditions are thought to change the weather conditions around the structure locally and over time, and ultimately affect the size and location of the estimated extent of snow damage. Furthermore, the features surrounding the structure refer to, for example, mountains, valleys, lakes, and rivers, which change a first wind direction to a different second wind direction, or which generate wind with a second wind direction in an area regardless of the wind with the first wind direction. Furthermore, the second wind direction may be determined based on observation data from the Japan Meteorological Agency, or may be determined based on actual measurements around the structure.

[0016] In the invention of the above configuration, in addition to the effects of the first or second invention, by including topographical conditions when generating the model or estimating the snow damage range, the snow damage range is estimated taking into account the weather conditions specific to the area.

[0017] The fourth invention is characterized in that, in the first or second invention, it includes a damage detection unit that detects pre-snow damage to a structure from pre-snow accumulation images, and the model generation unit uses as input data the pre-snow accumulation damage, weather conditions at the time of snow accumulation, and feature elements at the time of snow accumulation. In an invention with such a configuration, the model is generated by learning what kind of snow damage range was formed on the structure shown in the pre-snowfall image based on the actual damage that occurred before snow accumulation, the weather conditions at the time of snow accumulation, and the characteristic elements of the power lines at the time of snow accumulation. The damage that actually occurred before snow accumulation included damage caused by snow falling from power lines and damage caused by factors other than snow. As mentioned above, the range of snow damage only includes damage caused by snow falling from power lines.

[0018] Furthermore, in accordance with the model, when the estimation unit estimates, the damage detection unit detects post-snowmelt damage to the structure from the post-snowmelt image. Therefore, the estimation unit uses the post-snowmelt damage, the weather conditions at the time of a specified snowfall, and the feature elements at the time of the specified snowfall as input data for estimation.

[0019] In the invention of the above configuration, in addition to the effects of the first or second invention, the model learns the relationship between pre-snow damage, which includes damage caused by snow falling from electric wires as well as other causes, and the snow fall damage range, which includes only damage caused by snow falling from electric wires, taking into account weather conditions and characteristic elements of the electric wires.Therefore, the snow fall damage range corresponding to the input data for estimation does not include damage caused by causes other than snow falling from electric wires. [Effects of the Invention]

[0020] According to the first invention, the snow fall damage range used as output data for generating a model does not include the range where damage occurs only due to factors other than snow falling from electric wires, so the estimation unit can estimate the snow fall damage range that causes damage to a structure due to snow falling from electric wires.

[0021] According to the second invention, in addition to the effect of the first invention, the estimated snow damage range reflects actual weather conditions, etc., so that a highly accurate snow damage range can be obtained. Therefore, after the snow melts, it becomes possible to clearly determine whether the damage to the structure is due to snow falling from the electric wires or to other causes.

[0022] According to the third invention, in addition to the effects of the first or second invention, by including topographical conditions when generating a model or estimating the extent of snowfall damage, the extent of snowfall damage can be estimated taking into account the weather conditions specific to the area, making it possible to more accurately grasp the locations of damage to structures caused by snow falling from power transmission equipment.

[0023] According to the fourth invention, in addition to the effects of the first or second invention, the snow falling damage range corresponding to the estimation input data does not include damage caused by causes other than snow falling from the electric wires, so by comparing the snow falling damage range with damage after snow melting, damage caused by snow falling from the electric wires can be easily extracted. [Brief explanation of the drawings]

[0024] [Figure 1] 1 is a configuration diagram of a falling snow damage range estimation device according to an embodiment. [Figure 2] This is a process diagram of a model generation method executed by a model generation unit that constitutes the snow fall damage range estimation device of the embodiment. [Figure 3] 10 is a table for extracting the temperature of the electric wire. [Figure 4] 1 is a list of training data used in a model generation method executed by a model generation unit. [Figure 5] This is an image of the training data showing damage before snow accumulation and the extent of damage caused by falling snow. [Figure 6] This is a process diagram of the estimation method performed by the snow fall damage range estimation device of the embodiment. [Figure 7] This is a post-snowmelt image showing the damage caused by snowmelt and the estimated extent of damage caused by falling snow. DETAILED DESCRIPTION OF THE INVENTION [Example]

[0025] A falling snow damage extent estimating device according to an embodiment of the present invention will be described in detail with reference to Figures 1 to 8. Figure 1 is a configuration diagram of a falling snow damage extent estimating device according to an embodiment. As shown in Fig. 1, the snow damage extent estimation device 1 according to the embodiment is a device, specifically a computer, that estimates the extent of snow damage to a ground structure caused by snow falling from a power transmission facility. In this embodiment, the structure is assumed to be an agricultural greenhouse.

