Vegetation management system and vegetation management method

Abstract

[Problem] To predict the impact of vegetation on a feature with high accuracy. [Solution] A vegetation management system for managing the impact of vegetation on a specified feature is characterized by comprising an acquisition unit that acquires remote sensing image data of the vegetation, a classification unit that classifies trees included in the vegetation according to growth activity that represents future growth potential based on the remote sensing image data, a growth prediction unit that predicts the growth of the trees based on the classification results by the classification unit, a risk assessment unit that determines the risk of contact with the specified feature, and a visualization unit that outputs and visualizes the assessment results by the risk assessment unit.

Description

[Technical Field]

[0001] The present invention relates to a vegetation management system and a vegetation management method. [Background technology]

[0002] Conventional vegetation management systems for maintaining power facilities predict vegetation growth to prevent contact with power lines. Detailed tree species information is essential for predicting vegetation growth.

[0003] Japanese Patent Laid-Open Publication No. 2019-144607 (Patent Document 1) describes a technology for estimating the specific tree species. This publication states that "the surface of the ground to be analyzed is photographed via satellite, a panchromatic image including trees is generated, trees are automatically extracted from the generated panchromatic image based on image features, the trees are surrounded by a circle, a co-occurrence matrix of the area inside the extracted circle is calculated, multispectral normalized data at the center of the circle is obtained, normalization processing is performed, training data is matched with the extracted trees, a tree species estimation model is created using a multivariate analysis model, and the tree species of the trees extracted from the panchromatic image of the ground surface to be analyzed is estimated using the analysis model."

[0004] Furthermore, in recent years, efforts are underway to automate survey work due to factors such as labor shortages. Many power transmission lines are installed in places that are difficult for people to access, such as mountainous regions. For this reason, remote sensing technology is attracting attention as a way to remotely monitor tree contact with distribution and transmission lines. Typical remote sensing methods include the use of artificial satellites and drones. Research and development is also underway to determine vegetation contact through three-dimensional measurement using LIDAR sensors.

[0005] Japanese Patent Laid-Open Publication No. 2016-123369 (Patent Document 1) describes a technology relating to vegetation growth assessment using remote sensing image data. This publication states, "A plant growth analysis system for analyzing plant growth based on remote sensing images, comprising: a plant growth model in which changes in feature amounts of the plant's growth are registered corresponding to time information indicating the elapsed time from a predetermined time for the plant; a feature amount calculation unit that calculates the feature amounts based on images corresponding to multiple growth locations of the plant that are part of the remote sensing images; a growth difference correction unit that refers to the plant growth model and corrects the feature amounts of the multiple growth locations to feature amounts corresponding to reference time information; and an image generation unit that generates new images so that the images corresponding to the multiple growth locations that are part of the remote sensing images exhibit the corrected feature amounts." [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2019-144607 A [Patent Document 2] Japanese Patent Application Laid-Open No. 2016-123369 Summary of the Invention [Problem to be solved by the invention]

[0007] In conventional power facility maintenance work, workers regularly conduct on-site inspections along the routes of power facilities (such as distribution and transmission lines), and in areas where problems may occur, perform tasks such as removing tree branches and using herbicides. During tree inspections, workers with specialized tree knowledge visit the site and inspect each tree one by one, identifying the tree's species and predicting future growth risks. If any trees that appear to be dangerous are discovered, they report them to a tree removal company. However, identifying the species of trees is extremely difficult, and even experts often have difficulty with this.

[0008] Methods using drones and helicopters are also beginning to be introduced. These methods involve developing platforms that handle large volumes of heterogeneous, time-series geographic information data, and providing support for operations such as operation and visualization. However, these methods become extremely costly when the shooting area is wide or the shooting period is frequent. There is also a method of generating a 3D map of trees by performing 3D reconstruction from data obtained by a LiDAR sensor. However, when analyzing vegetation using this 3D map, the characteristics of the data make it difficult to perform highly accurate analysis of vegetation classification and growth prediction. Furthermore, due to the difficulty of distinguishing vegetation, the reliability of the correct data is not high, making vegetation classification modeling itself difficult.

[0009] For these reasons, it has become an important issue to accurately predict the impact of vegetation on power facilities. This issue is not limited to power facilities, but also arises in the same way for other features. Therefore, the objective is to accurately predict the impact of vegetation on features. [Means for solving the problem]

[0010] In order to achieve the above-mentioned object, one representative vegetation management system of the present invention is a vegetation management system that manages the impact of vegetation on a specified feature, and is characterized by comprising an acquisition unit that acquires remote sensing image data of the vegetation, a classification unit that classifies trees included in the vegetation according to their growth activity, which represents their future growth potential, based on the remote sensing image data, a growth prediction unit that predicts the growth of the trees based on the classification results by the classification unit, a risk assessment unit that assesses the risk of contact with the specified feature, and a visualization unit that outputs and visualizes the assessment results by the risk assessment unit. Furthermore, one representative vegetation management method of the present invention is a vegetation management method using a vegetation management system that manages the impact of vegetation on a specified feature, characterized in that the vegetation management system includes the steps of: acquiring remote sensing image data of the vegetation; classifying trees included in the vegetation based on the remote sensing image data according to their growth activity, which represents their future growth potential; predicting the growth of the trees based on the classification results from the classification step; a risk assessment step that assesses the risk of contact with the specified feature; and outputting and visualizing the assessment results from the risk assessment step. [Effects of the Invention]

