Facility structure estimation device and facility structure estimation program

Abstract

To suppress a work amount required for creating a three-dimensional model of a structure having a main body and a wide part.SOLUTION: A central processing unit in a computer defines a structure with a wide part protruding in a width direction from a main body extending upward, as a target, and creates a three-dimensional model having framed the structure, on the basis of three-dimensional coordinates of each point in a three-dimensional point group representing the structure. The central processing unit executes, as processing for creating the three-dimensional model, in the following order: extraction processing; data conversion processing; feature point estimation processing, data restoration processing, and framing processing. A three-dimensional model created in this way is a result of performing framing with a main body frame corresponding to the main body and a protrusion frame corresponding to the wide part. With regard to creation of a three-dimensional model having framed the structure, a work amount can be carried out while suppressing it to be small through each of the processing based on data thereof, even when the data of a three-dimensional point group representing the structure is enormous.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to an equipment structure estimation device and an equipment structure estimation program. [Background technology]

[0002] 2. Description of the Related Art Some structures have a main body extending upward and a wide portion protruding in a width direction from the main body. An example of such a structure is a power transmission tower. The above-mentioned structure may be inspected using an aircraft such as a drone as follows. That is, the aircraft is moved along a predetermined flight path set around the structure, and the structure is photographed with a camera mounted on the aircraft. Then, based on the image of the structure photographed by the camera, the structure is inspected for any abnormalities. In addition, the inspection of the above-mentioned structure may include checking whether an appropriate distance is secured between the structure and objects such as trees around the structure.

[0003] When determining the flight path of the aircraft or checking the distance between the structure and surrounding objects, it is possible to use a three-dimensional model of the structure. In order to create such a three-dimensional model, data representing the structure as a three-dimensional point cloud is used. Patent Document 1 describes the acquisition of a three-dimensional point cloud representing a structure. If a three-dimensional model of the structure is created based on the three-dimensional point cloud data acquired in this way, it becomes possible to use the three-dimensional model to determine the flight path of the aircraft or check the distance between the structure and surrounding objects. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2023-4929 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, the three-dimensional point cloud data representing the above-mentioned structure having a main body and a wide portion is enormous, so if one were to create a three-dimensional model based on the data by checking the data one by one, the amount of work required to create the three-dimensional model would be large. [Means for solving the problem]

[0006] The means for solving the above problems and their effects will be described below. The facility structure estimation device for solving the above problem includes a control unit that targets a structure having a wide portion protruding in the width direction from a main body extending upward, and creates a three-dimensional model in which the structure is framed based on the three-dimensional coordinates of each point in a three-dimensional point cloud representing the structure. The control unit sequentially executes an extraction process, a data conversion process, a feature point estimation process, a data restoration process, and a framing process. The extraction process extracts each point representing an extraction point, which is a portion corresponding to the wide portion and the wide portion in the main body, from the three-dimensional point cloud representing the structure. The data conversion process converts each point representing the extraction point into each point in a two-dimensional coordinate system excluding the height. The feature point estimation process estimates a plurality of feature points that are bending points in the outer shape of the extraction point represented by each point using each point in the two-dimensional coordinate system. The data restoration process determines the height of the wide portion at the extraction point based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the structure, and converts the two-dimensional coordinates of the plurality of feature points into three-dimensional coordinates by adding the determined height as height information of the plurality of feature points. The framing process creates a three-dimensional model of the structure based on the three-dimensional coordinates of the feature points and the three-dimensional coordinates of a point corresponding to the upper end of the structure among the three-dimensional point group representing the structure, as follows: That is, as the three-dimensional model of the structure, a three-dimensional model is created having a main body frame extending from the point corresponding to the upper end along the outer shape of the main body, and a protruding frame extending in the width direction of the main body from the main body frame to the protruding end of the wide portion relative to the main body.

