Ship hold estimation device, ship hold estimation method, ship hold estimation program, and unloading device

Abstract

To provide a ship hold estimation device and the like that can effectively estimate the ship hold using a range sensor.SOLUTION: A ship hold estimation device 300 according to the present invention comprises: a range-finding point cloud acquisition unit 310 that acquires a range-finding point cloud 320 on the surface of the ship hold by range sensors 18, 19; a ship hold model generation unit 330 that generates a ship hold model 340 that defines the state of the ship hold; a deviation evaluation unit 360 that evaluates a deviation between each of the range-finding points in the range-finding point cloud 320 and the ship hold model 340; and a ship hold model adoption unit 370 for adopting the ship hold model 340 deviation of which falls within an acceptable range as the one representing the actual state of the ship hold. The ship hold model 340 includes at least one of a ship hold upper surface sub-model 341 that defines the upper surface facing the edge of the ship hold, a ship hold side wall surface sub-model 342 that defines the side wall surface facing the edge of the ship hold, and a cargo upper surface sub-model 343 that defines the upper surface of the cargo stored in the ship hold.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present invention relates to a hold estimation device that can be used in unloading equipment for unloading cargo from a ship. [Background technology]

[0002] As an unloading device for unloading cargo or load from a ship, a lifting device that unloads the cargo loaded on the ship onto land is known. Among such lifting devices, those that load and unload bulk cargo or bulk materials such as coal and iron ore are also called unloaders. They are also sometimes called continuous unloaders or continuous ship unloaders (Continuous Ship Unloaders) because they continuously load and unload bulk materials loaded on the ship. In this specification, the abbreviation CSU is used.

[0003] Patent Document 1 discloses a technology for deriving the relative position between the unloader device and the ship based on the results of edge detection of the upper part of the hold by a distance measurement sensor (laser sensor). In this technology, the edge is detected based on a series of distance measurement points acquired through laser light that is continuously irradiated by the distance measurement sensor onto the upper surface and side wall surface facing the edge of the hold. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-131394 A Summary of the Invention [Problem to be solved by the invention]

[0005] The distance measurement sensor is provided in a scraper or bucket elevator that is inserted into the hold to scrape off bulk cargo such as coal. Therefore, there is a risk that the laser light of the distance measurement sensor will be blocked by dust or steam from the bulk cargo, disrupting the continuity of the distance measurement point cloud and making it impossible to properly detect the edge of the hold.

[0006] The present invention has been made in consideration of these circumstances, and aims to provide a hold estimation device etc. that can effectively estimate a hold using a distance measuring sensor. [Means for solving the problem]

[0007] In order to solve the above problems, a hold estimation device of one embodiment of the present invention comprises a ranging point cloud acquisition unit that acquires a ranging point cloud on the surface of the hold using a ranging sensor, a hold model generation unit that generates a hold model that defines the state of the hold, a deviation evaluation unit that evaluates the deviation between each ranging point in the ranging point cloud and the hold model, and an estimation unit that estimates the actual state of the hold based on the deviation.

[0008] In this embodiment, the deviation of each measurement point in the measurement point cloud from the hold model is evaluated. Therefore, even if a small number of measurement points are disturbed by dust or steam from bulk cargo stored in the hold, the actual state of the hold can be appropriately estimated based on a comparison (deviation evaluation) of the remaining (typically a large number) normal measurement points with the hold model.

[0009] Another aspect of the present invention is a hold estimation method, which includes acquiring a range point cloud on a surface of a hold by a range sensor, generating a hold model defining a state of the hold, evaluating a deviation between each range point in the range point cloud and the hold model, and estimating an actual state of the hold based on the deviation.

[0010] Yet another aspect of the present invention is an unloading device comprising: a range point cloud acquisition unit that acquires a range point cloud on a surface of a hold by a range sensor, a hold model generation unit that generates a hold model that defines a state of the hold, a deviation evaluation unit that evaluates a deviation between each range point in the range point cloud and the hold model, an estimation unit that estimates a hold model that represents an actual state of the hold based on the deviation, and an unloading unit that unloads cargo stored in the hold using the hold model estimated by the estimation unit.

[0011] Any combination of the above components, or any expression of these components converted into a method, device, system, recording medium, computer program, or the like, is also included in the present invention. Effect of the Invention

[0012] According to the present invention, the cargo hold can be effectively estimated using distance sensors. [Brief description of the drawings]

[0013] [Figure 1] FIG. 2 is a front view showing the overall configuration of the lifting device. [Diagram 2] FIG. 2 is a perspective view showing the overall configuration of the lifting device. [Diagram 3] The detailed configuration of the lifting section is shown. [Figure 4] The external appearance of the distance measuring sensor is shown. [Diagram 5] FIG. 4 is a top view showing an example of the arrangement of distance measuring sensors. [Figure 6] FIG. 2 is a functional block diagram of the hold estimation device. [Figure 7] A model of the ship's hold is shown diagrammatically. [Figure 8] The ship's hold model is shown diagrammatically. [Figure 9] The relationship between the CSU coordinate system, hatch coordinate system, and edge coordinate system is shown diagrammatically. [Figure 10] Schematic diagram of the evaluation of the distance between each measurement point in the measurement point cloud and the hold top surface sub-model. [Figure 11] Schematic diagram of the evaluation of the distance between each measurement point in the measurement point cloud and the hold side wall sub-model. [Figure 12] Schematic diagram showing the evaluation of the distance between each measurement point in the measurement point cloud and the cargo top surface sub-model. [Figure 13] 13 is a flowchart showing an example of a hold estimation process performed by the hold estimation device. [Figure 14] A modified version of the ship's hold model is shown. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Hereinafter, a form for carrying out the present invention (hereinafter also referred to as an embodiment) will be described in detail with reference to the drawings. In the description and / or drawings, the same or equivalent components, members, processes, etc. will be given the same reference numerals, and duplicated explanations will be omitted. The scale and shape of each part shown in the drawings are set for convenience in order to simplify the explanation, and are not to be interpreted as being limiting unless otherwise specified. The embodiment is an example, and does not limit the scope of the present invention in any way. All features described in the embodiment and their combinations are not necessarily essential to the present invention.

[0015] FIG. 1 shows the overall configuration of a lifting device 1 as an unloading device according to an embodiment of the present invention. The lifting device 1 is a continuous unloader or a ship continuous unloader that unloads bulk cargo M as cargo or ship cargo loaded on a ship 200 to land. Hereinafter, the lifting device 1 is also referred to as a CSU1. The CSU1 continuously carries out bulk cargo M stored in a hold 201 of a ship 200 berthed at a quay 101 of a wharf 102 of a port or the like to land. Examples of the bulk cargo M include coal, coke, ore, etc. The CSU1 is operated by an operator in a main operation room 16 provided in the main body of the CSU1. The operation room for operating the CSU1 may be provided in a location other than the CSU1, or may be provided in any location on land outside the CSU1.

[0016] The wharf 102 where the ship 200 comes alongside constitutes the land where the bulk cargo M is unloaded, and is made of high-strength materials such as reinforced concrete. As shown in the perspective view of FIG. 2, the wharf 102 is provided with a pair of parallel rails 3 as a track along the longitudinal direction (perpendicular to the paper surface of FIG. 1) of the ship 200 that is berthed and anchored at the quay 101. The rails 3 constitute a track along which the traveling part 2 as a moving part of the CSU 1 can move or run. The rails 3 enable the CSU 1 to move relative to the anchored ship 200. As shown in FIG. 2, the installation direction of the rails 3 is preferably aligned with the longitudinal direction of the anchored ship 200 or the quay 101, but may be any other direction. The rails 3 may also include curved or bent parts. When unloading cargo from the ship 200, the CSU 1 moves on the rails 3 to a position close to the opening 21 of the hold 201 from which the cargo is to be unloaded. Thereafter, the traveling section 2, the revolving frame 5 (revolving section), and the lifting section 9 (unloading section or unloading device) are driven to unload the bulk cargo M from the hold 201.