[0011] According to the present invention, the influence of vegetation on features on the ground can be predicted with high accuracy. Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a schematic diagram illustrating a configuration example of a vegetation management system for maintaining electric power facilities according to a first embodiment of the present invention. [Figure 2] 1 is a configuration diagram illustrating a hardware configuration of a vegetation management system according to a first embodiment. [Figure 3] 1 is a flowchart illustrating the overall processing of the system 100 according to the first embodiment. [Figure 4] 10 is a flowchart illustrating a process of a dead tree extraction unit in the first embodiment. [Figure 5] 10 is a flowchart illustrating the processing of a tree height estimation unit in the first embodiment. [Figure 6] 10 is a flowchart illustrating the processing of a crown extraction unit in the first embodiment. [Figure 7] 10 is a flowchart illustrating a process of a time series analysis unit in the first embodiment. [Figure 8] 10 is a flowchart illustrating the processing of the vegetation classification unit in the first embodiment. [Figure 9]10 is a flowchart illustrating a process of a growth prediction unit in the first embodiment. [Figure 10] 10 is a flowchart illustrating a process of a risk determination unit in the first embodiment. [Figure 11] FIG. 2 is a schematic diagram illustrating the processing of a dead tree extraction unit in the first embodiment. [Figure 12] 3 is a schematic diagram illustrating the processing of a tree height estimation unit in Example 1. FIG. [Figure 13] 3 is a schematic diagram illustrating the processing of a tree crown extraction unit in the first embodiment. FIG. [Figure 14] 3 is a schematic diagram illustrating the processing of the vegetation classification unit in the first embodiment. FIG. [Figure 15] 4 is a schematic diagram illustrating an intrusion risk process performed by a risk determination unit in the first embodiment. FIG. [Figure 16] 10 is a schematic diagram illustrating the toppling risk processing of the risk determination unit in the first embodiment. FIG. [Figure 17] 10 is a schematic diagram illustrating a jump-out risk process performed by a risk determination unit in the first embodiment. FIG. [Figure 18] FIG. 2 is a schematic diagram illustrating a database according to the first embodiment. [Figure 19] 10 is an example of a GUI schematic diagram showing visualization and maintenance instructions in the first embodiment. [Figure 20] 1 is a schematic diagram illustrating an example of a maintenance instruction according to the first embodiment. [Figure 21] 10 is a flowchart illustrating a process of a maintenance instruction unit in the first embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, an embodiment will be described with reference to the drawings. [Example]

[0014] A first embodiment of the present invention will be described below, in which a vegetation management system and method for power facility maintenance is applied. First, an overview will be given. The vegetation management system disclosed in the embodiment generates a height map of a vegetation area using measurement information such as remote sensing, geographic information, and environmental information, extracts tree crowns from the height map, classifies vegetation, predicts growth based on the classification results, and determines the risk of contact with and damage to power facilities. However, if detailed classification of tree species is used in the conventional way to predict vegetation growth, a bottleneck will arise because conventional classifications use the botanical order and family to which the tree belongs, making it difficult to distinguish between them. The disclosed system proposes a new tree classification method to solve the bottleneck of the past. Specifically, the disclosed system classifies trees based on their actual activity (the degree of potential for future growth). In the example, trees are classified into "dead trees (no activity)," "slow-growing trees (low activity)," "fast-growing trees (medium activity)," and "special-growth trees (high activity)" based on activity. Tree activity classification is estimated by analyzing remote sensing data and introducing machine learning models. Furthermore, in the present invention, the criteria for growth prediction and wire contact are varied depending on the activity level.

[0015] 1 is a block diagram showing the configuration of a vegetation management system 100 for power facility maintenance according to a first embodiment. The vegetation management system 100 for power facility maintenance includes a plurality of different data 101 (remote sensing image data 101a, vegetation information data 101b, geographic information data 101c, environmental information data 101d, and management information data 101e), a remote sensing data acquisition unit 102, a geographic information acquisition unit 103, an environmental information acquisition unit 104, a management information acquisition unit 105, a database generation unit 106, a dead tree extraction unit 107, a tree height estimation unit 108, a tree crown extraction unit 109, a time series analysis unit 110, a vegetation classification unit 111, and a growth prediction unit 112. The vegetation management system 100 also acquires time series remote sensing image data, vegetation information data, geographic information data, and environmental data for observing vegetation, and provides the data to a database 116. The remote sensing image data, vegetation information, geographic information data, and environmental data are explained in FIG. 1, and the database is explained in FIG.