[0007] According to the above configuration, even if the data of the three-dimensional point cloud representing the structure having the main body and the wide portion is huge, the three-dimensional model of the structure is created based on the data while keeping the amount of work to a minimum. That is, the three-dimensional model created is a framed model having a main body frame and a protruding frame through extraction processing, data conversion processing, feature point estimation processing, data restoration processing, and framing processing. The creation of the framed three-dimensional model of the structure can be performed based on the data through each of the above processes while keeping the amount of work to a minimum, even if the data of the three-dimensional point cloud representing the structure is huge. Therefore, when creating a three-dimensional model of a structure having a main body and a wide portion, the amount of work required for the creation can be kept to a minimum.

[0008] The structure for which the facility structure estimation device creates a three-dimensional model is a power transmission tower. The power transmission tower has the main body in a rectangular prism shape, and has multiple cross arms protruding in the width direction of the main body as wide parts. The extraction process extracts points representing the extraction points corresponding to the multiple cross arms from the three-dimensional point cloud representing the power transmission tower. The feature point estimation process estimates, as the feature points, feature points corresponding to the four corners of the main body at the extraction points and feature points corresponding to the protruding ends of the cross arms relative to the main body at the extraction points. The data restoration process determines the height of the cross arms for each extraction point based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the power transmission tower. Furthermore, the data restoration process assigns the height thus determined as height information of the multiple feature points at the extraction points corresponding to the same height, thereby converting the two-dimensional coordinates of the multiple feature points into three-dimensional coordinates.

[0009] According to the above configuration, for a transmission tower having a wide section protruding in the width direction from a rectangular prism-shaped main body and made of multiple cross arms, a three-dimensional model of the transmission tower is created as a frame through the above processes. Therefore, even if the three-dimensional point cloud data representing a transmission tower having multiple cross arms becomes huge, the amount of work required to create a three-dimensional model of the transmission tower based on that data can be kept low.

[0010] The facility structure estimation program for solving the above problem targets a structure having a wide portion protruding in the width direction from a main body extending upward, and causes a computer to execute an extraction process, a data conversion process, a feature point estimation process, a data restoration process, and a framing process in order as a process for creating a three-dimensional model in which the structure is framed based on the three-dimensional coordinates of each point in a three-dimensional point cloud representing the structure. The extraction process extracts each point representing an extraction point, which is a portion corresponding to the wide portion and the wide portion in the main body, from the three-dimensional point cloud representing the structure. The data conversion process converts each point representing the extraction point into each point in a two-dimensional coordinate system excluding the height. The feature point estimation process estimates a plurality of feature points that are bending points in the outer shape of the extraction point represented by each point using each point in the two-dimensional coordinate system. The data restoration process determines the height of the wide portion at the extraction point based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the structure, and converts the two-dimensional coordinates of the plurality of feature points into three-dimensional coordinates by adding the determined height as height information of the plurality of feature points. The framing process creates a three-dimensional model of the structure based on the three-dimensional coordinates of the feature points and the three-dimensional coordinates of a point corresponding to the upper end of the structure among the three-dimensional point group representing the structure, as follows: That is, as the three-dimensional model of the structure, a three-dimensional model is created having a main body frame extending from the point corresponding to the upper end along the outer shape of the main body, and a protruding frame extending in the width direction of the main body from the main body frame to the protruding end of the wide portion relative to the main body.

[0011] According to the above configuration, even if the data of the three-dimensional point cloud representing the structure having the main body and the wide portion is huge, the three-dimensional model of the structure is created based on the data while keeping the amount of work to a minimum. That is, the three-dimensional model created is a framed model having a main body frame and a protruding frame through extraction processing, data conversion processing, feature point estimation processing, data restoration processing, and framing processing. The creation of the framed three-dimensional model of the structure can be performed based on the data through each of the above processes while keeping the amount of work to a minimum, even if the data of the three-dimensional point cloud representing the structure is huge. Therefore, when creating a three-dimensional model of a structure having a main body and a wide portion, the amount of work required for the creation can be kept to a minimum.