[0017] At the wharf 102, a belt conveyor 45 is provided between the pair of rails 3 as a conveyor for transporting the unloaded bulk goods M in a certain direction. As shown in FIG. 2, the installation direction of the belt conveyor 45, i.e., the transport direction, is preferably aligned with the installation direction of the rails 3, but may be any other direction. The belt conveyor 45 may also include curved or bent portions. The belt conveyor 45 needs to be provided between the pair of rails 3 at the location where the bulk goods M unloaded from the CSU 1 are received, but may be provided outside the pair of rails 3 at other locations.

[0018] The CSU 1 comprises a traveling section 2 serving as a moving section movable relative to the ship 200, a revolving frame 5 constituting a rotating section capable of rotating relative to the traveling section 2, and a lifting section 9 provided at the tip of the revolving frame 5 and serving as an unloading section or unloading device for unloading bulk cargo M. The revolving frame 5 is supported on the traveling section 2 so as to be rotatable about a rotation axis in the vertical direction (the up-down direction in FIG. 1). The revolving frame 5 is provided with a boom 7 extending laterally and intersecting the rotation axis, and a bucket elevator constituting the main part of the lifting section 9 is supported at its tip.

[0019] The lifting unit 9 maintains a vertical position regardless of the hoisting angle of the boom 7 (the rotation angle around the hoisting axis perpendicular to the plane of FIG. 1) by a parallel link mechanism formed between the revolving frame 5, the boom 7, and the parallel link 8. A counterweight 13 is provided at the rear end of the revolving frame 5 opposite the tip of the boom 7. The counterweight 13 is connected to the tip of the boom 7 via a balancing lever 12. The action of this counterweight 13 puts the lifting unit 9 in a substantially unloaded state, achieving a stable load balance. Note that the main components of the revolving unit, such as the revolving frame 5, the boom 7, the balancing lever 12, and the counterweight 13, may be collectively referred to as the main body below.

[0020] A cylinder 15 is provided to adjust the hoisting angle of the boom 7. When the cylinder 15 is at its standard length, the hoisting angle is 0 degrees, i.e. the boom 7 is parallel or horizontal to the ground (left and right in Figure 1). When the cylinder 15 is extended beyond its standard length, the tip of the boom 7 rises, resulting in a positive hoisting angle. When the cylinder 15 is shortened beyond its standard length, the tip of the boom 7 descends, resulting in a negative hoisting angle. The lifting unit 9 supported at the tip of the boom 7 rises while maintaining a vertical position when the hoisting angle of the boom 7 increases, and descends while maintaining a vertical position when the hoisting angle of the boom 7 decreases.

[0021] A main operation room 16 for operating the CSU 1 is provided in the main body. Specifically, the main operation room 16 is provided on the lifting unit 9 side of the revolving frame 5. An operator in the main operation room 16 can safely operate the CSU 1 while visually checking the lifting unit 9. Parameters related to the position and attitude of the CSU 1, such as the position of the traveling unit 2, the rotation angle of the revolving frame 5, and the elevation angle of the boom 7, are controlled according to the operation of the main operation room 16. In addition, the operation of carrying out the bulk load M by the lifting unit 9 can also be operated from the main operation room 16.

[0022] The lifting section 9 includes a scraping section 11 that scrapes off bulk goods M, and a bucket elevator as an elevator section that transports the scraped off bulk goods M upward. The scraping section 11 is provided at the bottom of the lifting section 9, and continuously excavates and scrapes off the bulk goods M in the hold 201 with a number of buckets 27 (see FIG. 3) that are movably provided along its periphery. The scraped off bulk goods M are transported upward together with the buckets 27 by the bucket elevator.

[0023] Fig. 3 shows a detailed configuration of the lifting section 9. The bucket elevator comprises a cylindrical elevator body 14 extending in the vertical direction, and a chain bucket 29 that moves in circles relative to the elevator body 14. The chain bucket 29 comprises a pair of roller chains 25, each of which is an endless chain, and a plurality of buckets 27, both sides of which are supported by the pair of roller chains 25. Specifically, the pair of roller chains 25 are arranged side by side in a direction perpendicular to the plane of Fig. 3(B), and each bucket 27 is attached so as to be suspended between the pair of roller chains 25.

[0024] The bucket elevator includes a driving roller 31a that guides the stretched roller chain 25, driven rollers 31b and 31c, and a deflecting roller 33. The driving roller 31a is provided at the top 9a of the bucket elevator, and is rotated by a motor or the like (not shown) to cause the chain bucket 29 to move in a circular motion. The driven roller 31b is provided in front of the scraping unit 11 (left side in FIG. 3(B)), and the driven roller 31c is provided behind the scraping unit 11 (right side in FIG. 3(B)), and each of them guides the circularly moving chain bucket 29. The deflecting roller 33 is a driven roller provided below the driving roller 31a, and guides the circularly moving chain bucket 29 and changes the direction of its movement. An expandable cylinder 35 is provided between the driven roller 31b and the driven roller 31c. When the cylinder 35 expands and contracts, the distance between the axes of the driven rollers 31b, 31c changes, changing the trajectory of the orbital motion of the chain bucket 29. The expansion and contraction control of the cylinder 35 may be performed by operation of the main operation room 16, or may be performed automatically according to a program by a computer built into the CSU 1. Since two roller chains 25 are provided, two each of the drive roller 31a, driven rollers 31b, 31c, and deflection roller 33 are also provided, and are arranged side by side in a direction perpendicular to the plane of the paper in FIG. 3(B).

[0025] By the rotational drive of the drive roller 31a, the chain bucket 29 moves in an orbit relative to the elevator body 14. For example, the chain bucket 29 moves in an orbit counterclockwise direction along the arrow W shown in Fig. 3(B). At this time, the chain bucket 29 reciprocates between the scraping unit 11 provided at the bottom of the bucket elevator and the drive roller 31a provided at the top 9a of the bucket elevator.

[0026] Each bucket 27 of the chain bucket 29 rises inside the elevator body 14 while maintaining a posture in which its opening faces upward. When each bucket 27 passes through the driving roller 31a at the top 9a of the bucket elevator, its direction of motion changes from upward to downward, and the opening of each bucket 27 also turns from upward to downward. A discharge chute (not shown) is provided below the opening of each bucket 27 thus turned downward, and the bulk goods M picked up by each bucket 27 are discharged. The discharge chute discharges the bulk goods M onto a rotary feeder 37 (FIG. 1) provided on the outer periphery of the upper part of the lifting section 9.

[0027] The rotary feeder 37 rotates around a rotation axis in the extension direction of the elevator body 14, i.e., the vertical direction, and transfers the bulk goods M discharged from the discharge chute to a boom conveyor 39 of the boom 7. The boom conveyor 39 transports the bulk goods M within the boom 7 to the vicinity of the rotation axis of the rotating frame 5 and supplies them to a hopper (not shown) provided there. An internal conveyor 43 that receives the bulk goods M is provided within the running section 2 below the discharge opening of the hopper. The internal conveyor 43 transfers the bulk goods M to the aforementioned belt conveyor 45 provided on the wharf 102 as land.

[0028] Next, a description will be given of a basic lifting operation of the CSU 1 having the above configuration. In this lifting operation, the lifting unit 9 and / or the CSU 1 function as a discharge device that discharges bulk material M (cargo) in the hold 201 of the ship 200 out of the hold 201.

[0029] The operator of the CSU1 operates the CSU1 in the main operation room 16. First, the traveling section 2 travels on the rail 3 and moves it to a position close to the opening 21 of the hold 201 to be unloaded. Next, the revolving frame 5 is rotated about a vertical rotation axis provided at a position overlapping the traveling section 2 in a top view (when viewed from above in FIG. 1), and the lifting section 9 provided at the tip of the boom 7 is moved above the opening 21 of the hold 201 to be unloaded. Here, it is preferable to perform the traveling operation and the revolving operation with the lifting section 9 raised by raising the boom 7 in the positive direction (clockwise direction in FIG. 1) so that the lifting section 9 does not collide with the wharf 102 or the ship 200. Next, the boom 7 is raised and lowered in the negative direction (counterclockwise direction in FIG. 1), and the scraping section 11 provided at the tip of the lifting section 9 is inserted into the hold 201 through the opening 21. The movement of the traveling section 2, the rotation of the revolving frame 5, and the raising and lowering of the boom 7 may be performed simultaneously.

