Measurement method and program

Abstract

An object of the present disclosure is to provide a measurement method capable of accurately specifying an edge position of an individual suspended load even when a plurality of suspended loads are arranged without a gap and even under an environmental condition of nighttime or rainy weather.SOLUTION: A measurement method of the present disclosure is a measurement method for specifying an edge position of each of a plurality of suspended loads having a known width and the same shape using a three dimensional sensor in a state where the suspended loads are arranged on a plane in the same direction, and includes acquiring point cloud data obtained by measuring a distance to an object including the suspended loads by the three dimensional sensor from a side opposite to the plane with respect to the suspended loads, specifying a point cloud of a top surface of each of the suspended loads from the point cloud data, specifying a center position of the top surface of each of the suspended loads based on the number of the suspended loads, and specifying the edge position of the suspended load from the center position based on the width of the suspended load.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a measurement method and a program. [Background technology]

[0002] In order to reduce fixed costs such as labor costs, the automation of cranes in billet yards is being considered. To achieve crane automation, the crane must be able to recognize the location of the billet specified by the operator. In addition, there are up to five billets lined up as loads, and sometimes only a specified number of billets must be lifted. In order to lift the desired number of billets by their edges (corners), the position of the billet edges must be identified accurately.

[0003] A 3D sensor attached to the crane can be used to identify this edge position. Several methods have been proposed for detecting an object to be lifted using a 3D sensor attached to a crane.

[0004] Patent Document 1, which is listed as a prior art document, shows a method of measuring the depth of the top surface of an object from above the object using a 3D sensor, and detecting the boundary between the floor and the object from this and an image from an RGB camera, thereby identifying the position coordinates of three or more vertices of the object (corners of the object) and calculating the vertical and horizontal widths of the object.

[0005] Patent Document 2 discloses a method for generating a plane cluster by clustering point cloud data detected by a 3D sensor and presenting easily recognizable guide information to a crane operator. This presentation method involves estimating a plane, dividing the point cloud data into layers of a predetermined thickness and assigning them to multiple groups. A plane is estimated for each group. Specifically, this plane estimation is performed using the following procedure: First, any two points are selected from the point cloud data. If the distance between them is less than a threshold, they are considered to be in the same plane, and a center of gravity is set. The center of gravity and one nearby point are selected and subjected to the same process to update the center of gravity. When the update stops, the point clouds used to calculate the center of gravity are determined to be in the same plane.

[0006] Patent Document 3 shows a device for measuring the position of an object lifted by a container crane. This device uses a three-dimensional non-contact sensor to measure the container below the crane, and estimates points where the height coordinate value increases and decreases in the measurement data as the edges of the container's top surface. If the distance of the estimated edge position roughly matches the container width, it is identified as a container. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Application Publication No. 2023-005398 [Patent Document 2] Patent Publication No. 2021-103530 [Patent Document 3] Japanese Patent Application Laid-Open No. 2006-312521 Summary of the Invention [Problem to be solved by the invention]

[0008] In billet yards, multiple loads (billets) may be lined up with no gaps between them. In this case, the depth difference between the loads may appear small to a 3D sensor. The methods described in Patent Documents 1 and 3 identify the vertices of an object based on the depth difference measured by the 3D sensor, making it difficult to identify the vertex of each load when the depth difference is small. Furthermore, the method described in Patent Document 2 may treat the points of adjacent loads as planes of the same object when updating the center of gravity.

[0009] Work may be performed even under environmental conditions such as at night or in rainy weather. The method described in Patent Document 1 uses an RGB camera, which may reduce visibility from the camera's captured image under the above-mentioned environmental conditions, resulting in a risk of reduced accuracy in identifying edge positions. Furthermore, under conditions such as rain, 3D sensors are generally prone to partial loss of measurement data and increased measurement errors. In particular, data loss at the edge of an object may result in a large error in the measured top surface of the object. The methods described in each patent document do not mention any countermeasures, which may result in a risk of reduced accuracy in identifying edge positions in such cases.