[0016] The vegetation management system 100 provides data managed in the database 116 to the risk determination unit 113 to determine the risk of contact between the power facility and vegetation, and visualizes the contact risk using the visualization unit 114. The vegetation management system 100 also provides the information stored in the risk determination unit 113 and the visualization unit 114 to the maintenance instruction unit 115, which uses the information for facility maintenance support.

[0017] FIG. 2 shows the overall hardware configuration of the above-mentioned vegetation management system. A remote sensing observation device 120 captures the remote sensing images. The remote sensing observation device 120 is not limited to a specific type, and may be, for example, an observation satellite or an aircraft-mounted imaging device. This embodiment will be described using satellite images acquired by an observation satellite. A computer system 121 acquires remote sensing image data and operates as the vegetation management system 100 of FIG. 1. This computer system 121 has a typical hardware configuration including a CPU, RAM, and a storage unit.

[0018] Returning to FIG. 1, the explanation continues. The remote sensing image data 101a is image data obtained from a remote sensing sensor and is any aerial photograph. In this embodiment, the time-series low-resolution remote sensing images and the high-resolution remote sensing images are shown as multiple different remote sensing images, but the type of data is not limited to a specific type. Time-series images from the same remote sensing image may also be used. The difference in resolution is for relative comparison.

[0019] The vegetation information data 101b includes vegetation distribution information, spectrum information, type information, crown height information, etc., and the type of data is not limited. The vegetation distribution information includes information such as latitude and longitude, distribution shape and position. Spectral information is spectral and spatial information (color and shape) obtained by optical sensors. Optical sensors, in particular, are passive sensors that obtain information from light from the object being observed. Like a camera, they have a structure and function similar to that of the human eye, where an optical system (crystalline lens) such as a lens collects light from the object and forms an image on the detection system (retina). While the eye can only capture visible light, optical sensors can detect a wide range of wavelengths, from visible light to infrared. This allows for the acquisition of a wealth of useful information that cannot be interpreted by the human eye, such as the identification of minerals and vegetation, the temperature of the earth's surface, land use, and the water and plankton resources of oceans and lakes. This information can also be obtained as two-dimensional images over a wide area. Vegetation type and tree height data were obtained through on-site sampling and measurements.

[0020] Geographic information data 101c is essential data for sharing local data and satellite data, such as polygons and location information data. Examples of environmental information data 101d include soil data, meteorological data, elevation data, and slope data, but there are many others. Examples of meteorological data include AMeDAS, MODIS surface temperature satellite, and meteorological station data, but the types are not limited to these. Management information data 101e is operational data such as power company operation and maintenance notes, logs, and history. Other data may also be used.

[0021] The remote sensing data acquisition unit 102, geographic information acquisition unit 103, environmental information acquisition unit 104, management information acquisition unit 105, database generation unit 106, dead tree extraction unit 107, tree height estimation unit 108, tree crown extraction unit 109, time series analysis unit 110, vegetation classification unit 111, growth prediction unit 112, risk assessment unit 113, and visualization unit 114 are implemented as a combination of multiple CPUs and RAMs, each divided according to its role, to perform their respective calculation processes. Each of the above units uses an external storage device such as a hard disk or USB memory.

[0022] FIG. 3 is a flowchart illustrating an example of the overall processing of the vegetation management system 100 for power facility maintenance in the first embodiment. First, in S302, the vegetation management system 100 inputs remote sensing image data and vegetation information data, acquired as different data items, to the remote sensing data acquisition unit 102. Geographic information data is input to the geographic information acquisition unit 103. Environmental information data is input to the environmental information acquisition unit 104. In this embodiment, the environmental information mainly includes meteorological data, but may also include other data. The meteorological data includes temperature, precipitation, and solar radiation, but may also include other data. Management information data is input to the management information acquisition unit 105. In this embodiment, the management information data includes a monitoring log, a felling log, a maintenance location, and a time history, but may also include other data.

[0023] In S303, the vegetation management system 100 provides all of the input data to the database generation unit 106. The geographic information data includes location information and shape information, and masking of a location of interest in the remote sensing data is performed based on the location and shape. A masking part is extracted, and the remote sensing data and environmental data for that part are generated by the database generation unit and provided to the database 112. The management information data is provided to the database 116.

[0024] In S304, the vegetation management system 100 extracts dead tree areas from the remote sensing image data, vegetation geographic information data, and vegetation information data stored in the database 116, and provides the extracted dead tree area map to the database 116. The extraction method will be described in the description of the dead tree extraction unit 107.

[0025] In S305, the vegetation management system 100 uses the geographic information data of classified vegetation stored in the database 116, remote sensing image data representing a digital surface model, multi-band satellite image data, meteorological data, elevation model data, etc. to remove the dead tree area map extracted by the dead tree extraction unit 107 stored in the database 116, construct a model for estimating the crown height for the vegetation area, and provide the model and the crown height map for the vegetation area to the database 116. Details of the model construction and height estimation methods will be described later.

