Underwater cleaning device

Abstract

To provide an underwater cleaning device capable of efficiently cleaning.SOLUTION: An underwater cleaning device comprises moving means to move underwater, cleaning means to clean underwater cleaning targets, one more acquisition means to acquire the image of the cleaning target or the surface friction coefficient of the cleaning target, and moving control means to control the moving means based on an image or the friction coefficient.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an underwater cleaning device. [Background technology]

[0002] As a document disclosing the background art, there is JP2021-521057A (Patent Document 1). This document states that "In FIG. 1C, a method for cleaning the hull of a vessel 140 is provided. In 142, an underwater hull cleaning machine (e.g., 100) can be deployed on the surface of the hull. In 144, a real-time image of a portion of the surface can be captured (e.g., by a camera or CCTV) to determine whether the portion of the surface should be cleaned. In 146, if it is determined that the portion of the surface should be cleaned, at least one brush (e.g., 108) of the underwater hull cleaning machine 100 can be lowered to an extended position. In the extended position, the at least one brush 108 can extend downward toward a base region (e.g., 104) of the underwater hull cleaning machine 100 and can be in contact with a portion of the surface" (see paragraph 0060).

[0003] Another document is JP 2015-157587 A (Patent Document 2), which states that "transparent front cover 31 and rear cover 33 are provided at the front end (front end in the traveling direction) and rear end (rear end in the traveling direction) of the housing body 3, respectively, a front camera 35 is accommodated in the front cover 31, and a rear camera 37 is accommodated in the rear cover 33, and the movement or running of the submersible cleaning machine 1 in water and on the bottom and side of the ship is controlled by the controller 27 while checking the front and rear images sent from the front camera 35 and the rear camera 37 and displayed on the control and monitor computer 29" (see paragraph 0017).

[0004] Yet another document is JP 2011-088485 A (Patent Document 3), which states, "When such a pair of thrusters / brushes is not grounded simultaneously, but one of them is grounded on the bottom or side of the ship, a pitch motion will occur due to the reaction moment, but the pitch motion (motion around the z-axis) is controlled by the differential or combination of the multiple side thrusters and the multiple thrusters at the rear, as described above. Such control is performed by remote control using cameras mounted on the front and rear of the device" (see paragraph 0017). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Special Publication No. 2021-521057 [Patent Document 2] JP 2015-157587 A [Patent Document 3] JP 2011-088485 A Summary of the Invention [Problem to be solved by the invention]

[0006] The above Patent Document 1 describes lowering a brush toward a ship's hull surface based on an image. The above Patent Document 2 describes a person controlling the movement of an underwater cleaning machine based on a camera image. The above Patent Document 3 describes a person controlling a device (underwater robot) based on a camera image. However, neither of the patent documents discloses automatic control of the movement of an underwater cleaning device based on an image. Furthermore, none of the patent documents discloses automatic control of the movement of an underwater cleaning device based on the friction coefficient of the surface to be cleaned. The present invention has been made in consideration of the above circumstances, and provides a submersible cleaning device capable of efficient cleaning. [Means for solving the problem]

[0007] In order to solve the above problems, for example, the configurations described in the claims are adopted. The present application includes multiple means for solving the above problems, and one example is an underwater cleaning device comprising: a moving means for moving underwater, a cleaning means for cleaning a cleaning target underwater, one or more acquisition means for acquiring an image of the cleaning target or a friction coefficient of the surface of the cleaning target, and a movement control means for controlling the moving means based on the image or the friction coefficient. Effect of the Invention