[0010] The present disclosure has been made based on the above-mentioned circumstances, and aims to provide a measurement method and program that can accurately identify the edge position of each hanging load, even when multiple hanging loads are lined up closely together, and even in environmental conditions such as at night or in rainy weather. [Means for solving the problem]

[0011] A measurement method according to one aspect of the present disclosure is a measurement method that uses a three-dimensional sensor to identify the edge position of each of a plurality of hanging loads, each having a known width and the same shape, arranged in the same orientation on a plane. The measurement method includes an acquisition process that acquires point cloud data measuring the distance to an object including the plurality of hanging loads using the three-dimensional sensor from the opposite side of the plane for the plurality of hanging loads; a top surface identification process that identifies a point cloud of the top surfaces of the plurality of hanging loads from the point cloud data; a center identification process that identifies the center position of the top surface of each hanging load based on the number of the hanging loads; and an edge identification process that identifies the edge position of the hanging load from the center position based on the width of the hanging load.

[0012] Another aspect of the present disclosure is a program for causing a computer to identify the edge position of each of a plurality of hanging loads having known widths and identical shapes, arranged in the same orientation on a plane, using a three-dimensional sensor. The program includes an acquisition step for acquiring point cloud data measuring the distance to an object including the plurality of hanging loads from the opposite side of the plane using the three-dimensional sensor for the plurality of hanging loads; a top surface identification step for identifying a point cloud of the top surfaces of the plurality of hanging loads from the point cloud data; a center identification step for identifying the center position of the top surface of each hanging load based on the number of the hanging loads; and an edge identification step for identifying the edge position of the hanging load from the center position based on the width of the hanging load. [Effects of the Invention]

[0013] The measurement method and program disclosed herein can accurately identify the edge position of each load even when multiple loads are closely spaced and in environmental conditions such as at night or in rainy weather. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is a schematic side view showing the vicinity of a lifting mechanism of a crane in which a measurement method according to one embodiment of the present disclosure is used. [Figure 2] FIG. 2 is a flow chart showing a method for measuring the edge position of a load used in the crane of FIG. [Figure 3] FIG. 3 is an image diagram showing an example of point cloud data of a suspended load acquired and extracted in the acquisition step of FIG. [Figure 4] FIG. 4 is an explanatory diagram showing a method for identifying the top surface using a unit normal vector in the top surface identification step. [Figure 5] FIG. 5 is an explanatory diagram showing a method for identifying the top surface based on the difference in distance from the plane between each point and its adjacent point in the top surface identification step. [Figure 6] FIG. 6 is a schematic diagram showing an example of a top surface identified in the top surface identifying step. [Figure 7]FIG. 7 is an explanatory diagram showing one method for identifying the center position in the center identifying step. [Figure 8] FIG. 8 is an explanatory diagram showing the effect of identifying the center position. DETAILED DESCRIPTION OF THE INVENTION

[0015] [Description of the embodiments of the present disclosure] (1) A measurement method according to one aspect of the present disclosure is a measurement method for using a three-dimensional sensor to identify the edge position of each of multiple loads, each having a known width and the same shape, arranged in the same orientation on a plane. The measurement method includes an acquisition process for acquiring point cloud data measuring the distance to an object including the multiple loads using the three-dimensional sensor from the opposite side of the plane for the multiple loads; a top surface identification process for identifying a point cloud of the top surfaces of the multiple loads from the point cloud data; a center identification process for identifying the center position of the top surface of each load based on the number of the loads; and an edge identification process for identifying the edge position of the load from the center position based on the width of the load.

[0016] Generally, in measurements to identify the edge position of each suspended load, multiple suspended loads are often identical in shape and orientation, and their widths are known. This measurement method utilizes these characteristics of the suspended loads to improve the accuracy of edge position identification. In other words, in this measurement method, even if the top surfaces of adjacent suspended loads are identified as the same top surface when identifying the point cloud of the top surfaces of multiple suspended loads in the top surface identification step, the center position of each suspended load's top surface is identified in the center identification step based on the width of the suspended load. Therefore, even if a portion of the measurement data is missing using a 3D sensor, for example, errors are unlikely to occur. Therefore, this measurement method can accurately identify the edge position of each suspended load even when multiple suspended loads are closely spaced, and even under environmental conditions such as at night or in the rain.

