System and program for managing weed in premise of power facility

Abstract

To enable proper management of weeds in the premises of power facilities.SOLUTION: A system for managing weeds in the premises of power facilities comprises: a growth information acquisition task 35 in which growth information including the growth areas, kinds and growth situations of the weeds growing the premises of the power facilities is acquired on the basis of an image obtained by photographing the premises of the power facilities; a weed information database 341 in which at least one of a herbicidal method and a herbicidal period appropriate for each kind of the weeds is stored as weed information; and a management plan task 37 in which a management plan including the method and the period of weed removal to be performed in the future is formulated on the basis of the growth information and the weed information.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an electric power facility premises weed management system and an electric power facility premises weed management program for managing weeds within the premises of an electric power facility. [Background technology]

[0002] For example, weeds grow and overgrow within substations, and they need to be removed from the perspective of preserving the premises and the surrounding environment. For this reason, in the past, on-site surveys were conducted to determine the growth and overgrowth of weeds, and the timing and method of weeding were determined and an order was placed with a weeding company. However, surveys were required over a wide area of ​​the substation premises, and when managing multiple substations, the on-site surveys required a great deal of time and effort. In addition, there was a problem that people with little knowledge or experience of weeds and their removal did not know the optimal time or method for weeding.

[0003] On the other hand, a quantitative monitoring method for weeds that can effectively suppress the occurrence of weed damage is known (for example, see Patent Document 1). In this method, the number and coverage of weeds at the seedling stage are observed by multiple visual methods to determine the level of dominance of weeds, and the weeding method is changed depending on the dominance level and weed density. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Special Publication No. 2022-536301 Summary of the Invention [Problem to be solved by the invention]

[0005] Incidentally, substations sometimes have ballast (gravel) laid on the ground, which makes it unsuitable for weeding using a weeder, and substations adjacent to fields or residential areas have unique circumstances in which it is not possible to spray herbicides. In addition, substations are located in a variety of areas, and a variety of weeds grow there, so weeding must be done at the right time and with the right method for each type of weed, and weeds at multiple substations must be managed efficiently (weeding, etc.).

[0006] In contrast, the method of Patent Document 1 simply changes the weeding method depending on the dominance and density of weeds, and does not allow for planning appropriate weeding methods or timing based on the specific circumstances of a substation.

[0007] Therefore, an object of the present invention is to provide an electric power facility premises weed management system and an electric power facility premises weed management program that enable proper management of weeds within the premises of an electric power facility. [Means for solving the problem]

[0008] In order to solve the above problem, the invention of claim 1 is a weed management system for an electric power facility, comprising: a growth information acquisition means for acquiring growth information including the growth area, type, and growth status of weeds growing within the electric power facility based on images taken within the electric power facility; a weed information storage means for storing at least one of an appropriate weeding method and weeding timing for each type of weed as weed information; and a management planning means for formulating a management plan including the weeding method and timing to be carried out in the future based on the growth information and the weed information.

[0009] The invention of claim 2 is characterized in that, in the power facility premises weed management system described in claim 1, it is provided with a facility information storage means for storing surrounding information including whether or not there are fields around the power facility premises as facility information, and the management planning means formulates the management plan based on the growth information, the weed information, and the facility information.

[0010] The invention of claim 3 is characterized in that, in the power facility premises weed management system described in claim 1, it is provided with a facility information storage means for storing ground laying information, including whether the ground within the power facility premises is laid with ballast or not, as facility information, and the management planning means formulates the management plan based on the growth information, the weed information and the facility information.

[0011] The invention of claim 4 is characterized in that, in the power facility premises weed management system described in claim 1, it is provided with a herbicide information storage means for storing appropriate application targets and application times for each type of herbicide as herbicide information, and the management planning means formulates the management plan based on the growth information, the weed information, and the herbicide information.

[0012] The invention of claim 5 is characterized in that, in the power facility premises weed management system described in claim 1, it is provided with a management history storage means for storing a management history including the areas, methods and timing of weeding previously carried out on the power facility premises, and the management planning means formulates the management plan based on the growth information, the weed information and the management history.

[0013] The invention of claim 6 is characterized in that, in the power facility premises weed management system described in claim 1, it is provided with a weather information acquisition means for acquiring weather information within the power facility premises, and the management planning means formulates the management plan based on the growth information, the weed information and the weather information.