[0008] According to the present invention, a submersible cleaning device capable of efficient cleaning can be provided. Problems, configurations and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a top perspective view showing an example of the appearance of an ROV 100. [Diagram 2] FIG. 2 is a perspective view showing an example of the appearance of the ROV 100, as viewed from below. [Diagram 3] FIG. 3 is a plan view showing an example of the appearance of the ROV 100. [Figure 4] FIG. 4 is a front view showing an example of the appearance of the ROV 100. [Diagram 5] FIG. 5 is a side view showing an example of the appearance of the ROV 100. [Figure 6] FIG. 6 is a rear view showing an example of the appearance of the ROV 100. [Figure 7] FIG. 7 is a bottom view showing an example of the appearance of the ROV 100. [Figure 8] FIG. 8 is a perspective view seen from above, showing an example of the appearance of the frame 101. [Figure 9] FIG. 9 is a perspective view showing an example of the appearance of the frame 101, as viewed from below. [Figure 10] FIG. 10 is a plan view showing an example of the appearance of the frame 101. As shown in FIG. [Figure 11] FIG. 11 is a side view showing an example of the appearance of the frame 101. As shown in FIG. [Figure 12] FIG. 12 is a bottom view showing an example of the appearance of the frame 101. As shown in FIG. [Figure 13] FIG. 13 is a diagram showing an example of a cross section of the camera unit 103. As shown in FIG. [Figure 14] FIG. 14 is a diagram showing an example of a vertical cross section of the camera unit 103. As shown in FIG. [Figure 15] FIG. 15 shows an example of the electrical configuration of the ROV 100. [Figure 16] FIG. 16 shows an example of how the ROV 100 can be operated. [Figure 17] FIG. 17 shows an example of how the ROV 100 can be operated. [Figure 18] FIG. 18 shows an example of how the ROV 100 is operated. [Figure 19] FIG. 19 shows an example of how the ROV 100 is operated. [Figure 20] FIG. 20 shows an example of how the ROV 100 is operated. [Figure 21] FIG. 21 shows an example of how the ROV 100 can be operated. [Figure 22] FIG. 22 shows an example of how the ROV 100 can be operated. [Figure 23] FIG. 23 shows an example of how the ROV 100 is operated. [Figure 24] FIG. 24 shows an example of a control flow during cleaning. [Diagram 25] FIG. 25 shows an example of a defect detection process flow. [Figure 26] FIG. 26 is a top perspective view showing an example of the appearance of ROV100A. [Figure 27] FIG. 27 is a perspective view showing an example of the appearance of the ROV 100A, as viewed from below. [Figure 28] FIG. 28 is a plan view showing an example of the appearance of the ROV 100A. [Figure 29] FIG. 29 is a front view showing an example of the appearance of the ROV 100A. [Diagram 30]FIG. 30 is a side view showing an example of the appearance of ROV100A. [Diagram 31] FIG. 31 is a rear view showing an example of the appearance of ROV100A. [Diagram 32] FIG. 32 is a bottom view showing an example of the appearance of ROV100A. [Diagram 33] FIG. 33 is a perspective view seen from above, showing an example of the appearance of the frame 101A. [Diagram 34] FIG. 34 is a perspective view showing an example of the appearance of the frame 101A, as viewed from below. [Diagram 35] FIG. 35 is a plan view showing an example of the appearance of the frame 101A. [Diagram 36] FIG. 36 is a side view showing an example of the appearance of the frame 101A. [Figure 37] FIG. 37 is a bottom view showing an example of the appearance of the frame 101A. [Figure 38] FIG. 38 shows an example of a cross section of the camera unit 2601. [Figure 39] FIG. 39 shows an example of a vertical cross section of the camera unit 2601. [Diagram 40] FIG. 40 is a diagram comparing the thickness of the camera unit 2601 and the camera unit 103. [Diagram 41] FIG. 41 is a diagram comparing the thickness of ROV100A and ROV100. [Diagram 42] FIG. 42 shows an example of the electrical configuration of the ROV100A. [Diagram 43] FIG. 43 is a top perspective view showing an example of the appearance of ROV 100B. [Diagram 44] FIG. 44 is a perspective view showing an example of the appearance of ROV100B, seen from below. [Diagram 45] FIG. 45 is a plan view showing an example of the appearance of the ROV 100B. [Figure 46] FIG. 46 shows an example of the electrical configuration of ROV100B. [Figure 47] FIG. 47 shows an example of a control flow during cleaning. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, the embodiments will be described with reference to the drawings. 1. First Example 1-1. Overview Traditionally, inspection and cleaning of the bottom of a ship is mainly done manually, either when the ship is docked every 2-5 years, or when a visual inspection by a diver indicates that cleaning of marine mullingos (in other words, marine biological fouling) is necessary. The proliferation of mullingos reduces a ship's fuel efficiency by 5-15%. In addition, they can bring organisms from other regions attached to the bottom of the ship into port.