[0012] The structure for which the facility structure estimation program creates a three-dimensional model is a power transmission tower. The power transmission tower has the main body in a rectangular prism shape, and has multiple cross arms protruding in the width direction of the main body as wide parts. The extraction process extracts points representing the multiple cross arms for each extraction point corresponding to the multiple cross arms from the three-dimensional point cloud representing the power transmission tower. The feature point estimation process estimates, as the feature points, feature points corresponding to the four corners of the main body at the extraction points and feature points corresponding to the protruding ends of the cross arms relative to the main body at the extraction points. The data restoration process determines the height of the cross arms for each extraction point based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the power transmission tower. Furthermore, the data restoration process assigns the height thus determined as height information of the multiple feature points at the extraction points corresponding to the same height, thereby converting the two-dimensional coordinates of the multiple feature points into three-dimensional coordinates.

[0013] According to the above configuration, for a transmission tower having a wide section protruding in the width direction from a rectangular prism-shaped main body and made of multiple cross arms, a three-dimensional model of the transmission tower is created as a frame through the above processes. Therefore, even if the three-dimensional point cloud data representing a transmission tower having multiple cross arms becomes huge, the amount of work required to create a three-dimensional model of the transmission tower based on that data can be kept low. [Brief description of the drawings]

[0014] [Figure 1] 1 is a schematic diagram showing the configuration of a computer. [Diagram 2] FIG. 2 is a side view showing a three-dimensional model of a transmission tower. [Diagram 3] FIG. 3 is a plan view showing a three-dimensional model of the transmission tower in FIG. 2. [Figure 4] 2 is a flowchart showing an execution procedure of a facility structure estimation program executed by the computer of FIG. 1. [Diagram 5] FIG. 1 is a side view of a power transmission tower represented by a three-dimensional point cloud. [Figure 6] FIG. 6 is an enlarged view showing an extraction point on the transmission tower in FIG. 5. [Figure 7] 7 is a plan view showing a point cloud when a three-dimensional point cloud at the extraction location in FIG. 6 is two-dimensionalized. FIG. [Figure 8] FIG. 8 is a schematic diagram showing a part of the point cloud of FIG. 7. [Figure 9] FIG. 8 is an explanatory diagram of evaluation of point N in the point cloud of FIG. 7. [Figure 10] FIG. 8 is an explanatory diagram for evaluation of point N1 in the point cloud of FIG. 7. [Figure 11] FIG. 8 is a plan view showing feature points E, F, M, and N tentatively estimated based on the point group of FIG. 7. [Figure 12] 12 is an explanatory diagram of provisional estimation of feature points A, B, C, and D based on feature points M and N in FIG. 11. [Figure 13] FIG. 13 is an explanatory diagram for correcting the feature points A, B, C, and D in FIG. [Figure 14] FIG. 12 is a plan view showing feature points A, B, C, D, E, and F estimated based on the point group of FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Hereinafter, an embodiment of an equipment structure estimation device and an equipment structure estimation program will be described with reference to FIGS. The computer 11 shown in Fig. 1 functions as a facility structure estimation device that creates a three-dimensional model of a structure. The above-mentioned structure may have a main body extending upward and a wide portion protruding in the width direction from the main body. For example, a transmission tower is a structure that has the above-mentioned main body in the shape of a rectangular prism, and has multiple cross arms protruding in the width direction of the main body as the wide portion.

[0016] The computer 11 includes a central processing unit 12, a storage unit 13, and a communication unit 14. The central processing unit 12 controls various devices in the computer 11, such as the storage unit 13 and the communication unit 14, and performs calculations and the like based on various data. The storage unit 13 is for storing various data. The communication unit 14 is for exchanging data with the outside of the computer 11.

[0017] The central processing unit 12 serves as a control unit that creates a three-dimensional model of the transmission tower framed based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the structure, i.e., the transmission tower. The three-dimensional model of the transmission tower created in this manner is shown in Figures 2 and 3. Figure 2 shows the three-dimensional model of the transmission tower as seen from the side, and Figure 3 shows the three-dimensional model as seen from above.