[0014] The invention of claim 7 is characterized in that, in the power facility premises weed management system described in claim 1, the growth information acquisition means uses a learning model for growth information that has been machine-learned based on past performance data so that when the image is input, the growth information is output.

[0015] The invention of claim 8 is characterized in that, in the power facility premises weed management system described in claim 1, the management plan means uses a learning model for the management plan that has been machine-learned based on past performance data so that when the growth information and the weed information are input, the management plan is output.

[0016] The invention of claim 9 is characterized in that, in the power facility premises weed management system described in claim 1, the growth information acquisition means acquires the growth information for each of the power facility premises based on images of the multiple power facility premises, and the management planning means formulates the management plan for the multiple power facility premises based on the growth information and the weed information for each of the power facility premises.

[0017] The invention of claim 10 is a weed management program for an electric power facility, characterized in that it causes a computer to function as: a growth information acquisition means for acquiring growth information including the growth area, type and growth status of weeds growing within the electric power facility based on images taken within the electric power facility; a weed information storage means for storing at least one of an appropriate weeding method and weeding timing for each type of weed as weed information; and a management planning means for formulating a management plan including the weeding method and timing to be carried out in the future based on the growth information and the weed information.

[0018] The invention of claim 11 is characterized in that, in the power facility premises weed management program described in claim 10, the growth information acquisition means uses a learning model for growth information that has been machine-learned based on past performance data so that when the image is input, the growth information is output.

[0019] The invention of claim 12 is characterized in that, in the weed management program for an electric power facility premises described in claim 10, the management plan means uses a learning model for the management plan that has been machine-learned based on past performance data so that when the growth information and the weed information are input, the management plan is output. Effect of the Invention

[0020] According to the inventions described in claims 1 and 10, a management plan is automatically formulated as to when and how to remove weeds based on growth information including the growth area, type, and growth status of weeds growing within the power facility, and weed information including a weeding method and weeding time suitable for the type of weed. This makes it possible to remove weeds within the power facility at the right time and in the right way, that is, to properly manage weeds, without relying on human knowledge or experience. Moreover, since the management plan is formulated based on the weeding method and weeding time suitable for the type of weed growing within the power facility, it becomes possible to formulate an appropriate management plan for the power facility, which exists in a variety of regions and where a variety of weeds grow.

[0021] According to the invention described in claim 2, a management plan is formulated based on surrounding information such as whether there are fields around the electric power facility, making it possible to formulate an appropriate management plan for electric power facility premises that are located in various regions and may have various restrictions depending on the surrounding environment. For example, for an electric power facility premises that has fields around it, it is possible to formulate a management plan that uses a weed control method that does not use herbicides, thereby reducing the impact on the fields.

[0022] According to the invention described in claim 3, a management plan is formulated based on ground construction information such as whether the ground in the electric power facility premises is laid with ballast or not, so it is possible to formulate an appropriate management plan for electric power facility premises that are located in various regions and have various ground constructions. For example, for an electric power facility premises with a ballast ground, a management plan can be formulated using a weeding method that does not use a weeder, and appropriate weeding can be performed.

[0023] According to the invention described in claim 4, a management plan is formulated based on the appropriate target and timing of use for each type of herbicide, making it possible to formulate a management plan suitable for the herbicide used and to carry out appropriate weeding.

[0024] According to the invention described in claim 5, a management plan is formulated based on the management history of when, where, and how weeding was carried out, making it possible to formulate an appropriate management plan that takes into account past weeding actions.

[0025] According to the invention described in claim 6, since a management plan is formulated based on meteorological information within the electric power facility premises, it is possible to formulate an appropriate management plan suited to the weather within the electric power facility premises.

[0026] According to the inventions described in claims 7 and 11, growth information is output using a machine-learned learning model for growth information, making it possible to obtain more appropriate growth information and formulate a more appropriate management plan.

[0027] According to the inventions described in claims 8 and 12, a management plan is output using a machine-learned learning model for a management plan, making it possible to formulate a more appropriate management plan.

[0028] According to the invention described in claim 9, a management plan for multiple power facility premises is formulated based on weed growth information for each of the multiple power facility premises, making it possible to formulate an appropriate and efficient management plan for each of the multiple power facility premises. [Brief description of the drawings]

[0029] [Figure 1] 1 is a schematic diagram showing a weed management system for an electric power facility according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a schematic block diagram showing the configuration of a weed management computer of the electric power facility premises weed management system of FIG. 1. [Diagram 3] FIG. 3 is a functional block diagram showing a schematic configuration of a growth information learning model of the weed management computer of FIG. 2. [Figure 4] FIG. 3 is a functional block diagram showing a schematic configuration of a learning model for management planning in the weed management computer of FIG. 2. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0030] Hereinafter, the present invention will be described based on the illustrated embodiment.