[0026] The falling snow damage range estimation device 1 includes an acquisition unit 2, an output unit 3, a control unit 4, and a storage unit 9. The acquisition unit 2 acquires pre-snowfall images and post-snowmelt images from the imaging means 16a mounted on the aircraft 16 via a wireless communication channel 17. Furthermore, the acquisition unit 2 acquires meteorological conditions used as input data for training data and meteorological conditions used as input data for estimation from a meteorological data server 18 via a network N. This meteorological data server 18 stores observation results by the Japan Meteorological Agency, such as outside temperature, humidity, snowfall, first wind direction, and wind speed in the area including the region.

[0027] In addition, the acquisition unit 2 acquires the topographical conditions used as input data for the teacher data and the topographical conditions used as input data for estimation from the topographical data server 19. This topographical data server 19 stores information on facilities owned by the power transmission and distribution company (such as the elevation of the area and the direction in which the power lines are laid) and actual measurement results such as second wind direction and wind speed measured by wind vanes and anemometers installed independently by the power transmission and distribution company. The output unit 3 is, for example, a display screen or a printer, and outputs the estimation result of the estimation unit 8 included in the control unit 4.

[0028] The control unit 4 is a central control device that controls the operations of the acquisition unit 2 through the storage unit 9 and the aircraft 16, and includes a damage detection unit 5, a table search unit 6, a model generation unit 7, and an estimation unit 8. Of these, the damage detection unit 5 detects pre-snow accumulation damage and post-snow melt damage to the structure from the pre-snow accumulation image, which is the input data of the training data acquired by the acquisition unit 2, and the post-snow melt image, which is the input data for estimation. This pre-snow accumulation damage and post-snow melt damage are, for example, holes or discolored areas formed in the roof of the structure, and are found by performing edge processing on the pre-snow accumulation image and the post-snow melt image, respectively. Specifically, the pre-snow accumulation damage and post-snow melt damage are expressed as two-dimensional coordinates indicating the position on the pre-snow accumulation image and the post-snow melt image, respectively.

[0029] The table search unit 6 searches for the input data of the teacher data and the wire temperature among the characteristic elements of the wire included in the input data for estimation. This wire temperature is a value that depends on the outside air temperature included in the weather conditions and the wire thickness, wire current, etc. included in the characteristic elements, and is calculated in advance for combinations of outside air temperature, wire thickness, wire current, etc., and stored as a table. Therefore, the table search unit 6 searches this table and extracts the corresponding wire temperature. The contents of the table will be explained with reference to FIG.

[0030] The model generation unit 7 performs machine learning using training data consisting of input data and output data to generate a model for estimating the extent of snow damage. Specifically, the model generation unit 7 uses the pre-snow damage detected by the damage detection unit 5 from the pre-snow accumulation image, the weather conditions at the time of snow accumulation, and the feature elements at the time of snow accumulation as input data, and uses the snow damage extent corresponding to this input data as output data. The snow damage area, which is the output data, is obtained by identifying the maximum area that contains multiple snowfall traces from an image of the structure when snow has accumulated on it. This identification can be done using, for example, a known method that can convert point cloud data into a surface.

[0031] The estimation unit 8 estimates the extent of snow damage corresponding to the input data for estimation based on the model generated by the model generation unit 7. In detail, the estimation unit 8 uses the post-snowmelt damage detected from the post-snowmelt image by the damage detection unit 5, the weather conditions at the time of snow accumulation, and the characteristic elements at the time of snow accumulation as input data for estimation, and estimates the extent of snow damage corresponding to this input data for estimation.

[0032] Next, the storage unit 9 includes a teacher data storage unit 10, a damage storage unit 11, a table storage unit 12, a model storage unit 13, an acquired data storage unit 14, and an estimation result storage unit 15. Of these, the teacher data storage unit 10 stores the pre-snowfall images acquired by the acquisition unit 2 and weather conditions, which are its input data. Furthermore, the teacher data storage unit 10 stores characteristic elements of the electric wire, which are input data of the teacher data. These characteristic elements include the electric wire temperature, which is input data extracted by the table search unit 6.