[0018] 1 executes a facility structure estimation program for creating the above-mentioned three-dimensional model. This facility structure estimation program causes the computer 11 to execute extraction processing, data conversion processing, feature point estimation processing, data restoration processing, and framing processing in order, thereby creating a three-dimensional model of the above-mentioned transmission tower based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the above-mentioned transmission tower.

[0019] The flowchart in FIG. 4 shows the procedure for executing the facility structure estimation program. The central processing unit 12, as the process of step 101 (S101) in the above flowchart, imports three-dimensional point cloud data representing the above transmission tower, more specifically, the three-dimensional coordinates of each point in the three-dimensional point cloud. The above three-dimensional point cloud data can be acquired by laser surveying the transmission tower using an aircraft such as a drone. Note that the above three-dimensional point cloud data may be acquired by other methods.

[0020] Such three-dimensional point cloud data may be stored in advance in the memory unit 13 of the computer 11. In this case, the central processing unit 12 retrieves the three-dimensional point cloud data from the memory unit 13. The central processing unit 12 may also retrieve the three-dimensional point cloud data directly from the aircraft via the communication unit 14. The three-dimensional point cloud data may be stored in a server or the like. In this case, the central processing unit 12 may retrieve the three-dimensional point cloud data stored in the server via the network through a network connection made by the communication unit 14.

[0021] The process of S102, S103, S104, S105, and S106 in the above flowchart correspond to the extraction process, data conversion process, feature point estimation process, data restoration process, and framing process, respectively. Each of these processes will be described in detail below.

[0022] <Extraction process> 5 shows a three-dimensional point cloud representing the transmission tower. As an extraction process of S102, the central processing unit 12 extracts each point representing the extraction points A1, A2, and A3, which are points corresponding to the cross arms of the transmission tower and the cross arms of the main body, from the three-dimensional point cloud. In detail, the central processing unit 12 extracts each point representing the extraction points A1, A2, and A3 for each of the extraction points A1, A2, and A3 corresponding to the multiple stages of cross arms of the transmission tower.

[0023] The central processing unit 12 determines the extraction points A1, A2, and A3 as follows. That is, the central processing unit 12 divides the three-dimensional point cloud representing the transmission tower at a predetermined interval in the vertical direction. The central processing unit 12 judges whether the three-dimensional point cloud of each section includes each point representing an arm based on the horizontal spread of the three-dimensional point cloud of each section. Based on this judgment, the central processing unit 12 determines the extraction points A1, A2, and A3 including each point representing a plurality of arms. Then, the central processing unit 12 extracts each point of the extraction point A1, each point of the extraction point A2, and each point of the extraction point A3 as each point including a point representing an arm.

[0024] <Data conversion process> As the data conversion process of S103 in Fig. 4, the central processing unit 12 converts each point representing the extraction points A1, A2, A3 into each point of a two-dimensional coordinate system excluding the height for each of the extraction points A1, A2, A3. Fig. 6 shows each point representing the extraction point A1, and Fig. 7 shows each of the points after conversion into the two-dimensional coordinate system. By converting each point representing the extraction point A1 into each point of the two-dimensional coordinate system in this way, the density of each point on the two-dimensional coordinate system can be increased.

[0025] <Feature point estimation processing> As the feature point estimation process of S104 in Fig. 4, the central processing unit 12 estimates a plurality of feature points that are bending points in the outline of the extracted points A1, A2, and A3 represented by each point of the two-dimensional coordinate system. In detail, the central processing unit 12 estimates, as the feature points, feature points A to D that correspond to the four corners of the main body of the transmission tower at the extracted points A1, A2, and A3, respectively, and feature points E and F that correspond to the protruding ends of the cross arms relative to the main body at the extracted points A1, A2, and A3. These feature points A to F are shown in Fig. 3.