[0026] In S306, the vegetation management system 100 extracts the canopy of the vegetation using the canopy height map of the vegetation area stored in the database 116, and provides geographic information data of the canopy result to the database 116. Details of the extraction method will be described later.

[0027] In S307, the vegetation management system 100 calculates changes in tree crown and tree height using the results of executing the processes of S305 and S306 multiple times in a time series, and provides the calculation results to the database 116. Details of the time series analysis method will be described later.

[0028] In S308, the vegetation management system 100 classifies trees based on their growth activity using the crown changes and tree height changes stored in the database 116, and provides the geographic information data of the classification results to the database 116. Details of the classification method will be described later.

[0029] In S309, the vegetation management system 100 uses the tree species information stored in the database 116 to construct a different growth prediction model for each species to predict future growth, and provides the predicted future tree height and crown size to the database 116. Details of the growth prediction method for each species will be described later.

[0030] In S310, the vegetation management system 100 uses the predicted vegetation growth results, geographic information data on tree crown size and tree height data, and geographic information data on power facilities to evaluate the time-series fluctuations in the wide-area vegetation and the two-dimensional positional relationship of the power facilities using physical models for each of the risks of intrusion, protrusion, and falling, and evaluates the time-series fluctuations in vegetation tree height and the three-dimensional positional relationship of the power facilities. The vegetation management system 100 takes the evaluation results into consideration and determines the power facility risk for each classification based on tree growth activity, and provides the results to the database 116. Details of the risk determination model construction and calculation method will be described later.

[0031] In S311, the vegetation management system 100 maps the predicted vegetation growth results, the geographic information data of the tree crown size data and tree height data, the geographic information data of the power facility, and the geographic information data for the risk assessment results onto a remote sensing image. The time-series fluctuations in vegetation growth and the risk assessment results are displayed two-dimensionally or three-dimensionally at specific times or at designated time intervals. Details of the visualization will be described later.

[0032] In S312, the vegetation management system 100 presents maintenance instructions using the results of the analysis and the management information data stored in the database 116. All data is stored in the database. This data is shown in FIG. 18.

[0033] The remote sensing data acquisition unit 102 acquires remote sensing image data 101 a, accepts input of vegetation information data 101 b, and provides these to the database generation unit 106 .

[0034] The geographic information acquisition unit 103 acquires the geographic information data 101c and provides the acquired data to the database generation unit . The environmental information acquisition unit 104 acquires the environmental information data 101d and provides the acquired data to the database generation unit . The management information acquisition unit 105 acquires the management information data 101e and provides the acquired data to the database generation unit .

[0035] Next, the database generation unit 106 combines the remote sensing image data, vegetation information data, environmental data and geographic information data stored in the database 116 and maps them.

[0036] The dead tree extraction unit 107 extracts dead tree areas using the mapping results generated by the database generation unit 106. Fig. 4 is a flowchart illustrating the processing of the dead tree extraction unit 107 in the first embodiment.

[0037] In S402, the dead tree extraction unit 107 inputs the remote sensing image and the mapping data of geographic information generated by the database generation unit 106. In S403, the dead tree extraction unit 107 uses the remote sensing mapping data and vegetation information environmental data to generate a model for identifying areas corresponding to dead trees within the mapping range of the remote sensing image data. As an example of a dead tree area identification model, a machine learning technique is used to extract spectral information from a pixel-by-pixel target pixel, and texture information within a window range set within a certain range around the target pixel is calculated and used as a feature. Spectral information includes R, G, B, near-infrared, etc., and the spectral information varies depending on the remote sensing image used. Texture information refers to the texture, feel, and pattern of an object's surface. Texture analysis quantifies general textures such as rough, smooth, silky, and bumpy as a function of the spatial variation of pixel intensity across the image. An example of a calculation method is to first calculate a GLCM matrix, and then use the calculated GLCM matrix to calculate feature quantities such as entropy and energy. Texture information is not limited to GLCM, and other calculation methods can also be used. In this way, a model can be generated by learning three classifications: dead trees, healthy trees, and grass, using the feature values of the texture information within a certain range around the target pixel. Note that the classification items are not limited to these three classifications, and can include other classification items such as dead trees.

[0038] In S404, the dead tree extraction unit 107 extracts dead tree areas using the generated dead tree extraction model, and stores the extracted dead tree areas in the database 116. In S405, the dead tree extraction unit 107 converts the dead tree areas stored in the database 116 into geographic information data and provides the geographic information data to the database 116. FIG. 11 is a schematic diagram showing the dead tree area results. In FIG. 11, the dead tree areas are shown surrounded by white frames on the satellite image. In addition, FIG. 11 also shows areas surrounded by black frames. The black frames are dead tree areas added to the generated image by, for example, an operator.

[0039] The tree height estimation unit 108 excludes the dead tree areas using the generated geographic information data of the dead tree areas, estimates the crown height of the corresponding vegetation area using remote sensing data and multi-band satellite images representing the digital surface model stored in the database 116, and provides the tree height estimation model together with the estimated height to the database 116.