[0031] Fig. 1 is a schematic diagram showing an electric power facility premises weed management system 1 according to an embodiment of the present invention. This electric power facility premises weed management system 1 is a system for managing weeds W within the electric power facility premises, and in this embodiment, a case will be described in which the electric power facility premises is a substation premises G, but the electric power facility premises may be within the premises of other electric power facilities. Also, it is assumed that a plurality of substations are provided in different places and regions.

[0032] The power facility premises weed management system 1 includes a monitoring drone 2 and a weed management computer 3.

[0033] The surveillance drone 2 is an unmanned aerial vehicle that photographs the substation premises G while flying within the substation premises G, and in this embodiment, it automatically flies a predetermined flight route / flight from a predetermined takeoff and landing point within the substation premises G. That is, it is equipped with a GPS (Global Positioning System), and flies autonomously along a preset flight route while checking its own position with the GPS. At this time, it may be equipped with a function to avoid objects that become obstacles to flight depending on the flight environment, etc.

[0034] The surveillance drone 2 also includes a camera (imaging means) 21, and hovers at a preset imaging point to capture images of the ground side and weeds W side of the substation premises G. That is, the drone 2 hovers at a plurality of imaging points set on the flight route, starts up the camera 21, captures images of predetermined locations and imaging points in the substation premises G, and transmits the captured images together with location information of the imaging location (latitude and longitude corresponding to the imaging point) to the weed management computer 3. As will be described later, if possible, the entire substation premises G may be captured in one imaging session, in which case transmission of location information is not necessary.

[0035] Here, multiple such surveillance drones 2 may be provided and flown simultaneously along different flight routes, or instead of automatic flight, a person may remotely control the drones. Also, the surveillance drone 2 may continuously capture images of the ground side with the camera 21 while flying along the flight route, and the captured images and flight position (latitude, longitude) at each time may be stored sequentially, so that it is possible to know where each image is from.

[0036] In this embodiment, such a monitoring drone 2 is periodically started and flown, and periodically captures images of the ground (weeds W) and transmits them to the weed management computer 3. The received images are then stored in the weed management computer 3, and the stored images are used and referenced when a management planning task 37, which will be described later, is started. Alternatively, the monitoring drone 2 may be started and flown when the management planning task 37 is started, and the latest current state of the ground (weeds W) may be captured and the images may be transmitted to the weed management computer 3.

[0037] The weed management computer 3 is a computer that formulates a management plan for weeds W, and in this embodiment is located in a management center G that manages a plurality of substation premises G. As shown in Fig. 2, this weed management computer 3 mainly comprises an input unit 31, a display unit 32, a communication unit 33, a memory unit 34, a growth information acquisition task (growth information acquisition means) 35, a weather information acquisition task (weather information acquisition means) 36, a management plan task (management plan means) 37, a learning task 38, and a central processing unit 39 that controls these.

[0038] The input unit 31 is an interface for inputting various information and commands, and specifically, inputs a management history to a management history database 344 described later, and inputs a command to start a management plan task 37. The display unit 32 is a display for displaying various data and information, and specifically, displays growth information acquired and generated in a growth information acquisition task 35, and displays a management plan formulated in a management plan task 37. The communication unit 33 is an interface for communicating with the outside via the Internet network, a telephone communication network, and the like, and specifically, receives images from the monitoring drone 2, and receives weather information from a weather information server 4 described later.

[0039] The memory unit 34 mainly includes a weed information database (weed information storage means) 341, a herbicide information database (herbicide information storage means) 342, a facility information database (facility information storage means) 343, a management history database (management history storage means) 344, a learning model 345, and a performance database 346. Here, the databases 341, 342, 343, and 344 will be explained, and the learning model 345 and the performance database 346 will be explained later. As will be explained later, the learning model 345 includes a growth information learning model 345A and a management plan learning model 345B, and the performance database 346 includes a growth information performance database 346A and a management plan performance database 346B.