[0011] This embodiment has been developed in view of the above circumstances, and relates to an ROV (Remotely Operated Vehicle) for cleaning ship hulls. This ROV is equipped with a front camera and a rear camera arranged to sandwich a rotating brush along the traveling direction. The ROV determines the degree of dirt based on the image captured by the front camera, and changes the rotation speed of the rotating brush according to the determined degree of dirt. The ROV also detects areas that have not been cleaned based on the image captured by the rear camera, and if an area that has not been cleaned is detected, it retreats and scrubs again with the rotating brush.

[0012] First, this ROV can remove marine debris. As a result, it can prevent the ship's fuel consumption from decreasing and prevent organisms from other regions that are attached to the bottom of the ship from being brought into the port. Second, if any areas are missed after cleaning, they can be cleaned repeatedly to prevent this. In addition, redundant cleaning can be avoided to protect the painted surface of the hull. Third, it can record images before and after cleaning using two cameras.

[0013] 1-2.Configuration The configuration of the ROV 100 will be described. The ROV 100 is a wired and remotely operated unmanned underwater vehicle (in other words, an unmanned exploration vehicle, an underwater robot, or an underwater drone). The ROV 100 is an underwater cleaning device equipped with a rotating brush for cleaning the hull.

[0014] 1 to 7 show examples of the appearance of the ROV 100. Fig. 1 is a perspective view seen from above, Fig. 2 is a perspective view seen from below, Fig. 3 is a plan view, Fig. 4 is a front view, Fig. 5 is a side view, Fig. 6 is a rear view, and Fig. 7 is a bottom view.

[0015] The ROV 100 mainly comprises a frame 101, ten thrusters 102A to 102J, a front camera unit 103A, a rear camera unit 103B, three rotating brushes 104, a control unit 105, two drive wheels 106, and a caster 107. Each component will be described below.

[0016] First, frame 101 will be described. Figures 8 to 12 show examples of the appearance of frame 101. Figure 8 is a perspective view seen from above, Figure 9 is a perspective view seen from below, Figure 10 is a plan view, Figure 11 is a side view, and Figure 12 is a bottom view.

[0017] The frame 101 has a substantially rectangular frame body 801, a front beam portion 802 and a rear beam portion 803 that are spanned between the left and right frame edges of the frame body 801, a pair of L-shaped wheel mounting portions 804 extending from the underside of the front beam portion 802, a connecting plate 805 that connects the central portions of the front beam portion 802 and the rear beam portion 803, a front column portion 806 that is erected on the front of the upper surface of the connecting plate 805, a rear column portion 808 that is inserted into a cutout portion 807 formed at the rear of the connecting plate 805 and fixed to the front surface of the rear beam portion 803, a support plate 809 that is spanned between the front column portion 806 and the rear column portion 808 and is curved in an arc shape, and a hollow brush mounting portion 810 that is attached to the center of the underside of the connecting plate 805 and extends in the left-right direction. The connection between the brush attachment portion 810 and the connecting plate 805 may have elastic mobility in the pressing direction of the rotating brush 104 .