[0040] 5 is a flowchart illustrating the processing of the tree height estimation unit in Example 1. In S502, the tree height estimation unit 108 inputs remote sensing data multi-band satellite images representing a digital surface model stored in the database 116.

[0041] In S503, the tree height estimation unit 108 receives the generated geographic information data, environmental information data, and vegetation information data. In S504, the tree height estimation unit 108 receives the generated geographic information data of the dead tree area. In S505, the tree height estimation unit 108 excludes the dead tree area from the geographic information data of the corresponding dead tree area. In S506, the tree height estimation unit 108 generates a tree crown height estimation model using the generated remote sensing image data, geographic information data, and environmental data mapping. As an example of an estimation method, a random forest machine learning model may be used, with the value of the digital surface model as the objective variable and other spectral data, meteorological data, elevation data, vegetation data, etc. as the explanatory variables, to construct a tree crown height estimation model. The tree height estimation unit 108 estimates the tree crown height using the constructed tree height estimation model. Other prediction methods may also be used.

[0042] In S507, the tree height estimation unit 108 generates a tree height map using the generated model for estimating tree crown height and stores it in the database 116. Figure 12 is a schematic diagram showing the tree height estimation process. A mapping image showing tree height is generated from a normal satellite image, and changes in tree height are indicated by color changes from white to black.

[0043] The tree crown extraction unit 109 extracts the tree crowns using the tree height map in the vegetation area stored in the database 116. Fig. 6 is a flowchart illustrating the processing of the tree crown extraction unit in the first embodiment.

[0044] In S602, the crown extraction unit 109 inputs the tree height map within the vegetation area stored in the database 116. In S603, the tree crown extraction unit 109 detects the top points of trees using a tree height map of the vegetation area. The tree crown extraction unit 109 converts the detected top points into point geographic information data and stores the point geographic information data in the database 116. As an example of a method for detecting the top points of trees, the tree height map can be searched pixel by pixel, and the pixel point with the greatest height within a window range of a certain size can be set as the top point of the tree. Other methods for detecting the top points can also be used. The size of the search window is set in advance.

[0045] In S604, the tree crown extraction unit 109 uses the geographic information data of the top points of the trees stored in the database 116 to draw polygons of the tree crowns. In S605, the tree crown extraction unit 109 converts the drawn polygons into geographic information data and provides the geographic information data to the database 116. Figure 13 is a schematic diagram showing the tree crown extraction process. Fine tree crown polygons are extracted from the tree height map, converted into shape files, and mapped.

[0046] The time series analysis unit 110 analyzes the time series remote sensing image data generated by the database generation unit 106 using the canopy extraction method and vegetation area height estimation method stored in the database 116, and calculates changes in the canopy size and height of vegetation.

[0047] FIG. 7 is a flowchart illustrating the process of the time series analysis unit according to the first embodiment. In S702, the time series analysis unit 110 acquires the height at each time point in the time series. In S703, the time series analysis unit 110 calculates the crown size at each time point in the time series. In S704, the time series analysis unit 110 performs time series analysis of the calculated time series tree crown size data and time series height data. In S705, the time series analysis unit 110 calculates the change in height of the tree in question and provides the change in height to the database 116. In S706, the time series analysis unit 110 calculates the change in crown size of the tree in question and provides the change in crown size to the database 116.

[0048] The vegetation classification unit 111 uses the crown size change rate and height change rate stored in the database 116 to classify vegetation according to growth activity, which indicates future growth potential.

[0049] FIG. 8 is a flowchart illustrating the process of the vegetation classification unit in the first embodiment. In S802, the vegetation classification unit 111 inputs the crown size change rate and height change rate of the relevant tree stored in the database 116. In S803, the vegetation classification unit 111 determines the growth potential of the tree using the input crown size change rate and height change rate of the tree. An example of a method for determining the growth potential of a tree is shown below. If the calculated crown size change rate and height change rate of the tree show a tendency to decrease over time, the tree is likely to be reaching maturity and is therefore classified as a low-growth tree. On the other hand, if the calculated crown size change rate and height change rate of the tree show a tendency to increase over time, the tree is likely to be young and is therefore classified as a high-growth tree. Other methods for classifying trees into high-growth and low-growth may also be used.

[0050] In S804, the vegetation classification unit 111 extracts special growth tree areas using the felling records and felling location information data stored in the management database. In S805, the vegetation classification unit 111 generates a vegetation classification result based on growth activity using the dead tree areas extracted by the dead tree extraction unit 107, the special growth tree areas extracted in S804, and the high growth tree areas and low growth tree areas classified in S803, and provides this to the database 116. Figure 14 is a schematic diagram showing the vegetation classification process. Using the crown extraction results, each tree type is mapped in polygon form with a different color.

[0051] The growth prediction unit 112 uses the vegetation classification map stored in the database 116, the remote sensing image data, environmental data, and vegetation information data generated by the database generation unit 106 to predict future growth using different methods for each type, and provides the future crown size and height to the database 116.