[0017] (2) In the measurement method of (1) above, the number of the suspended loads may be determined from the width of the point cloud of the top surface and the width of the suspended loads. In this measurement method, when the point cloud of the top surfaces of multiple suspended loads is determined in the top surface determination step, even if the top surfaces of adjacent suspended loads are determined to be the same top surface, the number of suspended loads can be determined from the width of the suspended loads.

[0018] (3) In the measurement method (1) or (2), the top surface may be identified using the unit normal vector of each point of the point cloud data. By identifying the top surface using the unit normal vector of each point of the point cloud data in this way, the accuracy of identifying the top surface can be improved.

[0019] (4) In the top surface identification step of the measurement method (1) or (2), the top surface may be identified based on the difference in distance from the plane between each point and its adjacent point. Identifying the top surface based on the difference in distance from the plane in this way makes it easy to identify the top surface.

[0020] (5) In the center determining step of any one of the measurement methods (1) to (4) above, if there are multiple suspended loads, it is advisable to determine the center position from the suspended loads at both ends. By determining the center position from the suspended loads at both ends in this way, it is possible to further reduce errors in the center position.

[0021] (6) In the center identifying step of any of the measurement methods (1) to (4), the width of the point cloud on the top surface may be equally divided into sections equal to the number of the suspended loads, and the center of each section may be set as the center position. By dividing the width of the point cloud on the top surface into sections equal to the number of the suspended loads and setting the center positions in this way, the center position can be easily identified.

[0022] (7) Another aspect of the present disclosure provides a program for causing a computer to use a three-dimensional sensor to identify the edge position of each of a plurality of loads having known widths and identical shapes, arranged in the same orientation on a plane. The program includes an acquisition step for acquiring point cloud data measuring the distance to an object including the plurality of loads from the opposite side of the plane using the three-dimensional sensor for the plurality of loads; a top surface identification step for identifying a point cloud of the top surfaces of the plurality of loads from the point cloud data; a center identification step for identifying the center position of the top surface of each load based on the number of the loads; and an edge identification step for identifying the edge position of the load from the center position based on the width of the load.

[0023] In this program, when identifying the point cloud of the top surfaces of multiple loads in the top surface identification step, even if the top surfaces of adjacent loads are identified as the same top surface, the center identification step identifies the center position of the top surface of each load from the width of the load, so errors are unlikely to occur even if, for example, there is a partial loss of measurement data from a 3D sensor.As a result, the program can accurately identify the edge position of each load even if multiple loads are lined up closely together, or even in environmental conditions such as at night or in the rain.

[0024] Here, "load width" refers to the distance between the boundaries of adjacent loads when other loads are lined up on both sides of a certain load without any gaps. For example, if the load is a rectangular parallelepiped, it is the length of the side perpendicular to the adjacent side, and if the loads are cylindrical and lined up vertically, it is the diameter of the circle at the base.

[0025] [Details of the embodiments of the present disclosure] A measurement method and a program according to an embodiment of the present disclosure will be described in detail below.

[0026] A measurement method according to one embodiment of the present disclosure is used in a crane 1 as shown in FIG. 1, for example, and is a measurement method in which, in a state in which a plurality of suspended loads X of the same shape are aligned in the same direction on a plane P, a three-dimensional sensor 2 is used to identify the edge position of each of the suspended loads X.

[0027] <Hanging load> The load X is an object to be lifted by the crane 1, such as a billet or a container. The shape of the load X is preferably rectangular from the viewpoint of ease of lifting, but is not limited to a rectangular parallelepiped shape. The following description will be given taking as an example a case where the load X is a rectangular parallelepiped billet, but the load X is not limited to a billet.

[0028] As shown in Figure 1, the load X (billet) is stored inside a stand Y installed on floor G, for example. The loads X are long and are arranged so that their longitudinal directions are parallel. At this time, there may be a gap between adjacent loads X, but they may also be arranged without any gaps. As shown in Figure 1, they may be stacked in multiple levels (two levels in Figure 1). The load X in the first level is placed directly on floor G, which is a flat surface. The load X in the second level is stacked on top of the multiple loads X in the first level, but because the multiple loads X have the same shape and are arranged in the same direction, the top surface forms a flat surface P. In other words, both the first and second levels are a state in which multiple loads X are arranged in the same direction on a flat surface.

[0029] The plane is preferably a horizontal plane. The following description will be given on the assumption that the plane is horizontal, but this does not exclude the case where the plane is inclined. The measurement method will function even if the plane is inclined.