[0026] When estimating feature points A to F, first, for each of the extraction points A1, A2, A3, feature points E, F, M, and N are provisionally estimated when the outer shapes of the extraction points A1, A2, and A3 on the above-mentioned two-dimensional coordinate system are approximated as diamonds. These feature points E, F, M, and N are also shown in Figure 3. The provisional estimation of feature points E, F, M, and N is performed through the following (Step 1) to (Step 3).

[0027] (Step 1) For each of the extracted points A1, A2, and A3, four points are randomly selected from the points in the two-dimensional coordinate system corresponding to the extracted points A1, A2, and A3. Among the points shown in Figure 7, points M, N, E, and F indicate the four points selected in this way.

[0028] (Step 2) Select one of the points M, N, E, and F as a replacement candidate, and select one of the points in the two-dimensional coordinate system other than the points M, N, E, and F as a replacement candidate. For example, as shown in Fig. 8, select point N as the replacement candidate, and select point N1 as the replacement candidate.

[0029] (Step 3) Evaluate which of point N, which is a replacement candidate, and point N1, which is an alternative candidate, is more suitable as one of the above feature points E, F, M, and N. In detail, after calculating an evaluation value for point N and an evaluation value for point N1, the point N or point N1 with the larger evaluation value is set as the latest point N.

[0030] The evaluation value for point N is calculated as follows. One of points M, N, E, and F other than point N, for example point M, is selected. Then, as shown in Fig. 9, points M and N are shifted by a predetermined amount in a direction perpendicular to line segment MN to determine four points MA, MB, NA, and NB. The number of points present inside the range surrounded by points MA, MB, NA, and NB is determined as the evaluation value for point N.

[0031] The evaluation value for point N1 is calculated in the same manner as the evaluation value for point N. That is, as shown in Fig. 10, four points MA, NB, N1A, and N1B are determined by shifting points M and N1 by a predetermined amount in a direction perpendicular to line segment MN1. The number of points that exist inside the range surrounded by points MA, MB, N1A, and N1B is the evaluation value for point N1.

[0032] By repeating the above-mentioned (Step 2) and (Step 3) a predetermined number of times, for example, 1000 times, points M, N, E, and F are determined as shown in Fig. 11. At this time, points M, N, E, and F are provisionally estimated as feature points M, N, E, and F.

[0033] After feature points M, N, E, and F are provisionally estimated, feature points A, B, C, and D are provisionally estimated based on feature points M and N. In detail, as shown in Fig. 12, points A and D are placed at positions shifted from feature point M by a distance of 1 / 2 the line segment MN in a direction perpendicular to the line segment MN, and points B and C are placed at positions shifted from feature point N by a distance of 1 / 2 the line segment MN in a direction perpendicular to the line segment MN. Then, points A, B, C, and D are corrected as follows.

[0034] That is, one of points A, B, C, and D is selected as the correction target. For example, when point D is selected as the correction target, the average length AL of line segments DA, DB, and DC is calculated. This average length AL is indicated by an arrow in FIG. 13. Then, based on the average length AL, positions P1, P2, and P3 are determined as follows. Position P1 is determined at a position spaced apart by the average length AL in the direction from point D toward point A. Position P2 is determined at a position spaced apart by the average length AL in the direction from point D toward point B. Position P3 is determined at a position spaced apart by the average length AL in the direction from point D toward point C. Point D is then corrected so that the thus determined positions P1, P2, and P3 approach points A, B, and C, respectively.

[0035] By repeating such correction of points A, B, C, and D a predetermined number of times, for example 1000 times, points A, B, C, and D are determined as shown in Fig. 14. At this time, points A, B, C, and D are provisionally estimated as feature points A, B, C, and D.

[0036] After the feature points A to F have been provisionally estimated, the following (Step 5) and (Step 6) are repeated a predetermined number of times, for example 3000 times, to estimate the feature points A to F in earnest. (Step 5) One of the feature points A to F is selected as a replacement candidate, and one of the points in the two-dimensional coordinate system other than the feature points A to F is selected as a replacement candidate. (Step 6) Which of the replacement candidate and the replacement candidate is more suitable as one of the feature points A to F is evaluated using the same method as in (Step 3) described above, and the one with the higher evaluation value based on the evaluation is adopted as the latest feature point.