[0030] After the scraping unit 11 is inserted into the hold 201, the roller chain 25 is caused to make a circular motion along the arrow W. When a plurality of buckets 27 attached to the roller chain 25 make a circular motion integrally with the roller chain 25, they excavate and scrape off the bulk goods M stored in the hold 201. The bulk goods M scraped off by each bucket 27 are transported upward in the elevator body 14 as the roller chain 25 makes a circular motion.

[0031] The scraping unit 11 appropriately changes its three-dimensional position in the hold 201 in order to efficiently scrape off the bulk cargo M at various locations in the hold 201. For example, when the surface position of the bulk cargo M becomes lower as the lifting work progresses, the boom 7 is raised and lowered in the negative direction to lower the scraping unit 11. In addition, in order to scrape off the bulk cargo M near the wall of the hold 201, the running unit 2 and / or the rotating frame 5 may be operated to change the position of the scraping unit 11 in the horizontal plane. The scraping unit 11 can change not only its three-dimensional position but also its attitude and shape. For example, the scraping unit 11 can rotate around a rotation axis in the extension direction of the elevator body 14, i.e., the vertical direction, and its orientation can be changed arbitrarily. In addition, as shown by the dashed line in FIG. 3(B), the scraping unit 11 can take an inclined shape or a horizontally elongated shape that is contracted vertically and extended horizontally. As a result, even if the hold 201 has a large horizontal distance from the opening 21 to the wall, the scraping section 11 can be brought close to the wall and the bulk cargo M can be scraped off efficiently.

[0032] Changes in the position, posture, and shape of the scraping unit 11 (lifting unit 9) within the hold 201 in relation to the lifting operation of CSU1 as described above may be performed autonomously by CSU1 using a ranging sensor or camera described below (i.e., the lifting unit 9 and / or CSU1 may be operated automatically), or may be performed manually by an operator in the main operation room 16 while communicating with workers in the hold 201.

[0033] The bucket 27 that has scraped off the bulk goods M in the hold 201 rises in the elevator body 14 and turns from an upward to a downward direction when passing through the drive roller 31a at the top 9a. The bulk goods M that fall as the bucket 27 turns enters a discharge chute and is discharged onto a rotary feeder 37. Thereafter, the bulk goods M are transferred via a boom conveyor 39 and an internal conveyor 43 to a belt conveyor 45 provided on a wharf 102 as land. The above-mentioned carrying-out operation is repeated by the multiple buckets 27, so that the bulk goods M in the hold 201 are continuously unloaded.

[0034] Next, a description will be given of a distance measurement sensor provided in the CSU 1 to improve the safety and efficiency of unloading. The distance measurement sensor constitutes a hold detection unit or a position measurement unit that detects a part of the hold 201, such as the edge of the opening 21, the top / side surface facing the edge, the ceiling / wall / bottom of the hold 201, the position of structures inside the hold 201, etc.

[0035] As shown in FIG. 1, the upper part of the lifting section 9 is provided with a plurality of distance measuring sensors 19 for measuring the distance to the measurement objects below and to the sides. During lifting as shown in the figure, the measurement objects of the distance measuring sensors 19 are the edge of the opening 21, the ceiling / wall / bottom of the hold 201, bulk cargo M and other objects, people / structures in the hold 201, a bulldozer for bottom raking, the scraping section 11, the ship 200, other parts of the CSU 1 such as the boom 7 / rotating frame 5 / traveling section 2 / main operation room 16, the quay 101, the wharf 102, the rail 3, the belt conveyor 45, etc. The plurality of distance measuring sensors 19 may be disposed, for example, on the upper part of the cylindrical elevator body 14 so as to surround the outer periphery of the elevator body 14. Alternatively, the plurality of distance measuring sensors 19 may be provided on a flange section 91 that rotatably supports the upper part of the elevator body 14 so as to surround the outer periphery of the elevator body 14. It is preferable that the multiple distance measuring sensors 19 are provided below the connection part between the lifting unit 9 and the boom 7 so that the boom 7 does not enter the measurement range below and to the sides of the multiple distance measuring sensors 19. On the other hand, if the multiple distance measuring sensors 19 are provided above the connection part between the lifting unit 9 and the boom 7, each distance measuring sensor 19 should be provided at a position that does not overlap with the boom 7 when viewed from above (when viewed from above in Figure 1). An example of the arrangement of the multiple distance measuring sensors 19 when viewed from above will be described later. The number of distance measuring sensors 19 is arbitrary. For example, any number of distance measuring sensors 19 that measure distance mainly below the lifting unit 9 and any number of distance measuring sensors 19 that measure distance mainly to the sides of the lifting unit 9 may be provided.

[0036] The scraping section 11 at the bottom of the lifting section 9 is provided with a plurality of distance measuring sensors 18 for measuring the distance to the measurement object above, below, and to the side. During lifting as shown in the figure, the measurement objects of the distance measuring sensors 18 are the edge of the opening 21, the ceiling / wall / bottom of the hold 201, the bulk goods M and other objects, people / structures in the hold 201, a bulldozer for bottom raking, and other parts of the CSU 1 such as the boom 7. The distance measuring sensors 18 are provided at the front (left part of FIG. 1) and rear (right part of FIG. 1) of the scraping section 11, respectively. In order to avoid deterioration of the measurement accuracy due to dust of the bulk goods M scraped by the bucket 27 of the scraping section 11, it is preferable that the plurality of distance measuring sensors 18 are provided at a position (upper part of the scraping section 11) away from the place where the bucket 27 excavates the bulk goods M (lower part of the scraping section 11). The number of distance measuring sensors 18 is arbitrary. For example, any number of distance measuring sensors 18 that measure distances centered on the sides of the scraping unit 11 and any number of distance measuring sensors 18 that measure distances centered on the bottom of the scraping unit 11 may be provided.

[0037] 4 shows the appearance of the distance measuring sensors 18, 19. The distance measuring sensors 18, 19 are laser sensors capable of measuring distances, and include a laser emitting unit (not shown) as an emitting unit that emits laser light toward a measurement object including the hold 201, and a laser receiving unit (not shown) as a receiving unit that receives the laser light reflected by the measurement object, constituting a distance measuring unit that measures the distance to the measurement object. A light transmitting portion 171 through which the laser light can pass is formed in an endless band shape around the entire circumference of the side surface of the cylindrical housing 17 of the distance measuring sensors 18, 19.

[0038] A plurality of laser emission units are provided at positions facing the light-transmitting portion 171 in the housing 17, and emit linear laser light to the outside of the housing 17 through the light-transmitting portion 171. The laser emission units are arranged at a predetermined interval along the direction of the axis A of the housing 17 (the vertical direction in FIG. 4), but FIG. 4 shows the laser light emitted from one point for simplicity. As shown in the schematic diagram, the emission angles of the laser emission units are different from each other by about 0.1 degrees to 3 degrees. With this configuration, the distance measuring sensors 18 and 19 can irradiate laser light within a predetermined angle range (in the range of θ- to θ+ in the figure) above and below the reference plane S, which is a plane perpendicular to the axis A of the housing 17. θ- and θ+ can be designed arbitrarily, but hereinafter, -θ-=θ+=15 degrees. At this time, the distance measuring sensors 18 and 19 irradiate laser light within a range of ±15 degrees centered on the reference plane S. Moreover, these multiple laser emission units are integrally provided so as to be rotatable 360 ​​degrees around the axis A of the housing 17. With this configuration, the distance measuring sensors 18, 19 can irradiate laser light to all measurement objects around (to the sides of) the housing 17. Note that the laser emission units in the distance measuring sensors 18, 19 may irradiate laser light in any angular range less than 360 degrees around the axis A of the housing 17 (for example, an angular range of 180 degrees or less or an angular range of 120 degrees or less). Also, it is preferable to use laser light of an invisible wavelength such as near-infrared light so as not to disturb people inside or around the CSU 1 or the ship 200.