[0052] FIG. 9 is a flowchart illustrating the growth prediction process according to the first embodiment. In S902, the growth prediction unit 112 inputs the vegetation classification map stored in the database 116. In S903, the growth prediction unit 112 introduces different growth prediction methods for different vegetation classifications. The growth prediction unit 112 refers to the vegetation classification map stored in the database 116, and calculates future growth as zero for trees classified as dead trees because they have no growth activity, and provides the calculation results to the database 116.

[0053] The growth prediction unit 112 generates a growth prediction model (S904) for trees classified as fast-growing or slow-growing trees and predicts their future growth. An example of a growth prediction model uses deep learning to estimate future growth. Features include a height map obtained by time-series analysis, spectral information calculated from remote sensing image data, and a vegetation index (e.g., NDVI) calculated from remote sensing image data. These are just a few examples, and any feature can be used. Features can also include maps of climate data, environmental data, and topographical data generated by the database generation unit 106. The target variables are the tree height and crown size after growth. These data are ground truth data obtained through field surveys, and the model is trained using them. Future height and crown size can be estimated from the height map obtained by time-series analysis, and further corrections can be made to the future height and crown size using deep learning. An example of a prediction model uses an RNN (recurrent neural network) to predict time-series fluctuations. The satellite images from the past times t1 and t2 are used to predict large-scale fluctuations at or after t3, and a tree height map is generated from the large-scale map at or after t3, and tree height changes and crown size are predicted. This is not a limitation, and other models may also be used.

[0054] For trees that fall under the category of special growth trees, the growth prediction unit 112 inputs the growth rules stored in the management database and simulates future growth using the prediction method of S905. In S906, the growth prediction unit 112 converts the predicted future tree crown size and height into a geographic information data format and provides the converted data to the database 116.

[0055] The risk determination unit 113 uses the predicted tree growth status, crown size fluctuation, and tree height fluctuation to consider the positional relationship with the power facility, determine the contact risk, and provide the contact risk to the database 116. The contact risk can be classified into different physical models. Examples of contact risks include intrusion risk, lodging risk, and jumping out risk. The risk determination unit 113 applies different physical models to different types of trees classified by the vegetation classification unit.

[0056] FIG. 10 is a flowchart illustrating the process of the risk determination unit in the first embodiment. In S1002, the risk determination unit 113 receives the generated vegetation classification map and the predicted crown size change and height change at the specific time. In S1003, the risk determination unit 113 performs risk determination using different physical models for different types of trees. The risk determination unit 113 refers to the vegetation classification map stored in the database 116, and determines the risk of emergence and the risk of lodging for trees that fall into the dead tree category. The risk determination unit 113 also determines the risk of invasion and the risk of lodging for trees that fall into the fast-growing tree and slow-growing tree categories. The risk determination unit 113 also determines the risk of invasion for trees that fall into the special growth tree category.

[0057] The risk of jumping out is explained in S1004, the risk of intrusion in S1006, and the risk of falling over in S1005. In S1004, the risk determination unit 113 determines the risk of dead trees flying in from the predicted tree crown size and height for the specific time. The risk of dead trees flying in is the risk of dead trees flying in due to strong winds or the like and coming into contact with power facilities. The risk of dead trees flying in is determined based on factors such as wind strength and direction, and in the case of mountains, the orientation and slope of the mountain. These parameters are used to perform a simulation of dead trees flying in, and if there is a high possibility of them coming into contact with power facilities, they are highlighted and visualized. The risk determination unit 113 converts the detected high-risk areas into geographic information data and provides the geographic information data to the database 116. FIG. 17 is a schematic diagram illustrating the jump-out risk processing of the risk determination unit in the first embodiment.

[0058] In S1005, the risk determination unit 113 determines the risk of tree falling based on the predicted tree crown size and height for the specific time. In this determination, the predicted changes in tree crown size and height and the relative positions of power distribution or transmission lines are used to estimate locations where there is a high probability of contact in the event of a tree falling. As an example of an estimation method, a three-dimensional model of the power distribution and transmission lines and vegetation is used to determine the risk of tree falling, taking into account the tree height and the area after falling. The risk determination method is not limited to this embodiment, and other methods may also be used. The risk determination unit 113 converts the detected high-risk areas into geographic information data and provides the geographic information data to the database 116. FIG. 16 is a schematic diagram illustrating the tipping risk processing of the risk determining unit in the first embodiment.

[0059] In S1006, the risk determination unit 113 determines the intrusion risk based on the predicted tree crown size and height for the specific time. In this determination, the predicted tree crown size and height fluctuations and the positional relationship of the power distribution line transmission lines are used to estimate locations where there is a high probability of contact based on the size of the tree crown. As an example of an estimation method, a three-dimensional model of the power distribution line transmission lines and vegetation is used to determine the intrusion risk, taking into account the tree height and the area after the crown changes. The risk determination method is not limited to this embodiment, and other methods may also be used. The risk determination unit 113 converts the detected high-risk areas into geographic information data and provides the geographic information data to the database 116. FIG. 15 is a schematic diagram illustrating an intrusion risk process performed by the risk determination unit in the first embodiment.