[0037] <Data recovery process> The central processing unit 12, as the data restoration process of S105 in FIG. 4, assigns height information to the estimated feature points A to F, thereby converting the two-dimensional coordinates of the feature points A to F into three-dimensional coordinates. More specifically, the height of the cross arm is obtained for each of the extraction points A1 to A3 based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the transmission tower. That is, for each of the extraction points A1 to A3, a plurality of horizontal planes H1, H2, H3, . . . with different height positions are set as shown in FIG. 6. FIG. 6 shows the horizontal plane H1, H2, H3, . . . set at the extraction point A1. Among the plurality of horizontal planes H1, H2, H3, . . ., the height of the horizontal plane H1, H2, H3, . . . with the largest number of points is set as the height of the cross arm. The height of the cross arm is assigned as height information to the feature points A to F in the two-dimensional coordinate system, thereby converting the two-dimensional coordinates of the feature points A to F into three-dimensional coordinates.

[0038] <Framing process> As the framing process of S106 in Fig. 4, the central processing unit 12 creates a three-dimensional model of the transmission tower shown in Fig. 2 and Fig. 3 based on the three-dimensional coordinates of the characteristic points A to F and the three-dimensional coordinates of a point corresponding to the upper end of the transmission tower among the three-dimensional point cloud representing the transmission tower. In detail, the central processing unit 12 creates, as the above-mentioned three-dimensional model, a three-dimensional model having a main body frame 15 extending along the outer shape of the main body of the transmission tower from a point corresponding to the above-mentioned upper end of the transmission tower, and multiple stages of protruding frames 16 to 18 extending in the width direction of the main body from the main body frame 15 to the protruding ends of the multiple stages of cross arms relative to the main body.

[0039] According to the present embodiment described above in detail, the following advantageous effects can be obtained. Even if the amount of data of the three-dimensional point cloud representing the transmission tower, which is a structure having a rectangular prism-shaped main body and multiple cross arms as the wide portion, is huge, the three-dimensional model of the transmission tower is created based on such data while keeping the amount of work small. That is, the three-dimensional model created is a frame having a main body frame 15 and protruding frames 16 to 18 through extraction processing, data conversion processing, feature point estimation processing, data restoration processing, and framing processing. The creation of the framed three-dimensional model of the transmission tower can be performed based on the data through each of the above processes while keeping the amount of work small, even if the amount of data of the three-dimensional point cloud representing the transmission tower is huge. Therefore, when creating a three-dimensional model of the transmission tower, which is a structure having a main body and multiple cross arms as the wide portion, the amount of work required for the creation can be kept small.

[0040] The above embodiment can be modified, for example, as follows: The above embodiment and the following modified examples can be combined with each other to the extent that no technical contradiction occurs. In the data restoration process, the height information assigned to the estimated feature points A to F may be the height of each cross arm obtained from a design drawing of the tower.

[0041] The number of cross arms on a transmission tower may be any value other than three. The structure having a main body and a wide portion may be a structure other than a transmission tower. An example of such a structure is a support pole for a ropeway. [Explanation of symbols]

[0042] 11. Computer 12...Central processing unit 13...Storage section 14…Communications Department 15…Main frame 16~18…Protruding frame