[0039] Distance measuring sensors 18, 19 rotate multiple laser emitters together and emit pulsed laser light at each predetermined rotation angle. The pulsed laser light emitted by each laser emitter is reflected or scattered by the object to be measured and returns to distance measuring sensors 18, 19, where it is received by a laser receiver provided together with each laser emitter in housing 17. A calculation unit (not shown) of distance measuring sensors 18, 19 calculates the distance to the object to be measured based on the time from when the laser emitter emits a pulse of laser light to when the laser receiver receives the reflected pulse of laser light. This technology is also called LIDAR (Light Detection and Ranging or Laser Imaging Detection and Ranging).

[0040] Although a laser sensor has been given as an example of the distance measuring sensors 18 and 19 in the above, the distance measuring sensors 18 and 19 may be sensors that use other light or electromagnetic waves. For example, a millimeter wave sensor using so-called millimeter waves with a wavelength of about 1 mm to 10 mm may be used as the distance measuring sensors 18 and 19. Since the frequency of millimeter waves is high, about 30 GHz to 300 GHz, they have high linearity and can be treated like a laser. The millimeter wave sensor can be configured in the same manner as the laser sensor in FIG. 4, and a millimeter wave transmitter (broadly defined light emitting unit) that emits millimeter waves to the measurement object instead of the laser light emitting unit, and a millimeter wave receiver (broadly defined light receiving unit) that receives millimeter waves reflected by the measurement object instead of the laser light receiving unit may be provided. In addition, an optical sensor that uses light other than laser light, such as a Time of Flight (ToF) type image sensor, may be used as the distance measuring sensors 18 and 19. In addition, the distance measuring sensors 18 and 19 may not have a light emitting unit that emits light or electromagnetic waves to the measurement object. For example, the distance measuring sensors 18 and 19 may be stereo cameras capable of measuring distances by simultaneously photographing the object to be measured from different directions.

[0041] The distance measuring sensors 18 and 19 in Fig. 4 are attached to the CSU 1 in Fig. 1 in any position and in any posture according to the purpose of measurement. For example, the distance measuring sensor 18 of the scraping unit 11 is attached so that the axis A in Fig. 4 is vertical and the reference plane S is a horizontal plane. In this case, the distance measuring sensor 18 can measure the distance inside the hold 201 with the side of the scraping unit 11 as the center. The distance measuring sensor 18 may also be attached so that the axis A in Fig. 4 is horizontal and the reference plane S is a vertical plane. In this case, the distance measuring sensor 18 can measure the distance to the opening 21 above the scraping unit 11 and the bulk material M below the scraping unit 11. The direction of the axis A of the distance measuring sensor 18 is not limited to the vertical or horizontal direction and may be any direction.

[0042] The distance measuring sensor 19 on the top of the lifting section 9 is attached so that the axis A in FIG. 4 is horizontal and the reference plane S is vertical. In this case, the distance measuring sensor 19 can measure the edge of the opening 21 of the hold 201 below and the bulk cargo M in the hold 201. The distance measuring sensor 19 can also emit laser light upward, but since there is no measurement target above, the upward distance measurement may be disabled by covering the upper side of the distance measuring sensor 19 with a light-shielding cover. The distance measuring sensor 19 may also be attached so that the axis A in FIG. 4 is vertical and the reference plane S is parallel to the horizontal plane. In this case, the distance measuring sensor 19 can efficiently measure the distance to the measurement target outside the hold 201 to the side. The direction of the axis A of the distance measuring sensor 19 may be any direction, not limited to the horizontal or vertical direction, but the case of the horizontal direction will be described in detail below.

[0043] By providing the distance measuring sensors 18, 19 as described above in the lifting unit 9, it is possible to accurately grasp the positions of various measurement objects such as the edge of the opening 21, the ceiling / wall / bottom of the hold 201, bulk cargo M and other objects, people / structures in the hold 201, a bulldozer for raking the bottom, and the scraping unit 11. Therefore, it is possible to prevent the lifting unit 9 from colliding with other objects during lifting, and the bulk cargo M can be lifted efficiently.

[0044] FIG. 5 shows an example of the arrangement of the distance measuring sensor 19 from a top view. Three distance measuring sensors 191, 192, and 193 as the distance measuring sensor 19 are arranged to surround the outer periphery of the flange portion 91 or the elevator main body 14. The distance measuring sensor 191 is arranged so that the axis A in FIG. 4 is the left-right direction in FIG. 5, and the reference plane S1 corresponding to the reference plane S in FIG. 4 is the up-down direction in FIG. 5. The distance measuring sensor 191 measures distance by irradiating a laser beam within a range of ±15 degrees centered on the reference plane S1. The distance measuring sensors 192 and 193 are arranged so that the axis A in FIG. 4 is the up-down direction in FIG. 5, and the reference planes S2 and S3 corresponding to the reference plane S in FIG. 4 are the left-right direction in FIG. The distance measuring sensors 192 and 193 measure distance by irradiating a laser beam within a range of ±15 degrees centered on the reference planes S2 and S3. The reference planes S2 and S3 of the distance measuring sensors 192 and 193 are different planes parallel to each other and perpendicular to the reference plane S1 of the distance measuring sensor 191.

[0045] The CSU 1 unloads bulk cargo M from the hold 201 with the attitude shown in Fig. 5 as the basic attitude for unloading. In this basic attitude, the traveling part 2 is in a position shifted from the front position of the hold 201, and the rotating frame 5 and the boom 7 are in a rotating position forming an acute angle with the rail 3 that forms the track of the traveling part 2. At this time, the unloading part 9 is located above the hold 201 of the ship 200, and the scraping part 11 at its lower part is inserted into the hold 201 through the opening 21.

[0046] The opening 21 of the hold 201 is often rectangular and long in the direction of travel of the ship 200 (horizontal direction in FIG. 5). In this case, the upper edge E11 and the lower edge E12 of the opening 21 can be detected by the distance measuring sensor 191 that irradiates laser light parallel to the short side (vertical side in FIG. 5) of the opening 21. Note that the points shown at the centers of the edges E11 and E12 represent the positions where the laser light on the reference plane S1 of the distance measuring sensor 191 hits the edge of the opening 21, and the rectangle surrounding it typically represents the range where the laser light irradiated within a range of ±15 degrees from the reference plane S1 hits the edge of the opening 21. Hereinafter, the same notation is used for the distance measuring sensors 192 and 193.

[0047] Similarly, distance measuring sensors 192, 193 that irradiate laser light parallel to the long sides (the left and right sides in FIG. 5) of opening 21 can detect edges E21, E31 on the left side and edges E22, E32 on the right side of opening 21. By using two distance measuring sensors 192, 193, distance measurement can be performed with high accuracy even in the long direction, which is more difficult to measure distances in than the short direction. In this way, the arrangement of distance measuring sensors 191, 192, 193 in FIG. 5 is suitable for detecting the edges of opening 21 that has a shape that is long in one direction, such as a rectangle.

[0048] Even if CSU1 is not in the basic posture shown in Figure 5, if the loading section 9 is within the opening 21 when viewed from above, the three distance measuring sensors 191, 192, and 193 can acquire six distance measuring point groups on the edge of the opening 21 corresponding to E11, E12, E21, E22, E31, and E32, and the position of the opening 21 can be accurately determined.

[0049] 5, the basic position of the CSU1 during lifting may be such that the travel unit 2 is in front of the hold 201 and the revolving frame 5 and the boom 7 are perpendicular to the rail 3. In this case, the extension direction of the boom 7 coincides with the short side direction of the opening 21, so that the reference plane S1 of the distance measuring sensor 191 is parallel to the extension direction of the boom 7 and the reference planes S2 and S3 of the distance measuring sensors 192 and 193 are perpendicular to the extension direction of the boom 7. If the distance measuring sensors 191, 192, and 193 are integrally rotatable around the axis of the cylindrical elevator body 14, it is possible to easily arrange the distance measuring sensors 191, 192, and 193 in a manner suitable for the elongated opening 21, depending on the change in the basic position of the CSU1 during lifting.