[0033] The damage storage unit 11 stores the pre-snow accumulation damage and post-snow melt damage detected by the damage detection unit 5 from the pre-snow accumulation image and the post-snow melt image, respectively. As described above, the table storage unit 12 stores a table of wire temperatures calculated in advance.

[0034] The model storage unit 13 stores the model generated by the model generation unit 7. The acquired data storage unit 14 stores the post-snowmelt images acquired by the acquisition unit 2, which are input data for estimation, and meteorological conditions. Furthermore, the acquired data storage unit 14 stores characteristic elements of the electric wire, which are input data for estimation. These characteristic elements include the electric wire temperature, which is input data for estimation and extracted by the table search unit 6.

[0035] The estimation result storage unit 15 stores the snow damage range estimated by the storage unit 9. Specifically, the snow damage range is a set of consecutive position coordinates that indicate its boundary line, and can be displayed superimposed on the post-snowmelt image.

[0036] Next, we will explain the steps of the model generation method executed by the model generation unit 7. Figure 2 is a process diagram of the model generation method executed by the model generation unit that constitutes the falling snow damage range estimation device according to the embodiment. 2, the model generation method 20 includes a training data acquisition step S21, a pre-snow damage detection step S22, a first conductor temperature search step S23, and a model generation step S24. Each step will be described below in order.

[0037] The teacher data acquisition process in step S21 is a process in which the acquisition unit 2 acquires as input data the pre-snowfall image acquired by the photographing means 16a before snowfall, the weather conditions around the structure when snow fell, the characteristic elements of the power lines when snow fell, and the topographical conditions when snow fell as well, and further acquires as output data the extent of damage caused by falling snow that has already been obtained.

[0038] The above weather conditions are the outside temperature, humidity, snowfall amount, first wind direction, and wind speed. The characteristic elements are the wire diameter, the wire height above the ground, and the wire current. In detail, the input data acquired in this process does not include the wire temperature, but as will be described later, this wire temperature is extracted by the table search unit 6 and used as a characteristic element. The topographical conditions are the elevation of the area, the second wind direction of the wind generated depending on the features around the structure, and the direction in which the electric wires are stretched. The acquired or extracted teacher data is then stored in the teacher data storage unit 10.

[0039] The pre-snow accumulation damage detection process of step S22 is a process in which the damage detection unit 5 detects pre-snow accumulation damage to the structure from the pre-snow accumulation image. As described above, pre-snow accumulation damage is holes formed in the roof of the structure or discolored areas. The detected pre-snow accumulation damage is stored in the damage memory unit 11.

[0040] The first wire temperature search step of step S23 is a step in which the table search unit 6 searches the table of wire temperatures stored in the table storage unit 12 and extracts the wire temperature corresponding to the input data. The extracted wire temperature is stored in the teacher data storage unit 10 as part of the input data, which is teacher data. The table is a list of calculated wire temperatures for each combination of outside air temperature, wire diameter, wire current, etc. The contents of the table will be explained using FIG.

[0041] The model generation step of step S24 is a step in which the model generation unit 7 performs machine learning from the input data and output data to generate a model. The generated model is stored in the model storage unit 13. Note that the machine learning is performed by a known learning method.

[0042] Next, the table used in the first electric wire temperature search process in step S23 will be described with reference to FIG. As shown in Figure 3, the wire temperature varies roughly depending on the allowable current and wire thickness of the wire. In addition, the wire temperature varies in detail depending on the outside temperature, which is one of the weather conditions, the wire current and wire thickness, which are characteristic elements of the wire, and each of several types of feature items (A, B, etc.). Specifically, the wire temperature is calculated using a formula for calculating the allowable current for each combination of the outside air temperature, the wire current, and other characteristic quantities. Here, examples of the feature amount include conductor resistance, thermal resistance, etc. These are known or pre-calculated values ​​for each material and structure of the conductor and resin coating that make up the electric wire.

[0043] Therefore, when the outside air temperature, the wire current, and the wire thickness are acquired as input data, the table search unit 6 searches for and extracts the corresponding wire temperature from among the multiple types of wire temperatures shown in the table.