[0018] Three cylindrical openings 811 are formed in the bottom surface of the last brush attachment part 810 (see FIG. 12). The three cylindrical openings 811 are formed at the left and right ends and in the center of the brush attachment part 810. In addition, two cylindrical openings 812 are formed in the top surface of the brush attachment part 810 (see FIG. 10). The two cylindrical openings 812 are formed at positions that do not overlap with the three cylindrical openings 811 in a plan view.

[0019] Next, the ten thrusters 102A to 102J will be described with reference again to FIGS. Of the ten thrusters 102A-102J, eight thrusters 102A-102H are movement means for moving the ROV 100 in water. Of these, the thrusters 102A-102D are arranged at approximately equal intervals on the left frame side that constitutes the frame 801 (see FIG. 3). The thrusters 102E-102H are arranged at approximately equal intervals on the right frame side that constitutes the frame 801 (see FIG. 3). The thrusters 102B, 102C, 102F, and 102G may be arranged on the brush attachment portion 810.

[0020] The remaining thrusters 102I and 102J are means for discharging dirt removed by the rotating brush 104 above the ROV 100. These thrusters 102I and 102J are fitted into openings 812 (see FIG. 10) of the frame 101, respectively.

[0021] Next, the front camera unit 103A and the rear camera unit 103B will be described. The front camera unit 103A and the rear camera unit 103B have a substantially quadrangular pyramid shape. Of the two camera units, the front camera unit 103A is attached to the frame 101 such that its upper portion is sandwiched between the front frame edge of the frame body 801 and the front beam portion 802 (see FIG. 1). On the other hand, the rear camera unit 103B is attached to the frame 101 such that its upper portion is sandwiched between the rear frame edge of the frame body 801 and the rear beam portion 803 (see FIG. 1). In the following description, unless there is a particular need to distinguish between the front camera unit 103A and the rear camera unit 103B, they will be collectively referred to as the "camera unit 103."

[0022] Fig. 13 is a cross-sectional view of camera unit 103. Fig. 14 is a vertical cross-sectional view of camera unit 103. The camera unit 103 roughly comprises a water box 1301, a glass plate 1302, a camera cover 1303, an image sensor board 1307, a measurement computer 1309, and two rows of illumination lights 1310. Each of the components will be described below.

[0023] The water box 1301 has a hollow shape of a roughly truncated quadrangular pyramid, with an open top and a bottom sealed with a rectangular glass plate 1302 . Camera cover 1303 has a hollow, roughly rectangular parallelepiped shape, and the edge of its lower opening is connected to the edge of the upper opening of water box 1301. This camera cover 1303 has a rectangular partition plate 1304 that divides the inside of the cover into upper and lower sections, a circular opening 1305 formed in the center of partition plate 1304, and a protective dome glass 1306 attached to the underside of partition plate 1304 so as to cover opening 1305.

[0024] Image sensor board 1307 is housed in the space above camera cover 1303 formed by partition plate 1304. The board is positioned so that lens 1308 looks downward through opening 1305. The measurement computer 1309 is housed in the space above the camera cover 1303 formed by the partition plate 1304 . Two rows of lamps 1310 are positioned along the leading and trailing edges of the glass pane 1302 .

[0025] The space between the glass plate 1302 and the protective dome glass 1306 of this camera unit 103 is filled with a highly transparent liquid (e.g., water). Therefore, the image sensor board 1307 can capture an image of the hull surface through the highly transparent liquid. Note that the image sensor board 1307 referred to here is a means for acquiring an image of the hull surface.

[0026] Next, the three rotating brushes 104 will be described. Each of the three rotating brushes 104 has a disk portion 701 and a spiral brush portion 702 extending downward from the bottom surface of the disk portion 701 (see FIG. 7).

[0027] Each rotating brush 104 is rotatably attached to an opening 811 of the frame 101. Each rotating brush 104 is driven by a brush motor (not shown).

[0028] Each rotating brush 104 attached to the frame 101 has a circular opening 703 (see FIG. 7) formed in the center of the disk portion 701 thereof, which communicates with an opening 811 (see FIG. 12) of the frame 101.