[0030] <Crane> The crane 1 includes, for example, a three-dimensional sensor 2, a crane body 3 which is a mechanism for lifting a load X, and a cab 4 for supporting and moving the crane body 3.

[0031] The three-dimensional sensor 2 can measure the distance to an object and obtain information about the depth at the measurement point. The type of the three-dimensional sensor 2 is not particularly limited, and a well-known LiDAR (Light Detection and Ranging) or light cutting sensor can be used.

[0032] The 3D sensor 2 is preferably installed in a location that moves following the crane 1, such as the crane body 3 or the cab 4. The 3D sensor 2 is also positioned so that one of its x, y, or z axes is parallel to the longitudinal direction of the suspended load X. In FIG. 1, the sensor is installed so that the z axis is the longitudinal direction.

[0033] Although the crane 1 in Figure 1 has one 3D sensor 2, multiple 3D sensors 2 may be used. By using multiple 3D sensors 2 in this way, it is possible to integrate the measured point cloud data, thereby expanding the imaging range or improving the measurement accuracy of the point cloud data. Furthermore, by installing the sensors at different locations on the crane 1, it is possible to configure the system to calculate the positional relationship between each location on the crane 1 and the suspended load X.

[0034] The form of the crane 1 is not limited to the configuration shown in Fig. 1, but it is configured to be able to acquire its own position coordinates, and therefore a device such as a laser rangefinder may be used.

[0035] <Measurement method> As shown in FIG. 2, the measurement method includes a preparation step S1, an acquisition step S2, a top surface identification step S3, a center identification step S4, and an edge identification step S5.

[0036] (preparation process) In the preparation step S1, load information including the width of the load X is input. That is, in this measurement method, the width of the load X is known.

[0037] The method for inputting the suspended load information is not particularly limited, but for example, it can be input from a keyboard using a personal computer.

[0038] The above-mentioned suspended load information preferably includes the height of the suspended load X in addition to the width, which is essential. Other information may be included, such as the number of tiers in which the suspended loads X are stacked, the number of suspended loads X lined up horizontally on the top tier, the height of the three-dimensional sensor 2 from the ground, the height of the floor G, and the width of the platform Y. This information can reinforce the validity of the measurement results.

[0039] (Acquisition process) In the acquisition step S2, point cloud data D is acquired by measuring the distance to the object including the multiple loads X using the 3D sensor 2 from the opposite side of the plane P for the multiple loads X (see Figure 3). In this measurement method, the plane P is a horizontal plane, so specifically, in the acquisition step S2, the area including the multiple loads X from above to directly below the multiple loads X is scanned using the 3D sensor 2 to acquire the distance to the object (the distance in the y direction in Figure 1).

[0040] The scanning by the three-dimensional sensor 2 is not particularly limited, but for example, it can be performed continuously in the x direction while discretely changing the z coordinate.

[0041] Point cloud data D includes multiple suspended loads X, but may also include information on objects other than suspended load X. Information on these objects does not need to be processed. Also, because there is a risk that they may be mistaken for suspended load X depending on their shape, a portion including the top surface of the uppermost suspended load X is extracted from the scanned point cloud data D.

[0042] For example, if the load information includes the height of the 3D sensor 2 from the ground, the height of the floor G, the width of the platform Y, and the number of stacked levels of the load X, the extraction can be performed by estimating the range of the platform Y and the range including the top surface of the top level of the load X from the point cloud data D obtained based on this information, and extracting the information between them.

[0043] Some points may further be excluded from the extracted point cloud data D. Points to be excluded may include, for example, the point cloud at the end of the suspended load X, which has low measurement accuracy.

[0044] (Top surface identification process) In the top surface identification step S3, a point cloud of the top surfaces C of the multiple suspended loads X is identified from the point cloud data D.

[0045] The top surface C can be identified using the unit normal vector n of each point of the point cloud data D, as shown in FIG.