[0060] The visualization unit 114 performs processing to visualize the predicted vegetation crown size fluctuations, vegetation tree height fluctuations, and risk assessment results using the prediction results, which are stored in the database 116 .

[0061] The maintenance instruction unit 115 acquires and inputs the management information data stored in the database 116 . FIG. 20 is a flowchart illustrating the process of the maintenance instruction unit according to the first embodiment. In S1102, the maintenance instruction unit 115 inputs the remote sensing data stored in the database 116. The remote sensing data includes satellite images, drone images, vehicle-mounted camera images, etc., but other data may also be used. In S1103, the maintenance instruction unit 115 checks the risk assessment results using the input remote sensing data. Specifically, the maintenance instruction unit 115 checks the distribution of the risk assessment results visualized by the visualization unit 114 and selects a region of interest. The region of interest is selected using a threshold based on the quantification of the risk value, but other methods may also be used.

[0062] In S1104, the maintenance instruction unit 115 checks the detailed risk for the selected region of interest. The detailed risk judgment by the risk judgment unit and the detailed risk visualization by the visualization unit are acquired, and the detailed region is identified. To identify the detailed region, a threshold is used based on the quantification of the risk value, but other methods may also be used.

[0063] In S1105, the maintenance instruction unit 115 retrieves the selected detailed area, the contact risk assessment results, and the management information data from the database 116 and identifies high-risk locations requiring maintenance work as risk locations. The identified risk locations are then listed and the routes and times for dispatching personnel and transporting equipment are optimized. Figure 21 shows an example of a maintenance instruction diagram. The diagram shows the locations of an office and two equipment warehouses for four risk locations. Furthermore, the diagram searches for an optimal route, calculates the travel time and work time required for the optimal route, and displays the route and time instructions on the diagram. The maintenance instruction screen varies depending on the risk locations and the locations of the office and warehouses. The level or type of risk may also be used to search for the optimal route. For example, it is possible to search for a route that prioritizes eliminating the risk of tipping over.

[0064] The visualization unit 114 and the maintenance instruction unit 115 support maintenance work using displays on the management GUI. FIG. 19 is an example of a GUI schematic diagram showing visualization and maintenance instructions in the first embodiment. The screen shown in the figure has a display block showing changes in tree growth and changes in the risk of contact with power facilities, and can visualize the changes by moving along the time axis. It can also display a list of risk locations and display information such as latitude and longitude, risk level, and predicted time. The maintenance instruction block displays the maintenance instruction schematic diagram of FIG. 21. The detailed design and detailed items of the GUI may be different.

[0065] As described above, the disclosed vegetation management system is a vegetation management system that manages the impact of vegetation on a specified feature, and is characterized by comprising an acquisition unit that acquires remote sensing image data of the vegetation, a classification unit that classifies trees included in the vegetation according to growth activity that represents their future growth potential based on the remote sensing image data, a growth prediction unit that predicts the growth of the trees based on the classification result by the classification unit, a risk assessment unit that assesses the risk of contact with the specified feature, and a visualization unit that outputs and visualizes the assessment result by the risk assessment unit. According to this configuration and operation, by using a new classification method based on the growth activity of trees, it is possible to predict with high accuracy the influence of vegetation on features on the ground.

[0066] Furthermore, the disclosed vegetation management system varies the method used to predict the growth and / or determine the contact risk depending on the classification result. Therefore, by combining methods according to the growth activity of vegetation, it is possible to predict vegetation growth with high accuracy.

[0067] In addition, in the disclosed vegetation management system, the classification results include dead trees that have died and will not grow in the future, special growth trees that have been cut down and / or have had their branches removed, fast-growing trees that are evaluated to have high growth activity based on time-series changes in the remote sensing image data, and slow-growing trees that are evaluated to have low growth activity based on time-series changes in the remote sensing image data. In this way, by performing appropriate classification, vegetation growth can be predicted efficiently and with high accuracy.

[0068] In addition, the disclosed vegetation management system removes the dead trees from the prediction targets when predicting their growth, and uses a risk assessment model for lodging and jumping out when determining the risk of contact. The disclosed vegetation management system also predicts the growth of the fast-growing trees and the slow-growing trees, and uses a model for determining the risk of invasion and lodging that occurs due to changes in size caused by the predicted growth. The disclosed vegetation management system also uses a model for predicting growth of the special growth trees and determining the risk of invasion based on size changes. By performing growth predictions and risk assessments according to classification in this way, it is possible to predict with high accuracy the impact of vegetation on features.

[0069] The disclosed vegetation management system further includes a tree height estimation unit that estimates the canopy height from the vegetation area in the remote sensing image data, a crown extraction unit that extracts the canopy of the tree based on the estimated canopy height, and a time series analysis unit that calculates the time series changes in the size of the extracted canopy and the estimated canopy height, and the classification unit uses the time series changes to classify the trees according to their growth activity. In this way, by using the height and size of the tree crown, growth activity can be determined by using image processing of images taken from above.