[0050] The number and arrangement of the distance measuring sensors 19 described above are merely an example, and any number and arrangement may be adopted. The number of distance measuring sensors 19 is preferably at least two in order to efficiently measure the state of the hold 201 surrounding the lifting section 9 in a top view, such as its position, attitude, shape, etc. More preferably, it is three or more. The multiple distance measuring sensors 19 may be arranged at equal intervals along the outer periphery of the flange section 91 or the elevator body 14. In this case, the installation orientation of each distance measuring sensor 19 is arbitrary, but for example, each distance measuring sensor 19 is installed so that the reference surface S of the distance measuring sensor 19 is in contact with the outer periphery of the flange section 91 or the elevator body 14. With such a symmetrical arrangement, the state of the hold 201, such as its position, attitude, shape, etc. can be stably measured regardless of the attitude of the CSU 1 during lifting.

[0051] By controlling each movable part of the CSU 1, i.e., the movable traveling part 2, the rotatable rotating frame 5, the riseable boom 7, the rotatable and deformable scraping part 11, etc., according to the distance to the hold 201 itself or to an object inside or outside the hold 201 measured by the distance measuring sensors 18, 19 as described above, it is possible to prevent the lifting part 9 from colliding with the hold 201 itself or other objects inside or outside the hold 201 during lifting, and to efficiently lift the bulk cargo M. Note that in addition to or instead of the distance measuring sensors 18, 19, the hold 201 itself or objects inside or outside the hold 201 may be detected by an optical sensor such as an image sensor or a camera that photographs the object to be measured.

[0052] FIG. 6 is a functional block diagram of a hold estimation device 300 according to this embodiment, which can effectively estimate the state of the hold 201 using the distance measurement sensors 18 and 19. The hold estimation device 300 includes a distance measurement point cloud acquisition unit 310, a hold model generation unit 330, a coordinate conversion unit 350, a deviation evaluation unit 360, and a hold model selection unit 370 that may function as an estimation unit. Some of these functional blocks may be omitted as long as the hold estimation device 300 can achieve at least some of the actions and / or effects described below. These functional blocks are realized by the cooperation of hardware resources such as a central processing unit, memory, input device, output device, and peripheral devices connected to the computer, and software executed using them. Regardless of the type of computer or the location where the computer is installed, each of the above functional blocks may be realized by the hardware resources of a single computer, or by combining hardware resources distributed among multiple computers.

[0053] The distance measurement point cloud acquisition unit 310 acquires a distance measurement point cloud 320 on the surface of the hold 201 by the distance measurement sensors 18, 19. As shown in FIG. 7, the hold 201 is typically a cavity having a substantially rectangular parallelepiped shape. An opening 21 called a hatch is provided on the upper surface of the hold 201. In the example of FIG. 7, the shape and size of a cross section at any height (position in the z direction) of the hold 201 are the same as the shape and size of the hatch, but the shape of the hatch may be different from the shape of a cross section inside the hold 201, and the size of the hatch may be smaller than the size of a cross section inside the hold 201. The illustrated xyz coordinate system is a hatch coordinate system set in the hold 201. Here, the x direction is the long direction of the hold 201, the y direction is the short direction of the hold 201, and the z direction is the height direction of the hold 201. O is a representative point representing the position of the hold 201, and an arbitrary position (p x ,p y ,p z )

[0054] Examples of surfaces of the hold 201 from which the distance measurement point cloud acquisition unit 310 acquires the distance measurement point cloud 320 include top surfaces U1 to U4 facing the linear edges E1 to E4 at the upper part of the hold 201, side wall surfaces W1 to W4 (W1 and W4 are hidden in Figure 7) facing the edges E1 to E4, and the top surface of bulk cargo M (not shown in Figure 7) as cargo.

[0055] As will be described later, the measurement point cloud 320 acquired by the measurement point cloud acquisition unit 310 is compared with a plurality of hold models (hold models) 340 similar to that shown in Fig. 7, which are slightly different from each other in position, attitude, shape, etc. For this purpose, it is sufficient for the measurement point cloud acquisition unit 310 to measure the distance to a part of the surface of the hold 201 by the distance measurement sensors 18, 19. However, as described below, when the hold model 340 includes sub-models such as a hold top surface sub-model 341 corresponding to the top surfaces U1 to U4, a hold side wall surface sub-model 342 corresponding to the side wall surfaces W1 to W4, and a cargo top surface sub-model 343 corresponding to the top surface of bulk cargo M, it is preferable that the measurement point cloud acquisition unit 310 acquires a sufficient number and range of measurement point clouds 320 on each surface of the hold 201 to which each sub-model 341, 342, 343 corresponds (top surfaces U1 to U4, side wall surfaces W1 to W4, top surface of bulk cargo M) for comparison with each of the sub-models 341, 342, 343.

[0056] The hold model generating unit 330 generates a hold model 340 that defines the state of the hold 201. Here, the state of the hold 201 is a concept that includes at least one of the position, posture, and shape of the hold 201 itself and / or cargo such as bulk cargo M stored in the hold 201. The hold model 340 may include at least one of a hold upper surface sub-model 341 that defines upper surfaces U1 to U4 facing the edges E1 to E4 of the hold 201, a hold side wall surface sub-model 342 that defines side wall surfaces W1 to W4 facing the edges E1 to E4 of the hold 201, and a cargo upper surface sub-model 343 that defines the upper surface of the bulk cargo M stored in the hold 201.

[0057] 8 diagrammatically shows a hold upper surface sub-model 341, a hold side wall surface sub-model 342, and a cargo upper surface sub-model 343 that may be included in the hold model 340. The hold model 340 is a model that can be generated arbitrarily by the hold model generation unit 330 in order to simulate or estimate the state of the hold 201. Therefore, hereinafter, the same reference numerals are used for the hold 201 and the hold model 340 for convenience.

[0058] The hold model 340 is constructed based on various illustrated parameters that can be arbitrarily generated by the hold model generation unit 330. Specifically, the position (x, y, z) and the attitude (θ x ,θ y ,θ z The hold model 340 is constructed based on various parameters such as the width w of the hold model 340 (hold 201) in the y direction, the width γ of the hold top surface sub-model 341 (top surfaces U1 to U4) in the y direction, the height β of the hold side wall surface sub-model 342 (side wall surfaces W1 to W4) in the z direction, and the depth α of the cargo top surface sub-model 343 (top surface of bulk cargo M) in the z direction.

[0059] The position (x, y, z) of the representative point O of the hold model 340 is the same as the position (p x ,p y ,p z In this embodiment, the x-coordinate x(p x ) is, for example, the x-coordinate t of the bucket elevator (loading unit 9) known in the CSU 1. u,BE In this case, the hold model generation unit 330 uses the yz coordinates (p y ,p z ), the yz coordinates (y, z) of the representative point O of the hold model 340 are generated.

[0060] The attitude of the representative point O of the hold model 340 (θ x ,θ y ,θ z ) is used to estimate the attitude of the representative point O of the hold 201. x ,θ y ,θ z) are rotation angles (θ x ,θ y ,θ z )

[0061] The width w of the hold model 340 in the y direction is used to estimate the width of the hold 201 in the y direction, which is one aspect of the shape of the hold 201. As shown in the example, the y coordinate y(p y ) is preferably set to half the y-directional width w of the hold model 340 (i.e., y=w / 2), but when both the y-coordinate y and the width w have not yet been estimated, the respective values ​​may be set independently of each other by the hold model generation unit 330.

[0062] The hold upper surface sub-model 341, which has a width γ in the y direction, is intended to estimate the upper surfaces U1 to U4 of the hold 201. The hold upper surface sub-model 341 has, for example, the same z coordinate (hereinafter, for convenience, this may be treated as "0") as the representative point O of the hold model 340, and in the illustrated example, defines a line segment or plane extending with width γ on the negative side of the y direction (which also extends in the x direction perpendicular to the paper surface of FIG. 8).