[0046] In this method, first, the unit normal vector n is estimated from the three-dimensional point cloud data D. A general normal estimation method using plane detection can be used to estimate the unit normal vector n. Of the estimated normal vectors n(nx, ny, nz), for example, the plane where -0.5≦nx≦0.5 can be determined to be the top surface C of the suspended load X. Note that, as shown in Figure 4, multiple planes may fit the criteria. In this case, the plane with the larger y coordinate corresponds to the top surface. However, it is also possible that multiple top surfaces C are identified with the same y coordinate.

[0047] If the distance between point clouds in the longitudinal direction (z direction) is too long, general normal estimation methods may not be usable. In such cases, it is advisable to perform coordinate transformation to compress the longitudinal distance, making the distance between point clouds in the longitudinal direction appear shorter, and then use the normal estimation method.

[0048] In this way, it is possible to improve the accuracy of identifying the top surface C by identifying the top surface C using the unit normal vector n of each point of the point cloud data D. In particular, this method provides high accuracy in identifying the top surface C even when the top surface is uneven or contains noise.

[0049] After the top surface C is identified, the point cloud data D is organized in a direction other than the longitudinal direction for the next edge identification step S4. The method for organizing the point cloud data D in a direction other than the longitudinal direction is not particularly limited, but it is preferable to delete the z coordinate and leave only the x and y coordinates, for example.

[0050] The top surface C can also be identified based on the difference in distance from the plane P between each point and its adjacent point, as shown in FIG.

[0051] In this method, first, the point cloud data D is grouped in a direction other than the longitudinal direction (x and y coordinates only). The method of grouping the point cloud data D is the same as when using the unit normal vector n. Next, for example, each point is searched in ascending order of x coordinate, and the difference s in y coordinate between the point and the adjacent point in the positive direction of the x axis is calculated. i Calculate s from a predetermined threshold. i If is large, y i and yi+1 This is done in order, and the remaining point cloud is taken as the point cloud of the top surface C.

[0052] By identifying the top surface C based on the difference in distance from the plane P in this way, the top surface C can be identified easily.

[0053] Figure 6 shows an example of a top surface C identified in the top surface identification step S3. In the example of Figure 6, three top surfaces C1, C2, and C3 are identified. When adjacent suspended loads X are lined up without any gaps, the top surface C may be identified as a single continuous surface. Top surface C1 is such an example.

[0054] (center identification process) In the center identifying step S4, the center position of the top surface C of each of the suspended loads X is identified based on the number of the suspended loads X.

[0055] First, the point cloud data D is classified into clusters by clustering. This clustering determines the point clouds that belong to each top surface C. Existing clustering methods such as DBSCAN and k-means can be used.

[0056] Next, the number of loads X included in each top surface C is identified. The number of loads X can be provided in advance from outside as known data, but it can also be identified from the width of the point cloud of the top surface C and the width of the loads X. In this measurement method, when identifying the point cloud of the top surfaces C of multiple loads X in the top surface identification step S3, even if the top surfaces C of adjacent loads X are identified as the same top surface, the number can be identified from the width of the loads X.

[0057] When determining the number of loads X from the width of the point cloud of top surface C and the width of loads X, the number of loads X can be calculated by dividing the length of each top surface C in the x-axis direction (maximum x-coordinate value - minimum x-coordinate value) by the width of load X. For example, in the example of FIG. 6, the number of loads X on top surface C1 is 3, and the number of loads X on top surfaces C2 and C3 is 1. Note that the above division may not result in an integer value due to partial loss of measurement data from 3D sensor 2 or measurement errors, but even in such cases, the number will be relatively close to a specific integer value, such as 2.98, and therefore the number of loads X can be easily determined by rounding, for example.

[0058] Next, identify the center position of the top surface C of each suspended load X. Hereinafter, an example will be described using FIG.

[0059] For top surfaces C2 and C3 where the number of suspended loads X is 1, center positions M2 and M3 can be identified as positions that equally divide the length in the x-axis direction (the midpoint between the maximum and minimum x-coordinate values).

[0060] When there are multiple loads X, it is advisable to identify the central position M starting from the loads X at both ends. Taking the top surface C1 with three loads X as an example, and explaining this using Figure 7, first, the positions 1 / 2 the width W of the loads X inside from both ends of the top surface C1 (the position with the maximum x coordinate and the position with the minimum x coordinate) are set as the central positions M1 and M3 of the two loads X at both ends of the top surface C1.