[0040] The weed information database 341 is a database that stores at least one of an appropriate weeding method and a suitable weeding time for each type of weed W as weed information. That is, in this embodiment, for each type of weed W, sample images of the weed W for each growth state (germination state, leafy state, etc.) and whether the weed W is an annual weed or a perennial weed are stored. Furthermore, in the case of annual weeds, it is preferable to weed them before they produce seeds (within a specified period after they flower), and in the case of perennial weeds, it is stored that the weeding method is to remove them by the roots. Here, the types of weeds W stored are those that are generally assumed to grow naturally in the substation premises G.

[0041] The herbicide information database 342 is a database that stores, as herbicide information, the target and timing of application for each type of herbicide. That is, it stores information on which type of weed W, what type of growth condition of the weed W each herbicide is suitable for, when it is suitable to use the herbicide, or when it may be used. Specifically, the following herbicide information is stored.

[0042] "Soil treatment agent" is sprayed on the soil surface to suppress the germination of weeds W or kill them immediately after germination. The best time to use is in early spring (February to March) and in autumn (September to October). "Foliage treatment agent" is applied directly to the leaves and stems of already growing weeds (W) to kill them. The best time to use it is from April to October, when dead weeds need to be removed. "Non-selective herbicide" - kills all plants it comes into contact with "Selective herbicides" are those that kill only certain plant species. The facility information database 343 is a database that stores geographical information and structural information (facility information) related to weeds W and weeding in each substation premises G. That is, information about the surroundings of the substation premises G (surrounding information), such as whether there are fields or houses around the substation premises G, and information about the ground of the substation premises G (laying information), such as whether the ground of the substation premises G is covered with ballast or soil, are stored as facility information. In addition, location information, such as whether the substation premises G is in a mountainous area or an urban area, is stored as facility information.

[0043] The management history database 344 is a database that stores management history including the area (including size), method, and time of weeding that was previously performed in each substation premises G. In other words, for each weeding, management history is stored, such as where in the substation premises G, how much area, what weeding method, what herbicide was used, and when (date) the weeding was performed. In addition, images taken at the time of planning and images taken at the time of weeding may be stored as management history, so that the suitability of the plan can be confirmed in the management plan task 37 described later.

[0044] The growth information acquisition task 35 is a task program that acquires growth information including the growth area, type, and growth status of the weeds W growing in the substation premises G based on an image captured within the substation premises G. In other words, based on the image of the substation premises G received and stored from the monitoring drone 2 and the location information of the photographing location (photographing point), it determines where and how far within the substation premises G the weeds W are growing, what type of weeds W are growing (annual weeds or perennial weeds), and what state the weeds W are growing in (whether they are in a germinated state, a state where leaves are starting to appear, a state where the leaves are lush, etc.).

[0045] At this time, the image of the substation premises G is compared with sample images for each type and growth state of the weeds W stored in the weed information database 341, and the weed information (whether each type is an annual weed or a perennial weed) in the weed information database 341 is referred to. In addition, a map of the substation premises G and which part of the substation premises G is photographed at each photographing point of the substation premises G (position information of the photographing location received from the monitoring drone 2) are stored in advance, so that it is possible to obtain where and to what extent the weeds W are growing in the substation premises G. Here, when the entire substation premises G is photographed in one photographing, the photographing point is not necessary, and where and to what extent the weeds W are growing is obtained based on the map of the substation premises G (or an image without the weeds W) and the photographed image.

[0046] The growth information acquisition task 35 is also started when a management planning task 37, which will be described later, is started, and acquires growth information for one or more substation premises G designated when the management planning task 37 is started. When multiple substation premises G are designated, the growth information is acquired for each of the designated substation premises G based on images of the multiple substation premises G or the like.

[0047] In this type of growth information acquisition task 35, when an image and a photographing point (as described above, there may be no photographing point) are input, a growth information learning model 345A that is machine-learned based on past performance data is used so that growth information is output. This growth information learning model 345A is created by the learning task 38.

[0048] That is, the learning task 38 creates a growth information learning model 345A by a known machine learning algorithm such as a neural network using past performance data recorded and accumulated in a growth information performance database 346A. This growth information performance database 346A is a database in which performance data including growth information determined by experts in weed (plant) management and care and experts in managing weeds W in the substation premises G, etc., based on images and their photographing points as input information. The past performance data includes data created based on actual images and photographing points, and growth information actually determined by experts and experts, as well as data created by advance training, etc.