[0029] The three rotating brushes 104 described above are cleaning means for cleaning the hull surface. These three rotating brushes 104 and the above-mentioned front camera unit 103A and rear camera unit 103B are arranged in a line along the traveling direction of the ROV 100 (see FIG. 7). In addition, the three rotating brushes 104 are arranged so as to be sandwiched between the front camera unit 103A and the rear camera unit 103B along the traveling direction of the ROV 100.

[0030] The cleaning ranges of the three rotating brushes 104 (more specifically, the cleaning ranges extending in a direction substantially perpendicular to the traveling direction) correspond to the image acquisition ranges of the front camera unit 103A and the rear camera unit 103B (more specifically, the image acquisition ranges extending in a direction substantially perpendicular to the traveling direction) (see FIG. 7). In other words, the cleaning ranges and the image acquisition ranges appear to overlap when viewed from the front.

[0031] Next, the control unit 105 will be described. The control unit 105 has a substantially cylindrical case, and houses various devices such as a control computer and a position measuring device inside this case. The control unit 105 is mounted on a mounting plate 809 of the frame 101 (see FIG. 5).

[0032] Next, the two drive wheels 106 will be described. Each drive wheel 106 is rotatably attached to a wheel mounting portion 804 of the frame 101 (see FIG. 7), and is driven by a wheel motor 704. The drive wheels 106 may also have a suspension mechanism. Each drive wheel 106 is a means of movement for the ROV 100 to move underwater. The drive wheels 106 allow the ROV 100 to move linearly on the surface of the hull, performing reciprocating motion. The direction of movement can be changed by providing a rotation difference between the two drive wheels.

[0033] Finally, the caster 107 will be described. The caster 107 is attached to the rear end of the rear pillar 808 of the frame 101 (see FIG. 2). The caster 107 is a driven wheel. The caster 107 may have a suspension mechanism or a rotation angle detection function.

[0034] Next, a description will be given of the electrical configuration of the ROV 100. FIG.

[0035] The ROV 100 generally includes a control computer 1501, a position measuring device 1502, a wheel motor 704, thrusters 102A to 102J, a brush motor 1503, measurement computers 1309A and 1309B, cameras 1307A and 1307B, and lighting fixtures 1310A and 1310B. Each component will be described below.

[0036] The control computer 1501 controls the wheel motor 704 using a driver (or device driver) 1504. The control computer 1501 also controls the thrusters 102A to 102J using a driver 1505. The control computer 1501 also controls the brush motor 1503 using a driver 1506. The control computer 1501 also acquires camera images from the measurement computers 1309A and 1309B.

[0037] The position measuring device 1502 is a measuring device for measuring the position of the ROV 100, and includes a depth sensor and an inclination sensor. The position measuring device 1502 outputs the measurement value to the control computer 1501.

[0038] The measurement calculator 1309A is the measurement calculator 1309 included in the front camera unit 103A. The camera 1307A is the image sensor board 1307 included in the front camera unit 103A. The illuminating lamp 1310A is the illuminating lamp 1310 included in the front camera unit 103A. The measurement calculator 1309A acquires a camera image from the camera 1307A. The measurement calculator 1309A also uses a driver 1507 to control the illuminating lamp 1310A.

[0039] The measurement computer 1309B is the measurement computer 1309 included in the rear camera unit 103B. The camera 1307B is the image sensor board 1307 included in the rear camera unit 103B. The illuminating lamp 1310B is the illuminating lamp 1310 included in the rear camera unit 103B. The measurement computer 1309B acquires a camera image from the camera 1307B. The measurement computer 1309B also uses a driver 1508 to control the illuminating lamp 1310B.