[0061] Next, if the positions inside the width W of the load X from both ends of the top surface C1 are defined as reference positions B1 and B2, then if the number of loads X is three, one load X (=3-2) exists between these two reference positions B1 and B2, and the position that equally divides the two reference positions B1 and B2 becomes the center position M2 of this one load X. Note that if the number of loads X is two, no load X exists between the two reference positions B1 and B2, and therefore it is not necessary to specify the center position between the two reference positions B1 and B2. Conversely, if the number of loads X is four or more, two or more loads X exist at the two reference positions B1 and B2. In this case, the distance between the two reference positions B1 and B2 is equally divided by (the number of loads X - 2), and the center position of each equally divided section is determined to be the center position.

[0062] In this way, by identifying the center position from the loads at both ends, the error in the center position can be further reduced. This method can reduce the error in the center position of at least the loads X at both ends that are lined up side by side. In particular, when there are two loads X, even a slight gap between the loads X can be taken into consideration.

[0063] The width of the point cloud on the top surface C1 may be equally divided into sections equal to the number of suspended loads X, and the centers of the respective sections may be set as center positions M1, M2, M3. By dividing the width of the point cloud on the top surface C1 equally into sections equal to the number of suspended loads X and setting the center positions M1, M2, M3 in this way, it is possible to easily identify the center positions.

[0064] (Edge identification process) In the edge identifying step S5, the edge positions of the load X are identified from the center positions M1, M2, and M3 based on the width W of the load.

[0065] Specifically, the edge positions of each load X can be identified as positions extending from the respective center positions M1, M2, and M3 by 1 / 2 of the width W of the load X on both sides.

[0066] Here, with reference to FIG. 8, the effect of identifying the edge position with reference to the center position M will be described by taking as an example a case where the number of suspended loads X is one.

[0067] The left diagram in Figure 8 assumes that there is a partial loss in the measurement data of the 3D sensor 2 at both ends of load X, causing the width of the top surface C to be shorter by ΔW at both ends. In this case, if the method of aligning the ends of load X with ends E1 and E2 of the top surface C is used, it will be identified as the position of load X1 in the middle of Figure 8, which will result in an error of ΔW from the actual position of load X. In contrast, if it is aligned with center position M, it will be identified as the position of load X2 in the bottom of Figure 8, which will match the actual position of load X.

[0068] The right diagram in Figure 8 assumes that there is a partial loss in the measurement data of the 3D sensor 2 at one end of the load X, and the width of the top surface C is shorter by ΔW at one end E1. In this case, if the end of the load X is aligned with the other end E2 of the top surface C, there will be no error with the actual position of the load X, but if it is aligned with one end E1, there will be an error of ΔW with the actual position of the load X. In contrast, if it is aligned with the center position M, the error is suppressed to ΔW / 2.

[0069] In this way, specifying the edge position based on the center position M can prevent large errors from occurring. Here, the case where there is one suspended load X has been described, but the same effect can be achieved even if there are two or more suspended loads. Furthermore, this applies not only to partial loss of measurement data, but also to measurement errors.

[0070] <Program> The acquisition step S2, the top surface identification step S3, the center identification step S4, and the edge identification step S5 can be executed by a program. That is, a program according to another embodiment is a program for causing a computer to identify the edge position of each of the loads X using a three-dimensional sensor 2 when multiple loads X of known widths and identical shapes are aligned in the same direction on a plane P, and includes the following steps: an acquisition step for acquiring point cloud data D obtained by measuring the distance to an object including the multiple loads X using the three-dimensional sensor 2 from the opposite side of the plane P for the multiple loads X; a top surface identification step for identifying a point cloud of the top surfaces C of the multiple loads X from the point cloud data D; a center identification step for identifying a center position M of the top surface C of each load X based on the width of the point cloud of the top surface C and the number of loads X identified from the width W of the loads; and an edge identification step for identifying the edge position of the load X from the center position M based on the width of the load X.

[0071] The program may also include a preparation step of inputting load information including the width of the load X corresponding to the preparation step S1.