[0070] The disclosed vegetation management system further includes a maintenance instruction unit that uses the growth activity of the trees and the positions of the predetermined features to formulate and present a maintenance plan. In addition, in the disclosed vegetation management system, the specified feature is an electric power facility, and the maintenance instruction unit formulates a work execution route for maintenance of the electric power facility's base equipment and power lines based on the growth activity of surrounding trees. This makes it possible to present efficient maintenance plans based on highly accurate predictions of the impact of vegetation. For example, it is possible to reduce maintenance costs by optimizing routes for introducing personnel and equipment in mountainous areas and other places that are difficult to access.

[0071] It should be noted that the present invention is not limited to the above-described embodiment, and includes various modifications. For example, the above-described embodiment has been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to an embodiment having all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, remove, or replace part of the configuration of each embodiment with other configurations. Furthermore, the above-described configurations, functions, processing units, processing means, etc. may be partly or entirely realized in hardware, for example, by designing them as integrated circuits. Furthermore, each of the above-mentioned configurations and functions may be realized by software, with a processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files that realize each function can be stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC (Integrated Circuit) card, SD card, or DVD (Digital Versatile Disc). In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and do not necessarily show all the control lines and information lines in the product. In reality, it can be assumed that almost all components are interconnected. [Explanation of symbols]

[0072] 100: Vegetation management system for power facility maintenance 101: Different input data 101a: Remote sensing image data 101b: Vegetation information data 101c: Geographical information data 101d: Environmental Information Data 101e: Management information data 102: Remote sensing data acquisition unit 103:Geographic information acquisition department 104:Environmental information acquisition department 105: Management information acquisition department 106: Database generation unit 107: Dead tree extraction part 108: Tree height estimation section 109: Tree crown extraction part 110: Time series analysis section 111: Vegetation Classification Section 112: Growth Forecasting Department 113: Risk Assessment Department 114: Visualization part 115: Maintenance instruction department 116: Database Department

Claims

1. 1. A vegetation management system for managing the impact of vegetation on a predetermined feature, an acquisition unit that acquires remote sensing image data of the vegetation; a classification unit that classifies trees included in the vegetation according to a growth activity that indicates a potential for future growth based on the remote sensing image data; a growth prediction unit that predicts the growth of the tree based on the classification result by the classification unit; a risk determination unit that determines a risk of contact with the predetermined feature; a visualization unit that outputs and visualizes the determination result by the risk determination unit; A vegetation management system comprising:

2. The vegetation management system according to claim 1, A vegetation management system characterized in that different methods are used to predict the growth and / or determine the contact risk depending on the classification results.

3. The vegetation management system according to claim 1, A vegetation management system characterized in that the classification results include dead trees that have died and will not grow in the future, special growth trees that have been cut down and / or have had their branches removed, fast growing trees that are evaluated to have high growth activity based on the time series changes in the remote sensing image data, and slow growing trees that are evaluated to have low growth activity based on the time series changes in the remote sensing image data.

4. The vegetation management system according to claim 3, For the dead tree, Regarding growth forecasts, they are removed from the forecasts, To determine the risk of contact, a risk determination model for falling and jumping out is used. A vegetation management system characterized by:

5. The vegetation management system according to claim 3, For the fast growing trees and the slow growing trees, Predict growth and use a model to assess the risk of invasion and lodging caused by changes in size due to predicted growth. A vegetation management system characterized by:

6. The vegetation management system according to claim 3, For said special growth trees, Use a model to predict growth and assess the risk of invasion based on size change A vegetation management system characterized by:

7. The vegetation management system according to claim 1, a tree height estimation unit that estimates a tree crown height from a vegetation area in the remote sensing image data; a crown extraction unit that extracts the crown of the tree according to the estimated crown height; a time series analysis unit that calculates time series changes in the extracted tree crown size and the estimated tree crown height, The vegetation management system is characterized in that the classification unit classifies trees according to the growth activity using the time-series changes.

8. The vegetation management system according to claim 1, A vegetation management system further comprising a maintenance instruction unit that uses the growth activity of the trees and the positions of the predetermined features to formulate and present a maintenance plan.

9. 9. The vegetation management system according to claim 8, the predetermined feature is an electric power facility, The vegetation management system is characterized in that the maintenance instruction unit formulates a work execution route for maintenance of the base equipment and power lines of the power facility based on the growth activity of surrounding trees.

10. A vegetation management method using a vegetation management system that manages the impact of vegetation on a predetermined feature, comprising: The vegetation management system comprises: acquiring remote sensing image data of the vegetation; a classification step of classifying trees included in the vegetation according to a growth activity that indicates a future growth potential based on the remote sensing image data; a step of predicting the growth of the tree based on the classification result obtained by the classification step; a risk determination step of determining a risk of contact with the predetermined feature; a step of outputting and visualizing the determination result obtained by the risk determination step; A vegetation management method comprising:

Images