[0040] The control computer 1501 of the ROV 100 is connected to a console 1510 via a communication cable 1520. The communication cable 1520 may supply power or may add a mark for measuring the amount of movement. The console 1510 is a device placed on the ship to be cleaned, and includes a PC 1511, a display 1512, and a controller 1513. An operator uses this console 1510 to operate and monitor the ROV 100.

[0041] Next, the functional configuration of the ROV100 will be described. The above-mentioned control computer 1501 includes a main memory device, an auxiliary memory device, and a processor. Among these, the main memory device stores programs such as an image correction module 1531, an image analysis module 1532, a movement control module 1533, and a cleaning control module 1534 (all not shown). The various functions of the ROV 100 are realized by the processor executing these programs.

[0042] Each module may be implemented in hardware by integration, etc. Each module may be an independent program or application, or may be implemented as a subprogram or function in one integrated program or application. The image correction module 1531 and the image analysis module 1532 may be provided on the measurement computer.

[0043] In this specification, each module is described as an entity (subject) that performs processing, but in reality, a processor that processes various programs, applications, etc. (modules) executes the processing. Each module will be described below.

[0044] The image correction module 1531 corrects the camera image acquired from the measurement computer 1309A or 1309B. Specifically, this module performs lens distortion correction, perspective distortion correction, and contrast correction. Among these, in the perspective distortion correction, since the optical axis is tilted about 15 degrees with respect to the measurement surface, the image is converted to a front image by projective transformation.

[0045] In contrast correction, an image and turbidity are input to a trained model, which outputs a corrected image. At this time, the trained model to which the image, etc. is input is generated in advance by having a machine learning model (e.g., a neural network) learn training data. The training data is, for example, a collection of data sets each consisting of an image degraded by murky water, an image before degradation, and turbidity.

[0046] The turbidity input into the trained model may be measured by a turbidity sensor, or may be determined by analyzing the image to be corrected.

[0047] The image analysis module 1532 analyzes the corrected image generated by the image correction module 1531 to detect defects. The defects referred to here are marine biological fouling (in other words, marine loss). Specifically, the module inputs the corrected image into a trained model and obtains an output value indicating the amount of defects. If the obtained output value exceeds a predetermined threshold, the module determines that there is a defect (in other words, cleaning is incomplete), and if the obtained output value does not exceed the predetermined threshold, the module determines that there is no defect (in other words, cleaning is complete).

[0048] The trained model to which the corrected image is input is generated in advance by having a machine learning model (e.g., a neural network) learn training data. The training data is, for example, a collection of data sets each of which is made up of an image of the hull surface and a value indicating the amount of defect.

[0049] The movement control module 1533 controls the wheel motor 704 based on the camera image (i.e., the camera image of the rear camera unit 103B) acquired from the measurement computer 1309B. Specifically, the module moves the ROV 100 forward when the rear camera image is analyzed by the image analysis module 1532 and it is determined that cleaning is complete. On the other hand, the module moves the ROV 100 backward when the image analysis module 1532 determines that cleaning is not complete. This allows the target area to be cleaned again.

[0050] The module also determines that cleaning is complete when the number of cleaning attempts exceeds a certain threshold, and moves the ROV100 forward to avoid redundant cleaning to protect the hull's painted surface.

[0051] The cleaning control module 1534 controls the brush motor 1503 based on the camera image (i.e., the camera image of the front camera unit 103A) acquired from the measurement computer 1309A. Specifically, the module controls the rotation speed of the rotating brush 104 according to the defect amount value outputted after the front camera image is analyzed by the image analysis module 1532. At that time, the module increases the rotation speed of the rotating brush 104 as the defect amount value increases. Alternatively, the module increases the thrust of the thrusters 102B, 102C, 102F, and 102G to increase the pressing force of the rotating brush 104. Alternatively, the moving speed is decreased. This makes it possible to realize cleaning according to the defect amount.

[0052] 1-3. Operation method Next, a description will be given of a method of operating the ROV 100. Fig. 16 to Fig. 23 show an example of a method of operating the ROV 100.