[0072] <Advantages> Generally, in measurements to identify the edge position of each suspended load, multiple suspended loads X have the same shape and are often stored with the same orientation, and the width W of the suspended load is also known. The measurement method and program utilize these characteristics of the suspended load X to improve the accuracy of identifying the edge position. That is, in the measurement method and program, when identifying the point cloud of the top surfaces C of multiple suspended loads X in the top surface identification process S3 (top surface identification step), even if the top surfaces C of adjacent suspended loads X are identified as the same top surface, the center position M of the top surface C of each suspended load X is identified in the center identification process S4 (center identification step) based on the width of the suspended load X. Therefore, even if there is a partial loss of measurement data from the 3D sensor 2, for example, errors are unlikely to occur. Therefore, the measurement method and program can accurately identify the edge position of each suspended load X even when multiple suspended loads X are closely spaced, or even under environmental conditions such as at night or in the rain.

[0073] [Other embodiments] The above-described embodiments do not limit the configuration of the present invention. Therefore, the above-described embodiments may include omissions, substitutions, or additions of components based on the description in this specification and common general technical knowledge, and all of these should be construed as falling within the scope of the present invention.

[0074] In the above embodiment, the case where the preparation step is included has been described, but the preparation step is not an essential step. For example, if the width of the suspended load is pre-coded in the program, the preparation step S1 does not need to be performed each time a measurement is made and can be omitted.

[0075] In the above embodiment, the case where point cloud data including multiple suspended loads is extracted in the acquisition step (hereinafter, this extraction is also referred to as the "extraction step") has been described, but this extraction step is not an essential step. As long as the top surface of the suspended load can be identified, the extraction step can be omitted. However, from the viewpoint of reducing the processing of unnecessary areas and improving the efficiency of the measurement method, it is preferable to include the extraction step. [Industrial Applicability]

[0076] The measurement method and program disclosed herein can accurately identify the edge position of each load even when multiple loads are closely spaced and in environmental conditions such as at night or in rainy weather. [Explanation of symbols]

[0077] 1 crane 2. 3D sensor 3 Crane body 4 Cab X, X1, X2 Hanging load Y-mounted stand G floor P plane C, C1, C2, C3 top D Point cloud data n unit normal vectors M, M1, M2, M3 center position B1, B2 reference position E1, E2 ends W Width of the suspended load

Claims

1. A measurement method for identifying the edge position of each of a plurality of hanging loads having known widths and identical shapes, arranged in the same direction on a plane, using a three-dimensional sensor, comprising: an acquisition step of acquiring point cloud data in which distances to objects including the plurality of hanging loads are measured by the three-dimensional sensor from the opposite side of the plane with respect to the plurality of hanging loads; a top surface identification process for identifying a point cloud of the top surfaces of the plurality of suspended loads from the point cloud data; a center identifying step of identifying the center position of the top surface of each of the suspended loads based on the number of the suspended loads; an edge identifying step of identifying an edge position of the suspended load from the center position based on the width of the suspended load; A measurement method comprising:

2. The measurement method according to claim 1 , wherein the number of the suspended loads is determined from the width of the point cloud on the top surface and the width of the suspended loads.

3. 2. The measurement method according to claim 1, wherein in the top surface specifying step, the top surface is specified using a unit normal vector of each point of the point cloud data.

4. 2. The measurement method according to claim 1, wherein in the top surface specifying step, the top surface is specified based on a difference in distance from the plane between each point and its adjacent point.

5. 5. The measurement method according to claim 1, wherein, in the center identifying step, when there are a plurality of suspended loads, the center position is identified from the suspended loads at both ends.

6. 5. The measurement method according to claim 1, wherein in the center identification step, the width of the point cloud on the top surface is divided into equal sections equal to the number of the suspended loads, and the center of each section is set as the center position.

7. A program for causing a computer to specify the edge position of each of a plurality of hanging loads having known widths and identical shapes, with the loads being aligned in the same direction on a plane, using a three-dimensional sensor, an acquisition step of acquiring point cloud data in which distances to objects including the plurality of hanging loads are measured by the three-dimensional sensor from the opposite side of the plane with respect to the plurality of hanging loads; a top surface identification step of identifying a point cloud of the top surfaces of the plurality of suspended loads from the point cloud data; a center identifying step of identifying the center position of the top surface of each of the suspended loads based on the number of the suspended loads; an edge identifying step of identifying an edge position of the suspended load from the center position based on the width of the suspended load; A program that includes:

Images