Claims

1. A control unit is provided for creating a three-dimensional model of a structure having a wide portion protruding in a width direction from a main body extending upward, based on three-dimensional coordinates of each point in a three-dimensional point cloud representing the structure, the three-dimensional model being framed by the structure; the control unit sequentially executes an extraction process, a data conversion process, a feature point estimation process, a data restoration process, and a framing process; The extraction process includes extracting points representing the wide portion and an extraction portion that is a portion of the main body corresponding to the wide portion from the three-dimensional point cloud representing the structure; The data conversion process converts each of the points representing the extracted portion into each of the points in a two-dimensional coordinate system excluding height, the feature point estimation process estimates a plurality of feature points that are bending points in an outline of the extracted portion represented by each of the points in the two-dimensional coordinate system; the data restoration process determines a height of the wide portion at the extraction location based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the structure, and assigns the determined height as height information of the plurality of feature points, thereby converting the two-dimensional coordinates of the plurality of feature points into three-dimensional coordinates; The framing process is an equipment structure estimation device that creates a three-dimensional model of the structure based on the three-dimensional coordinates of the feature points and the three-dimensional coordinates of a point corresponding to the upper end of the structure among the three-dimensional point cloud representing the structure, and has a main body frame extending from the point corresponding to the upper end along the outer shape of the main body, and a protrusion frame extending in the width direction of the main body from the main body frame to the protruding end of the wide portion relative to the main body.

2. The structure is a transmission tower, The transmission tower includes a rectangular prism-shaped main body and a plurality of cross arms protruding in a width direction of the main body as the wide portion, The extraction process includes extracting points representing the extraction locations corresponding to the cross arms of a plurality of stages from the three-dimensional point cloud representing the transmission tower, the feature point estimation process estimates, as the feature points, feature points corresponding to the four corners of the body at the extraction location and feature points corresponding to protruding ends of the brackets relative to the body at the extraction location; The facility structure estimation device according to claim 1, wherein the data restoration process calculates the height of the cross arm for each of the extracted locations based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the transmission tower, and assigns the calculated height as height information of the multiple feature points at the extracted locations corresponding to the same height, thereby converting the two-dimensional coordinates of the multiple feature points into three-dimensional coordinates.

3. A process for creating a three-dimensional model of a structure that has a wide portion protruding in a width direction from a main body extending upward, based on three-dimensional coordinates of each point in a three-dimensional point cloud representing the structure, in which the structure is framed, is performed by a computer in the following order: extraction process, data conversion process, feature point estimation process, data restoration process, and framing process; The extraction process includes extracting points representing the wide portion and an extraction portion that is a portion of the main body corresponding to the wide portion from the three-dimensional point cloud representing the structure; The data conversion process converts each of the points representing the extracted portion into each of the points in a two-dimensional coordinate system excluding height, the feature point estimation process estimates a plurality of feature points that are bending points in an outline of the extracted portion represented by each of the points in the two-dimensional coordinate system; the data restoration process determines a height of the wide portion at the extraction location based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the structure, and assigns the determined height as height information of the plurality of feature points, thereby converting the two-dimensional coordinates of the plurality of feature points into three-dimensional coordinates; The framing process is an equipment structure estimation program that creates a three-dimensional model of the structure based on the three-dimensional coordinates of the feature points and the three-dimensional coordinates of a point corresponding to the upper end of the structure among the three-dimensional point cloud representing the structure, and has a main body frame extending along the outer shape of the main body from the point corresponding to the upper end, and a protruding frame extending in the width direction of the main body from the main body frame to the protruding end of the wide portion relative to the main body.

4. The structure is a transmission tower, The transmission tower includes a rectangular prism-shaped main body and a plurality of cross arms protruding in a width direction of the main body as the wide portion, The extraction process includes extracting points representing the extraction locations corresponding to the cross arms of a plurality of stages from the three-dimensional point cloud representing the transmission tower, the feature point estimation process estimates, as the feature points, feature points corresponding to the four corners of the main body at the extraction location and feature points corresponding to protruding ends of the brackets relative to the main body at the extraction location; The facility structure estimation program according to claim 3, wherein the data restoration process calculates the height of the cross arm for each of the extracted locations based on the three-dimensional coordinates of each point in the three-dimensional point cloud representing the transmission tower, and assigns the calculated height as height information of the multiple feature points at the extracted locations corresponding to the same height, thereby converting the two-dimensional coordinates of the multiple feature points into three-dimensional coordinates.

Images