[0049] As shown in Fig. 3, this learning task 38 uses machine learning and deep learning using a neural network to create a neural network with, for example, an image and its photographing point as an input layer, growth information as an output layer, and analysis processing from the input layer to the output layer as an intermediate layer based on the performance data recorded in the growth information performance database 346A. Then, the learning task 38 uses the performance data of the growth information learning model 345A as learning data to learn various parameters in the intermediate layer. That is, the learning task 38 learns various parameters in the intermediate layer so that growth information is appropriately output based on the image and its photographing point.

[0050] The weather information acquisition task 36 is a task program for acquiring weather information for each substation premises G. That is, the weather information server 4, which provides weather information and weather forecasts for each region, is accessed via the communication unit 33, and short-term, medium-term, and long-term weather information and weather forecasts for each substation premises G are acquired. Here, the weather information includes information that affects the emergence and growth of weeds W, such as predicted temperature, humidity, rainfall, solar radiation, and drought. This weather information acquisition task 36 is started when the management planning task 37 is started, and the acquired weather information for the substation premises G is used and referred to in the management planning task 37.

[0051] The management plan task 37 is a task program that formulates a management plan including the method and timing of weeding to be performed in the future based on growth information, weed information, etc. That is, when a start command is input from the input unit 31 at any time, it is started, and a management plan is formulated for one or more substation premises G specified at this time. Here, in this embodiment, a management plan is formulated for weeding to be performed within a specified period from the current time, for example, within one year. Alternatively, it may be arranged to formulate a management plan for weeding to be performed within a period specified at the time of starting (for example, one month or six months from the current time).

[0052] First, the growth information acquisition task 35 and the weather information acquisition task 36 are started to acquire growth information and weather information for the target substation premises G as described above. Then, a management plan is formulated based on the acquired growth information and weather information, the weed information stored in the weed information database 341, the herbicide information stored in the herbicide information database 342, the facility information stored in the facility information database 343, and the management history stored in the management history database 344.

[0053] Specifically, first, based on the growth information and the weed information, a plan is made as to when and how to weed the target substation premises G. For example, the plan is made as follows. Depending on whether the growing weed W is an annual weed or a perennial weed, determine when it is best to remove it (for example, before annual weeds set seeds) and what method should be used to remove it (for example, by skipping out on perennial weeds). Based on the growth condition of the growing weeds W, it is determined whether early weeding is necessary or whether it is okay to wait a while before weeding, and further, a weeding method suitable for the growth condition is formulated. Based on the location and range / area where weed W is growing, determine how much time it will take to weed each method. Also, based on the location and range where weed W is growing, determine whether or not immediate weeding is required and when it is required. Furthermore, the plan is formulated by taking these factors into consideration comprehensively. For example, a weeding method appropriate for the type of weed W growing and its growing condition is formulated, and a weeding method and weeding timing that can efficiently weed the entire weed W is formulated according to the location and range / area where the weed W is growing.

[0054] In this case, the optimal weeding method and timing are determined based on the current growth state and taking into consideration how much the weeds will grow in the future. In addition, the weeding timing is determined taking into consideration the period from when the weeding is requested to a weeding company until it is actually carried out (waiting period), the time required for each weeding method, and post-processing. For example, the weeding timing is determined so that weeding can be carried out in a growth state suitable for the weeding method, taking into consideration the waiting period. In addition, it is taken into consideration that manual weeding takes 2 to 3 days per substation premises G, and that when weeding with herbicides, about 10 substation premises G can be weeded in one day, but dead leaves must be collected two weeks later, and only about 2 substation premises G can be collected in one day.

[0055] Next, the weeding method and timing are adjusted and formulated based on the facility information of the target substation premises G. For example, the method and timing are adjusted and formulated as follows: If there are fields or houses nearby, it is not advisable to use herbicides, so weeding should be done with a weeder or by hand. Also, since this takes a long time to weed, the timing of weeding should be adjusted. - When the ground is covered with gravel, a weeder cannot be used, so weeding must be done with herbicides or by hand. If hand-cutting is used, it will take a long time to weed, so the weeding period must be adjusted. In the case of weed control using herbicides, for substation premises G in mountainous areas, spraying will be done once a year since the impact of weeds W on the surrounding area is minimal, whereas for substation premises G in urban areas, spraying will be done twice a year, taking into consideration the impact of weeds W on the surrounding area (for environmental conservation purposes). Furthermore, consider cost-effectiveness and select a weeding method that is economically effective in the long term. For example, the cost per weed is higher for manual weeding by the roots than for weeding with herbicides, but manual weeding should be used if weeding with herbicides is required multiple times a year within the target substation premises G.