[0063] The hold side wall surface sub-model 342, which has a height β in the z direction, is intended to estimate the side wall surfaces W1 to W4 of the hold 201. The hold side wall surface sub-model 342 defines a line segment or plane (which also extends in the x direction perpendicular to the paper surface of FIG. 8) extending at a height β on the negative side in the z direction from the +y side end of the hold upper surface sub-model 341 (which corresponds to the edges E1 to E4 of the hold 201, and as will be described later, an edge coordinate system "r" is defined with this point as the origin of the yz plane).

[0064] The cargo upper surface sub-model 343, having a depth α in the z direction, is intended to estimate the upper surface of the bulk cargo M stored in the hold 201. The cargo upper surface sub-model 343 defines a line segment or plane (which also extends in the x direction perpendicular to the paper surface of FIG. 8) extending in the positive y direction from a position a depth α away from the +y side end (the origin of the edge coordinate system "r") of the hold upper surface sub-model 341 in the negative z direction.

[0065] The coordinate conversion unit 350 converts the coordinates of one of the measurement points 320 and the hold model 340 from one coordinate system to the other coordinate system in preparation for comparison between the measurement points 320 and the hold model 340 in the subsequent deviation evaluation unit 360. For example, the CSU coordinate q of the measurement points 320 acquired in the CSU coordinate system (hereinafter, represented by the subscript "u") set in the CSU1 is u,i (i=1, ..., M: M is the total number of measurement points included in the measurement point group 320) is expressed as hatch coordinate q in the hatch coordinate system (hereinafter represented by the subscript "h") set in the hold 201 (hold model 340). h,i and / or an edge coordinate q in an edge coordinate system (hereinafter denoted by the subscript "r") substantially equivalent to the hatch coordinate system in question r,i Convert to.

[0066] Figure 9 shows a schematic diagram of the relationship between the CSU coordinate system "u", the hatch coordinate system "h", and the edge coordinate system "r". The origin of the CSU coordinate system "u" and the origin of the hatch coordinate system "h" are related by a three-dimensional translation t hu and the three-dimensional rotation R hu =R(θ hu ) exists (θ hu is a 3D vector that represents a 3D rotation. Here, the x-position of the bucket elevator in the CSU coordinate system "u" is t u,BE The origin of the hatch coordinate system "h" is set at the center of the y direction of the hold model 340 (hold 201), and the z position of the origin of the hatch coordinate system "h" is set at the height of the upper surfaces U1 to U4 of the hold model 340 (hold 201). In this way, the translation t hu is (t u,BE ,y,z)T Based on these, the point q in the CSU coordinate system "u" is u Let us consider point q in the hatch coordinate system "h". h The coordinate transformation to convert to is "q h =R hu T (q u -t hu )".

[0067] In addition, there is a three-dimensional translation t between the origin of the hatch coordinate system "h" and the origin of the edge coordinate system "r". rh and the three-dimensional rotation R rh =R(θ rh ) exists (θ rh Here, the hatch coordinate system "h" and the edge coordinate system "r" have the same orientation (rotation), so R rh =I (I is a unit matrix). In addition, the displacement between the origin of the hatch coordinate system "h" and the origin of the edge coordinate system "r" is only half the width w in the y direction of the hold model 340 (hold 201). Therefore, t rh =(0,-w / 2,0) T Based on these, the point q in the hatch coordinate system "h" is h Let us consider point q in the edge coordinate system "r". r The coordinate transformation to convert to is "q r =R rh T (q h -t rh )".

[0068] As shown in the above coordinate transformation from the CSU coordinate system “u” to the hatch coordinate system “h” and / or the edge coordinate system “r”, the main conditions of the hold model 340 (hold 201) such as its position, attitude, shape, etc. are determined by the parameter set X=(θ ru T ,y,z,w) T It is determined by (θ ru θ hu and θ rh This is a three-dimensional rotation of the rh Since there is no rotation, θ huThe deviation evaluation unit 360, which will be described next, searches for a parameter set X (i.e., the hold model 340) that adequately represents the actual state of the hold 201.

[0069] The deviation evaluation unit 360 evaluates the deviation between each ranging point in the ranging point group 320 acquired by the ranging point group acquisition unit 310 and the hold model 340 generated by the hold model generation unit 330. In the example of this embodiment, the deviation evaluation unit 360 includes a first evaluation unit 361 that evaluates the deviation between each ranging point in the ranging point group 320 and the hold model 340 using a RANSAC (Random Sample Consensus) algorithm, and a second evaluation unit 362 that evaluates the deviation between each ranging point in the ranging point group 320 and the hold model 340 using an EM (Expectation Maximization) algorithm. However, the deviation evaluation unit according to the present invention is not limited to such a two-stage deviation evaluation, and may be, for example, a one-stage deviation evaluation or a three-stage or more deviation evaluation. Furthermore, the algorithm used by the deviation evaluation unit according to the present invention is not limited to the RANSAC algorithm or the EM algorithm, and may be any algorithm as long as the actions and / or effects of the present invention and / or this embodiment are at least partially realized.

[0070] The deviation evaluation unit 360 evaluates the distance between each of the measurement points in the measurement point group 320 and the hold model 340 as a deviation. Specifically, the deviation evaluation unit 360 may evaluate the distance between each of the measurement points in the measurement point group 320 for each of the sub-models 341, 342, and 343 included in the hold model 340. Note that the evaluation of the distance is performed in the edge coordinate system "r" shown in FIG. 8 (i.e., both the measurement point group 320 and the hold model 340 are described on the edge coordinate system "r" through the above-mentioned coordinate conversion unit 350). The y-axis and z-axis shown simply in FIG. 10 to FIG. 12 are the y-axis and z-axis in the edge coordinate system "r".

[0071] As described below, each sub-model 341 (subscript “a”), 342 (subscript “b”), and 343 (subscript “c”) and the distance d of the distance measurement point q are a (q), db (q), d c (q) is expressed by the following equations, respectively.

number

[0072] 10 is a schematic diagram showing an evaluation of the distance between each of the measurement points in the measurement point cloud 320 and the hold upper surface sub-model 341. The deviation evaluation unit 360 evaluates the distance d in the yz plane of the hold upper surface sub-model 341 between each measurement point q included in the deviation evaluation area (measurement point cloud extraction area) of the measurement point cloud 320 shown in FIG. a Calculate (q). As shown in the above formula, the y coordinate q y For a range point q with negative z-coordinate q z The absolute value of is d a (q), and the y coordinate is q y For a distance measurement point q whose y coordinate is q y and z coordinate q z The square root of the sum of the squares of is d a The result is (q).

[0073] 11 is a schematic diagram showing an evaluation of the distance between each of the measurement points in the measurement point cloud 320 and the hold side wall surface sub-model 342. The deviation evaluation unit 360 evaluates the distance d in the yz plane between each measurement point q included in the deviation evaluation area (measurement point cloud extraction area) of the measurement point cloud 320 shown in FIG. b Calculate (q). As shown in the above formula, z coordinate q z For a distance measurement point q whose y coordinate is q y and z coordinate q z The square root of the sum of the squares of is d b (q), and the z coordinate is q z For a distance measurement point q where is less than zero and is greater than or equal to α (a negative value), the y-coordinate q y The absolute value of is d b (q), and the z coordinate is q z For a range point q where α is less than α, an arbitrary constant d is large enough to be considered an outlier. b The result is (q).

[0074] 12 is a schematic diagram showing an evaluation of the distance between each of the measurement points in the measurement point cloud 320 and the cargo upper surface sub-model 343. The deviation evaluation unit 360 evaluates the distance d in the yz plane of the cargo upper surface sub-model 343 between each of the measurement points q included in the deviation evaluation area (measurement point cloud extraction area) of the measurement point cloud 320 shown in FIG. c Calculate (q). As shown in the above formula, the y coordinate q y For a range point q with negative y coordinate q y and z coordinate q z The square root of the sum of the squares of the differences between and α is d c (q), and the y coordinate is q y For a range point q whose z coordinate is equal to or greater than zero, z The absolute value of the difference between and α is d c The result is (q).