[0044] Next, the training data used in the model generation method will be described with reference to Figures 4 and 5. Figure 4 is a list of training data used in the model generation method executed by the model generation unit. As shown in FIG. 4, from the left, the first column is an image before snowfall, the second to fourteenth columns are input data for the training data, and the fourteenth column is output data for the training data. Of these, the units [degrees] for the first wind direction, second wind direction, and direction of the power line are 360 ​​degrees, increasing clockwise from north as 0 degrees (=360 degrees). In addition, the damage before snow accumulation in the second column is displayed as a group of two-dimensional coordinates indicating each location, and the range of snow damage in the rightmost column is displayed as a group of two-dimensional coordinates indicating its boundary line.

[0045] Next, Figure 5 shows the damage before snow accumulation in the training data and an image before snow accumulation showing the extent of damage caused by falling snow. As shown in FIG. 5, the pre-snowfall image F1 includes a rectangular structure 50, an area 51 on the ground that includes the structure 50, and electric wires 60 that cross the sky above the structure 50. Among these, a plurality of structures 50 are present and are arranged in parallel within an area 51.

[0046] Furthermore, pre-snow damage is indicated by the reference numeral 52. This pre-snow damage 52 is obtained by plotting the two-dimensional coordinates (see the second column) of the pre-snow damage in Fig. 4 on the pre-snow damage image F1. As mentioned above, the pre-snow damage 52 includes damage caused by snow falling from electric wires and damage caused by other causes. Furthermore, the falling snow damage range is indicated by the reference numeral 53. This falling snow damage range 53 is obtained by plotting the two-dimensional coordinate group (see the rightmost column) of the falling snow damage range in FIG. 4 on the pre-snow accumulation image F1.

[0047] Next, the estimation method executed by the falling snow damage extent estimation device will be described with reference to Fig. 6. Fig. 6 is a process diagram of the estimation method executed by the falling snow damage extent estimation device according to the embodiment. 6, the estimation method 30 includes an estimation input data acquisition step S31, a post-snow melt damage detection step S32, a second electric wire temperature determination step S33, and an estimation step S34. Each step will be described below in order.

[0048] The estimation input data acquisition process of step S31 is a process in which the acquisition unit 2 acquires, as estimation input data, the post-snowmelt image acquired by the photographing means 16a after the most recent snow melts, the weather conditions at the time of the most recent snowfall from the weather data server 18, the characteristic elements at the time of the most recent snowfall, and the terrain conditions at the time of the most recent snowfall from the terrain data server 19. In detail, the input data for estimation acquired in this step does not include the wire temperature, but as will be described later, this wire temperature is extracted by the table search unit 6 and used as a characteristic element of the input data for estimation. The acquired or extracted input data for estimation is stored in the acquired data storage unit 14.

[0049] The post-snow melting damage detection process of step S32 is a process in which the damage detection unit 5 detects post-snow melting damage to the structure from the post-snow melting image. Post-snow melting damage is the same type of hole or discoloration as pre-snow melting damage. The detected post-snow melting damage is stored in the damage memory unit 11.

[0050] The second electric wire temperature determination step of step S33 is a step in which the table search unit 6 searches the table of electric wire temperatures stored in the table memory unit 12, as in the first electric wire temperature search step of step S23, and extracts the electric wire temperature corresponding to the estimation input data. The extracted electric wire temperature is stored in the acquired data storage unit 14 as part of the input data for estimation.

[0051] The estimation step of step S34 is a step in which the estimation unit 8 estimates the falling snow damage range corresponding to the input data for estimation based on the model. The falling snow damage range is stored in the estimation result storage unit 15.

[0052] Next, the extent of damage caused by falling snow will be explained using Fig. 7. Fig. 7 is a post-snowmelt image showing post-snowmelt damage and the estimated extent of damage caused by falling snow. As shown in FIG. 7, the post-snow melting image F2 displays post-snow melting damage 54, 54' and a snow damage range 55 estimated by the estimation unit 8. Of these, the post-snow melting damage 54 shown with diagonal lines is located inside the falling snow damage range 55, while the post-snow melting damage 54' shown with white lines is located outside the falling snow damage range 55. Therefore, the post-snow melting damage 54 is post-snow melting damage caused by snow falling from the electric wire 60, and the post-snow melting damage 54' is post-snow melting damage caused by something other than snow falling from the electric wire 60. The post-snow melting image F2 is output by the output unit 3.