[0056] In addition, weeding methods and timing are adjusted and formulated based on herbicide information. For example, we adjust and formulate as follows: If the weed W is expected to have just germinated at the determined weeding period, or if the germination of the weed W is to be suppressed after weeding, select a soil treatment agent as the herbicide. Alternatively, if the germination of the weed W is to be suppressed, the weeding method should be weeding with a soil treatment agent, and the weeding period should be early spring (February to March) or autumn (September to October) as much as possible. If the weeds W have already grown or are expected to have grown at the time of the determined weeding period, the weeding method should be foliar treatment chemicals, and the weeding period should be from April to October if possible. Alternatively, the weeding period should be determined so that the weeds W have grown and the weeding method using foliar treatment chemicals is optimal. When multiple species of weeds W are mixed or the growth conditions of weeds W vary, a non-selective herbicide or selective herbicide should be used. However, when plants other than weeds W are also growing, a selective herbicide should be used.

[0057] In addition, weeding methods and timing are adjusted and determined based on management records and meteorological information. For example, we adjust and determine as follows: -The next weeding time is calculated by taking into account the most recent weeding time (date) in the management history and the growth rate of weed W. - If the management history shows that the weeding timing and cycle is later or earlier than normal (than the timing and cycle established above), this is considered to be a unique circumstance or tendency, and the timing and cycle should be adjusted taking this into account. - Formulate a weeding method taking into consideration past weeding methods in the management history. For example, if weed W is overgrown despite frequent herbicide use in the past, manual weeding is a candidate. If the weather information within the substation premises G predicts a lot of rainfall or high temperatures over the medium to long term, the emergence and growth of weeds W will be rapid, and a weeding period will be determined. If the weather information in the substation premises G predicts a cool summer or cold winter in the future, the emergence and growth of weeds W will be slowed down, and the weeding period will be determined. Based on the current growth status and future weather information, and considering future growth, determine the optimal weeding method and timing. Here, when a management plan is formulated for multiple substation premises G, a management plan is formulated for each substation premises G as described above, and a management plan for multiple power facility premises G is formulated collectively so that weeding can be performed efficiently as a whole. For example, considering the time required for weeding for each weeding method and the geographical information (distance, travel time, etc.) between each substation premises G, a plan is made so that weeding can be performed using the same weeding method for multiple substation premises G in one day or over several consecutive days. In addition, if efficient weeding can be performed as a whole by delaying weeding in a specific substation premises G, weeding in the specific substation premises G is delayed. Alternatively, if weeding needs to be performed urgently in a specific substation premises G, the weeding time is formulated by separating this substation premises G from other substation premises G.

[0058] In this management plan task 37, when growth information, weed information, herbicide information, facility information, management history, and weather information are input, a management plan is output using a management plan learning model 345B that is machine-learned based on past performance data. This management plan learning model 345B is created by a learning task 38.

[0059] That is, the learning task 38 uses past performance data recorded and accumulated in the management plan performance database 346B to create a management plan learning model 345B by a known machine learning algorithm such as a neural network. This management plan performance database 346B is a database in which performance data including management plans formulated by experts in weed (plant) management and care and experts in managing weeds W in the substation premises G, etc., are recorded and accumulated based on growth information, weed information, herbicide information, facility information, management history, and weather information as input information. The past performance data includes actual growth information and weed information, data created based on management plans actually formulated by experts and experts, and data created through advance training, etc.

[0060] As shown in Fig. 4, this learning task 38 uses machine learning and deep learning using a neural network to create a neural network based on the performance data recorded in the management plan performance database 346B, with growth information, weed information, herbicide information, facility information, management history, and weather information as an input layer, a management plan as an output layer, and an intermediate layer for analysis processing from the input layer to the output layer. Then, the learning task 38 uses the performance data of the management plan learning model 345B as learning data to learn various parameters in the intermediate layer. In other words, the learning task 38 learns various parameters in the intermediate layer so that an appropriate management plan is output based on the growth information, weed information, herbicide information, facility information, management history, and weather information.

[0061] The result of the management plan thus formulated is output. For example, a Gantt chart is created showing the date and schedule of the weeding work (work using the formulated weeding method) for each substation premises G (including the date of spraying and the date of collecting dead leaves in the case of weeding using herbicides), or a Gantt chart is created showing the target substation premises G and the date and schedule for each weeding method. These Gantt charts are then displayed on the display unit 32 or transmitted to a terminal or computer of a person in charge that has been preset. Growth information for each substation premises G may also be displayed on the display unit 32 or transmitted to a terminal or computer of a person in charge.