[0075] As described above, for each of the distance measurement points q in the distance measurement point group 320, in principle, the distance d a (q), d b (q), d c The deviation evaluation unit 360 calculates the three distances d a (q), d b (q), d c The measurement point q is determined to belong to the sub-model 341, 342, 343 that gives the smallest distance among (q). The measurement point q determined to belong to the hold top surface sub-model 341 is equivalent to being determined to be a measurement point on the top surfaces U1 to U4 of the hold 201, the measurement point q determined to belong to the hold side wall surface sub-model 342 is equivalent to being determined to be a measurement point on the side wall surfaces W1 to W4 of the hold 201, and the measurement point q determined to belong to the cargo top surface sub-model 343 is equivalent to being determined to be a measurement point on the top surface of the bulk cargo M stored in the hold 201.

[0076] In the deviation evaluation unit 360, a first evaluation unit 361 using the RANSAC algorithm further evaluates the distance da (q), d b (q), d c The number of measurement points q where (q) is within a predetermined tolerance may be counted. These measurement points q function as inliers that indicate a high degree of agreement between each of the sub-models 341, 342, and 343 and the respective estimation targets (the upper surfaces U1 to U4 of the hold 201, the side wall surfaces W1 to W4 of the hold 201, and the upper surface of the bulk cargo M).

[0077] The second evaluation unit 362 using the EM algorithm in the deviation evaluation unit 360 may secondarily evaluate one or more hold models 340 for which the count value (number of inliers) by the first evaluation unit 361 is equal to or greater than the allowable value. For example, the second evaluation unit 362 may further evaluate or adjust one hold model 340 for which the count value (number of inliers) by the first evaluation unit 361 is the largest. In this way, the typically large number of hold models 340 generated by the hold model generation unit 330 are narrowed down to a typically small number of hold models 340 for which a sufficient number of inliers have been obtained in the primary comparison with the range point cloud 320 in the first evaluation unit 361.

[0078] The second evaluation unit 362 adjusts the deviation of the hold model 340 narrowed down by the first evaluation unit 361 from each ranging point in the ranging point cloud 320 so as to reduce the deviation by a method different from that of the first evaluation unit 361. For example, the second evaluation unit 362 may optimize the deviation or distance between each ranging point in the ranging point cloud 320 and the hold model 340. Note that some or all of the functions of the second evaluation unit 362 and / or the first evaluation unit 361 in the deviation evaluation unit 360 may be realized in a hold model selection unit 370 (estimation unit) described later.

[0079] In the example of this embodiment, the first evaluation unit 361 uses the RANSAC algorithm, whereas the second evaluation unit 362 uses the EM algorithm. However, the algorithms used by each of the evaluation units 361, 362 are not limited to the RANSAC algorithm or the EM algorithm, and may be any algorithm as long as the evaluation accuracy or estimation accuracy in the second evaluation unit 362 is higher than the evaluation accuracy or estimation accuracy in the first evaluation unit 361. In other words, the first evaluation unit 361 primarily narrows down the hold model 340 with a relatively low accuracy, and the second evaluation unit 362 secondarily or finally generates the hold model 340 that represents the actual state of the hold 201 with a relatively high accuracy.

[0080] The EM algorithm in the second evaluation unit 362 aims to minimize the following negative log-likelihood L, using the evaluation result by the first evaluation unit 361 (RANSAC algorithm) as an initial value.

number

[0081] Here, Q represents the measurement point group 320 in the deviation evaluation area (FIG. 9). Also, u represents the probability that the measurement point group Q belongs to each of the sub-models 341, 342, and 343. Specifically, u is the probability u that the measurement point group Q belongs to the hold top surface sub-model 341, as shown in the following formula: a , the probability that the measurement point group Q belongs to the hold side wall sub-model 342, u b , the probability that the measurement point group Q belongs to the cargo top surface sub-model 343, u c is a vector whose elements are

number

[0082] σ cis the standard deviation that represents the variation in the depth α in the z direction of the cargo upper surface sub-model 343 (upper surface of bulk cargo M). Unlike the upper surfaces U1 to U4 and the side wall surfaces W1 to W4 of the hold 201 as a structure, the height of the upper surface of the bulk cargo M (depth α of the cargo upper surface sub-model 343) is expected to vary greatly, and is therefore considered as a main parameter of the EM algorithm. Similarly, the standard deviation σ for the hold upper surface sub-model 341 a , and the standard deviation σ for the hold side wall submodel 342 b may be taken into account in the EM algorithm.

[0083] Also, p a , p b , p c is expressed by the following formula:

number

[0084] The EM algorithm uses the expectation E z may be applied to.

number

number

[0085] In the M (Maximization) step of the EM algorithm, a set of parameters X=(θ hu ,t hu ,α,σ c ,w)=(θ ru ,y,z,α,σ c ,w) is optimized.

number

[0086] 6, the hold model selection unit 370 also functions as an estimation unit that estimates the actual state of the hold 201 based on the deviation between the measurement point cloud 320 and the hold model 340 evaluated by the deviation evaluation unit 360. For example, the hold model selection unit 370 selects or estimates the hold model 340 whose deviation from the measurement point cloud 320 evaluated by the deviation evaluation unit 360 is within an allowable range as representing the actual state of the hold 201. Specifically, the hold model selection unit 370 selects or estimates the hold model 340 whose count value (number of inliers) by the first evaluation unit 361 using the RANSAC algorithm is equal to or greater than an allowable value and which has been adjusted or optimized by the second evaluation unit 362 using the EM algorithm as representing the actual state of the hold 201.

[0087] The unloading section 9 as a discharge section treats the hold model 340 selected or estimated by the hold model selection section 370 as an actual hold 201, and discharges cargo such as bulk cargo M stored in the hold 201.

[0088] According to the embodiment described above, for each of the measurement points in the measurement point cloud 320 acquired by the measurement point cloud acquisition unit 310, the deviation from the hold model 340 generated by the hold model generation unit 330 is evaluated in the deviation evaluation unit 360. Therefore, even if a small number of measurement points are disturbed by dust or steam from cargo such as bulk cargo M stored in the hold 201, the hold model selection unit 370 can select an appropriate hold model 340 that represents the actual state of the hold 201 based on a comparison (deviation evaluation) between the remaining (typically a large number) normal measurement points and the hold model 340.

[0089] 13 is a flowchart showing an example of the estimation process of the hold 201 by the hold estimation device 300. In the flowchart, "S" means a step or process. In S1, the distance measurement point cloud acquisition unit 310 acquires the distance measurement point cloud 320 on the surface of the hold 201 by the distance measurement sensors 18, 19. In S2, the coordinate conversion unit 350 converts, for example, the CSU coordinate q of the distance measurement point cloud 320 acquired in the CSU coordinate system into u,i is expressed by the edge coordinate q in the edge coordinate system in which the hold model 340 is described. r,i Convert to.

[0090] In S3, the initial state of the hold 201, for example, the above-mentioned parameter set X=(θ hu ,t hu ,α,σ c If the detection of the initial state of the hold 201 has not been completed in S3, the process proceeds to the initial state detection process of S4 to S7.

[0091] In S4, the distance measurement point group 320 near the edges E1 to E4 (for example, within the deviation evaluation area in FIG. 9) that are the main targets of the initial state detection process is extracted. In S5, the first evaluation unit 361 uses the RANSAC algorithm to calculate the distance d between each of the distance measurement points in the distance measurement point group 320 and each of the sub-models 341, 342, and 343. a , d b , d c In the primary evaluation by the first evaluation unit 361, the variation (standard deviation σ c ) is a relatively large cargo top surface submodel 343(d c ) is excluded, and the sub-model for the hold 201 itself, specifically the hold top sub-model 341(d a ) and hold side wall submodel 342(d b ) may be the target.