[0053] As described above, the snow falling damage range estimation device 1 makes it possible to clearly determine whether the damage to the structure 50 after snow melting is due to snow falling from the electric wires 60 or to some other cause, depending on whether the post-snow melting damage 54 is located within the snow falling damage range 55, as shown in Figure 7. Therefore, the power transmission and distribution company can accurately present to the owner the post-snowmelt damage 54 caused by snow falling from the power lines 60. This will deepen the owner's understanding of the post-snowmelt damage 54, and it is expected that compensation negotiations will progress quickly.

[0054] Furthermore, the snowfall damage range 55 can be obtained without acquiring images from the most recent snowfall, so the burden of estimation is small. Furthermore, since the snow damage range 55 is estimated based on actual weather conditions, topographical conditions, etc., the post-snowmelt damage 54 caused by snow falling from the electric wire 60 can be grasped more accurately.

[0055] The snow damage range estimation device according to the present invention is not limited to the one shown in the embodiment. For example, if the current temperature is measured using a temperature sensor, the table search unit 6 and the table storage unit 12 are omitted. Furthermore, the first electric wire temperature search step of step S23 and the second electric wire temperature determination step of step S33 are also omitted. [Industrial Applicability]

[0056] The present invention can be used as a snow damage range estimation device that estimates the range of snow damage to a structure caused by snow falling from power transmission equipment. [Explanation of symbols]

[0057] 1...Snowfalling damage range estimation device 2...Acquisition unit 3...Output unit 4...Control unit 5...Damage detection unit 6...Table search unit 7...Model generation unit 8...Estimation unit 9...Memory unit 10...Teacher data memory unit 11...Damage memory unit 12...Table memory unit 13...Model memory unit 14...Acquired data memory unit 15...Estimation result memory unit 16...Air vehicle 16a...Photographing means 17...Wireless communication path 18...Weather data server 19...Topography data server 20...Model generation method 30...Estimation method 50...Structure 51...Area 52...Damage before snow accumulation 53...Snowfalling damage range 54, 54'...Damage after snowmelt 55...Snowfalling damage range 60...Electric wire

Claims

1. A snow damage range estimation device that estimates the range of snow damage caused by snow falling from power transmission equipment to a structure on the ground, a model generation unit that performs machine learning using training data that includes input data including a pre-snowfall image obtained by photographing the area on the ground including the structure before snowfall using an imaging means mounted on the aircraft, weather conditions around the structure when snow falls, and characteristic elements of the electric wires that constitute the power transmission facility when snow falls, and output data including the snow-fall damage range, to generate a model; A snow fall damage range estimation device characterized by having an estimation unit that estimates the snow fall damage range based on the model, corresponding to estimation input data including a post-snow melt image obtained by the photographing means by photographing the area after snow melting, the weather conditions at a specified time of snow accumulation, and the characteristic elements at the specified time of snow accumulation.

2. the weather conditions are an outside temperature, humidity, an amount of snowfall, a first wind direction, and a wind speed; 2. The device for estimating the extent of damage caused by falling snow according to claim 1, wherein the characteristic elements are at least the wire diameter, the wire height above the ground, and the wire current.

3. the input data includes terrain conditions during the snowfall; the topographical conditions include an elevation of the area, a second wind direction of wind generated in accordance with the surrounding features of the structure, and a tensioning direction in which the electric wire is tensioned; The snow damage range estimation device described in claim 1 or claim 2, characterized in that the estimation unit estimates the snow damage range corresponding to the estimation input data including the post-snowmelt image, the weather conditions at the specified time of snow accumulation, the characteristic elements at the specified time of snow accumulation, and the topographical conditions at the specified time of snow accumulation.

4. a damage detection unit that detects pre-snow accumulation damage to the structure from the pre-snow accumulation image, The snow damage range estimation device described in claim 1 or claim 2, characterized in that the model generation unit uses the damage before snow accumulation, the weather conditions at the time of snow accumulation, and the characteristic elements at the time of snow accumulation as input data.

Images