[0062] As described above, according to the electric power facility premises weed management system 1, a management plan is automatically formulated as to when and how to weed, based on growth information including the growth area, type, and growth status of the weeds W growing in the substation premises G, and weed information including a weeding method and weeding time suitable for the type of weed W. This makes it possible to weed the weeds W in the substation premises G in an appropriate manner at an appropriate time, that is, to properly manage the weeds W, without relying on human knowledge or experience. Moreover, since the management plan is formulated based on the weeding method and weeding time suitable for the type of weed W growing in the substation premises G, it becomes possible to formulate an appropriate management plan for the substation premises G, which exists in a variety of regions and where a variety of weeds W grow.

[0063] In addition, since the management plan is formulated based on surrounding information such as whether there are fields or not around the substation premises G, it becomes possible to formulate an appropriate management plan for the substation premises G that are located in various regions and may have various restrictions depending on the surrounding environment. For example, for the substation premises G that are surrounded by fields, it becomes possible to formulate a management plan that uses a weeding method that does not use herbicides, thereby reducing the impact on the fields.

[0064] Similarly, since a management plan is formulated based on ground layout information such as whether the ground in the substation premises G is ballast or not, it becomes possible to formulate an appropriate management plan for substation premises G that are located in various regions and have various ground layouts. For example, for a substation premises G with a ballast ground, a management plan can be formulated using a weeding method that does not use a weeder, making it possible to perform appropriate weeding.

[0065] In addition, since a management plan is formulated based on the appropriate target and application time for each type of herbicide, a management plan suitable for the herbicide to be used can be formulated, making it possible to carry out appropriate weeding. In other words, it becomes possible to use a herbicide at the appropriate time for the appropriate type and growth condition of weeds W.

[0066] Furthermore, since the management plan is formulated based on the management history of when, where, and how the weeding was carried out, it becomes possible to formulate an appropriate management plan that takes into account past weeding actions. Also, since the management plan is formulated based on meteorological information in the substation premises G, it becomes possible to formulate an appropriate management plan that is suited to the weather in the substation premises G (suitable for the growth rate of the weeds W according to the weather, etc.).

[0067] On the other hand, since growth information is output using the machine-learned growth information learning model 345A, it is possible to obtain more appropriate growth information and formulate a more appropriate management plan. Similarly, since a management plan is output using the machine-learned management plan learning model 345B, it is possible to formulate a more appropriate management plan.

[0068] Furthermore, when formulating management plans for multiple substation premises G, the management plans for the multiple substation premises G are formulated collectively based on the growth information of the weeds W for each of the multiple substation premises G. This makes it possible to formulate appropriate and efficient management and weeding plans for the multiple substation premises G.

[0069] Although the embodiment of the present invention has been described in detail above, the specific configuration is not limited to this embodiment, and even if there are design changes within the scope of the present invention, they are included in the present invention. For example, in the above embodiment, the weed management computer 3 is configured as an integrated computer, but it may be configured as multiple computers or servers, which may be located in different locations. In addition, although an image of the substation premises G taken by the camera 21 of the monitoring drone 2 is used, any image that shows the growth status of the weeds W in the substation premises G may be used, and for example, it may be an image taken by a fixed camera from a high place such as the substation premises G.

[0070] On the other hand, the above-mentioned electric power facility premises weed management system 1 and weed management computer 3 may be configured by installing the following electric power facility premises weed management program in a general-purpose computer.

[0071] That is, the computer is provided with a weed information database (weed information storage means) 341 which stores at least one of an appropriate weeding method and an appropriate weeding timing for each type of weed W as weed information, a herbicide information database (herbicide information storage means) 342 which stores, for each type of herbicide, an appropriate target for use and an appropriate time for use, as herbicide information, a facility information database (facility information storage means) 343 which stores, as facility information, surrounding information including whether there are any fields around the power facility premises and ground laying information including whether the ground within the power facility premises is covered with ballast, and a database of weeds that have been previously carried out within the power facility premises. a management history database (management history storage means) 344 for storing management history including the area, method, and timing; a growth information acquisition task (growth information acquisition means) 35 for acquiring growth information including the growth area, type, and growth status of weeds W growing within the power facility premises based on images taken within the power facility premises; a weather information acquisition task (weather information acquisition means) 36 for acquiring weather information within the power facility premises; and a management plan task (management plan means) 37 for formulating a management plan including the method and timing of weeding to be performed in the future based on the weed information, herbicide information, facility information, management history, growth information, and weather information, The growth information acquisition task 35 uses a growth information learning model 345A that has been machine-learned based on past performance data so that growth information is output when an image is input, and the management plan task 37 uses a management plan learning model 345B that has been machine-learned based on past performance data so that a management plan is output when weed information, herbicide information, facility information, management history, growth information, and weather information are input. Furthermore, when multiple power facility premises are to be formulated, the growth information acquisition task 35 acquires growth information for each power facility premises based on images of the multiple power facility premises, and the management plan task 37 formulates management plans for the multiple power facility premises based on the growth information, facility information, management history, weather information, weed information, and herbicide information for each power facility premises. [Explanation of symbols]