[0092] In S6, the second evaluation unit 362 optimizes the deviation of the hold models 340 narrowed down in S5 from each of the range points in the range point cloud 320 using the EM algorithm. In S7, the hold model selection unit 370 optimizes the hold models 340 optimized by the EM algorithm (S6), specifically the parameter set X=(θ hu ,t hu ,α,σ c , w) as an estimated result representing the actual state of the hold 201.

[0093] If the detection of the initial state of the hold 201 is completed in S3, the process proceeds to tracking processing in S8 to S10. hu ,t hu ,α,σ c , w), w (width in the y direction of the hold 201) related to the shape of the hold 201 is considered to be unchanged, and therefore can be excluded from the estimation target in the tracking process of S8 to S10. Therefore, the parameter group to be estimated in the tracking process of S8 to S10 is, for example, (θ hu ,t hu ,α,σ c ) In addition, in the tracking process of S8 to S10, the estimation results from the initial state detection process of S4 to S7 and the estimation results from the previous tracking process of S8 to S10 can be used, so the primary evaluation (S5) using the RANSAC algorithm can be omitted.

[0094] In S8, the measurement point cloud 320 near edges E1 to E4 (for example, within the deviation evaluation area in FIG. 9) which are the main targets of the tracking process is extracted. In S9, the second evaluation unit 362 uses the EM algorithm to optimize the deviation of each measurement point in the measurement point cloud 320 from the hold model 340 output in the initial state detection process or the previous tracking process. In S10, the hold model selection unit 370 outputs the hold model 340 optimized by the EM algorithm (S9) as an estimated result representing the actual state of the hold 201.

[0095] In S11, it is determined whether or not to end the process. If the determination in S11 is "no", the process returns to S1. If the determination in S11 is "yes", the process ends.

[0096] The present invention has been described above based on the embodiments. Various modifications are possible to the combinations of the components and processes in the exemplary embodiments, and it will be obvious to those skilled in the art that such modifications are included within the scope of the present invention.

[0097] Although an example of a hold model 340 is shown in Fig. 8, the present invention can also be applied to another hold model 340 as shown in Fig. 14, for example. This hold model 340, like Fig. 8, includes a hold upper surface sub-model 341, a hold side wall surface sub-model 342, and a cargo upper surface sub-model 343. Unlike Fig. 8, in this hold model 340, the lower part of the hold side wall surface sub-model 342 (side wall surfaces W1 to W4) with height β protrudes outward in the y direction by a width w'.

[0098] The deviation evaluation unit 360 that evaluates such a hold model 340 may evaluate the distance between each of the sub-models 341, 342, and 343 included in the hold model 340 and each of the measurement points in the measurement point group 320, as in the above-mentioned embodiment. a (q), d b (q), d c (q) is expressed by the following equations, respectively.

number

[0099] d a (q) and d b Since (q) is the same as in the above embodiment, the explanation will be omitted. c (q) is the y-coordinate q as shown in the above formula. yFor the distance measurement point q whose y coordinate is smaller than -w', y and the square of the sum of w' and the z coordinate q z and α, the square root of the sum of the squares of the differences, y For the distance measurement point q whose z coordinate is -w' or more, z and the absolute value of the difference between α.

[0100] The position of the lower edge E' shown in Fig. 14 can be estimated by, for example, calculating the distance between any straight line passing through a group of end points of the distance measurement points 320 that may constitute the lower edge E' and each of the end points. Specifically, for a plurality of straight lines that are candidates for the lower edge E', end points whose distances to the straight lines are less than a threshold are counted as inliers. Then, the straight line with the largest number of inliers is selected as the one that best represents the lower edge E'. Such a process of estimating the lower edge E' is executed by, for example, the first evaluation unit 361 that uses the RANSAC algorithm.

[0101] The present invention is not limited to the bucket elevator type continuous unloader described in the embodiment, but can also be applied to a spiral type unloader or an unloader equipped with an air conveying mechanism.

[0102] The configuration, action, and function of each device and method described in the embodiments can be realized by hardware resources or software resources, or by the cooperation of hardware resources and software resources. For example, a processor, ROM, RAM, and various integrated circuits can be used as hardware resources. For example, programs such as an operating system and applications can be used as software resources. [Explanation of symbols]

[0103] 1 Cargo lifting unit (CSU), 9 cargo lifting section, 11 scraping section, 18 distance measurement sensor, 19 distance measurement sensor, 21 opening, 201 hold, 300 hold estimation device, 310 distance measurement point cloud acquisition section, 320 distance measurement point cloud, 330 hold model generation section, 340 hold model, 341 hold top surface sub-model, 342 hold side wall surface sub-model, 343 cargo top surface sub-model, 350 coordinate conversion section, 360 deviation evaluation section, 361 first evaluation section, 362 second evaluation section, 370 hold model selection section.

Claims

1. a range measurement point cloud acquisition unit that acquires a range measurement point cloud on the surface of the hold by a range measurement sensor; a hold model generation unit that generates a hold model that defines a state of the hold; a deviation evaluation unit that evaluates a deviation between each measurement point in the measurement point cloud and the hold model; an estimation unit that estimates an actual state of the hold based on the deviation; A vessel hold estimation device comprising:

2. The hold estimation device according to claim 1 , wherein the deviation evaluation unit evaluates a distance between each measurement point in the range point cloud and the hold model as the deviation.

3. The deviation evaluation unit counts the number of ranging points among the ranging points whose distance from the hold model is within a tolerance value, The estimation unit estimates the hold model in which the count value is equal to or greater than a tolerance value as representing an actual state of the hold. A vessel hold estimation device according to claim 2.

4. The hold estimation device according to claim 2 , wherein the estimation unit estimates the hold model, adjusted so as to reduce the deviation evaluated by the deviation evaluation unit, as representing the actual state of the hold.

5. the deviation evaluation unit includes a first evaluation unit that counts the distances of the measurement points in the range point cloud whose distances to the hold model are within a tolerance, and a second evaluation unit that adjusts the deviations of the measurement points in the range point cloud from the hold model whose counts are equal to or greater than the tolerance, using a method different from that used by the first evaluation unit; The estimation unit estimates the hold model adjusted by the second evaluation unit as representing an actual state of the hold. A hold estimation device according to claim 3.

6. The hold estimation device according to claim 5 , wherein the first evaluation unit evaluates a deviation between each of the range points in the range point cloud and the hold model using a RANSAC (Random Sample Consensus) algorithm.

7. The hold estimation device according to claim 5 , wherein the second evaluation unit adjusts the deviation between each ranging point in the range point cloud and the hold model using an Expectation Maximization (EM) algorithm.

8. A hold estimation device as described in any one of claims 1 to 7, wherein the hold model includes at least one of a hold top surface sub-model that defines an upper surface facing the edge of the hold, a hold side wall surface sub-model that defines a side wall surface facing the edge of the hold, and a cargo top surface sub-model that defines an upper surface of cargo stored in the hold.

9. 8. A hold estimation device according to claim 1, wherein the state of the hold includes at least one of the position, attitude, and shape of the hold and / or the cargo stored in the hold.

10. acquiring a range point cloud on a surface of the hold by a range sensor; generating a hold model defining a state of the hold; Evaluating a deviation between each of the range points in the range point cloud and the hold model; estimating an actual state of the hold based on the deviation; and A hold estimation method that performs.

11. acquiring a range point cloud on a surface of the hold by a range sensor; generating a hold model defining a state of the hold; Evaluating a deviation between each of the range points in the range point cloud and the hold model; estimating an actual state of the hold based on the deviation; and A ship's hold estimation program that causes a computer to execute the above.

12. a range measurement point cloud acquisition unit that acquires a range measurement point cloud on the surface of the hold by a range measurement sensor; a hold model generation unit that generates a hold model that defines a state of the hold; a deviation evaluation unit that evaluates a deviation between each measurement point in the measurement point cloud and the hold model; an estimation unit that estimates the hold model representing an actual state of the hold based on the deviation; a discharge unit that discharges cargo stored in the hold by using the hold model estimated by the estimation unit; An unloading device comprising:

Images