[0072] 1. Weed management system for power facility premises 2. Surveillance drones 3. Weed management computer 341 Weed information database (weed information storage means) 342 Herbicide information database (herbicide information storage means) 343 Facility information database (facility information storage means) 344 Management history database (management history storage means) 345A Learning model for growth information 345B Learning Model for Management Planning 346A Growth information performance database 346B Performance database for management planning 35 Growth information acquisition task (growth information acquisition means) 36 Weather information acquisition task (Means of acquiring weather information) 37 Management Plan Tasks (Management Plan Tools) 38 Learning Tasks G Substation Premises (Electric Power Facility Premises) W Weed

Claims

1. a growth information acquiring means for acquiring growth information including a growth area, a type, and a growth state of weeds growing within the electric power facility based on an image captured within the electric power facility; a weed information storage means for storing at least one of an appropriate weeding method and a suitable weeding time for each type of weed as weed information; A management planning means for formulating a management plan including a method and timing of weeding to be carried out in the future based on the growth information and the weed information; A weed management system for an electric power facility comprising:

2. a facility information storage means for storing surrounding information including whether or not there is a field around the premises of the electric power facility as facility information; The management planning means formulates the management plan based on the growth information, the weed information, and the facility information. The weed management system for an electric power facility according to claim 1 .

3. a facility information storage means for storing ground surface information including whether the ground surface within the premises of the electric power facility is ballast as facility information; The management planning means formulates the management plan based on the growth information, the weed information, and the facility information. The weed management system for an electric power facility according to claim 1 .

4. a herbicide information storage means for storing herbicide information on a suitable application target and application time for each type of herbicide; The management planning means formulates the management plan based on the growth information, the weed information, and the herbicide information. The weed management system for an electric power facility according to claim 1 .

5. a management history storage means for storing a management history including an area, a method and a time of weeding carried out in the past within the electric power facility premises; the management planning means formulates the management plan based on the growth information, the weed information, and the management history; The weed management system for an electric power facility according to claim 1 .

6. A weather information acquisition means for acquiring weather information within the electric power facility premises, The management planning means formulates the management plan based on the growth information, the weed information, and the weather information. The weed management system for an electric power facility according to claim 1 .

7. The growth information acquisition means uses a growth information learning model that is machine-learned based on past performance data so that the growth information is output when the image is input. The weed management system for an electric power facility according to claim 1 .

8. the management planning means uses a learning model for management planning that is machine-learned based on past performance data so as to output the management plan when the growth information and the weed information are input; The weed management system for an electric power facility according to claim 1 .

9. The growth information acquisition means acquires the growth information for each of the electric power facility premises based on a plurality of images of the electric power facility premises, the management planning means formulates the management plans for the plurality of power facility premises based on the growth information and the weed information for each of the power facility premises; The weed management system for an electric power facility according to claim 1 .

10. Computer, a growth information acquiring means for acquiring growth information including a growth area, a type, and a growth state of weeds growing within the electric power facility based on an image captured within the electric power facility; a weed information storage means for storing at least one of an appropriate weeding method and a suitable weeding time for each type of weed as weed information; a management planning means for formulating a management plan including a method and timing of weeding to be carried out in the future based on the growth information and the weed information; A weed management program for an electric power facility, characterized in that the program functions as follows.

11. The growth information acquisition means uses a growth information learning model that is machine-learned based on past performance data so that the growth information is output when the image is input.

11. The weed management program for an electric power facility according to claim 10.

12. the management planning means uses a learning model for management planning that is machine-learned based on past performance data so as to output the management plan when the growth information and the weed information are input; 11. The weed management program for an electric power facility according to claim 10.

Images