[0053] First, as shown in Fig. 16, an operator lowers the ROV 100 to the water surface from near the bow using a small crane 1601. Note that reference numeral 1602 in the figure denotes a tether winding machine for winding up the communication cable 1520.

[0054] Next, as shown in FIG. 17, the operator controls the ROV 100 to move from near the bow to the cleaning start position.

[0055] 18, an operator lowers the relay roller 1801 to near the water surface and bends the communication cable 1520. Note that reference numeral 1802 in the figure indicates a roller suspension cable.

[0056] Next, as shown in FIG. 19, the operator secures the relay roller 1801 with a roller restraining rope 1901, avoiding the drawer line.

[0057] Next, the operator sends a command to start cleaning to the ROV 100. Then, as shown in FIG. 20, the ROV 100 automatically moves up and down and moves to the stern while extending the communication cable 1520.

[0058] During cleaning, the ROV100 moves vertically up and down with an accuracy of ±0.1 degrees using the tilt sensor within a depth range specified by the depth sensor. The movement path of the ROV100 may overlap. If dirt remains, the ROV100 will perform cleaning again. The communication cable 1520 connected to the ROV100 automatically maintains tension that does not strongly constrain the ROV100. No human supervision is required while the ROV100 is cleaning.

[0059] When the ROV 100 reaches the stern, it automatically returns to the bow, as shown in FIG.

[0060] When the ROV 100 returns to the bow, the operator releases the roller restraining rope 1901 and retrieves the relay roller 1801, as shown in FIG. 22 .

[0061] Finally, as shown in FIG. 23, an operator lowers a sling along the communication cable 1520 and retrieves the ROV 100 with a small crane 1601. This concludes the explanation of the operation method.

[0062] 1-4.Operation Next, we will explain the operation of the ROV 100. Specifically, we will explain the control flow during cleaning and the defect detection process flow. First, the control flow during cleaning will be described.

[0063] FIG. 24 is an activity diagram showing an example of a control flow during cleaning. In the flow 2400 shown in the figure, the movement control module 1533 of the ROV 100 first resets the resharpening counter (step 2401). Next, the module controls the wheel motor 704 to move the ROV 100 forward (step 2402). Next, steps 2403 to 2405 are executed in parallel.

[0064] First, in step 2403, the image analysis module 1532 analyzes the camera image (i.e., the camera image of the front camera unit 103A) acquired from the measurement computer 1309A to detect defects. Note that the defects referred to here are marine biological fouling (in other words, marine loss). Incidentally, the camera image analyzed here is a camera image corrected by the image correction module 1531 .

[0065] In step 2404, the cleaning control module 1534 controls the rotation speed of the rotating brush 104 in accordance with the amount of defects detected in step 2403. At that time, the cleaning control module 1534 increases the cleaning ability by increasing the rotation speed of the rotating brush 104 as the value of the amount of defects increases.

[0066] In step 2405, the image analysis module 1532 analyzes the camera image acquired from the measurement computer 1309B (i.e., the camera image of the rear camera unit 103B) to detect defects. Incidentally, the camera image analyzed here is the camera image corrected by the image correction module 1531.

[0067] If the analysis in step 2405 determines that there is no defect ("no marine life remaining"), the movement control module 1533 returns to step 2402 and controls the wheel motor 704 to move the ROV 100 forward. On the other hand, if the analysis in step 2405 determines that there is a defect ("marine life remaining"), the movement control module 1533 increments the re-scrubbing counter (step 2406). Then, if the incremented value is equal to or less than the upper limit value ("upper limit or less"), the movement control module 1533 controls the wheel motor 704 to move the ROV 100 backward (step 2407). This allows the target area to be cleaned again. After the ROV 100 retreats, steps 2403 to 2405 are executed in parallel again.

[0068] On the other hand, if the incremented value exceeds the upper limit value ("Exceeds Upper Limit Value"), the module returns to step 2401 and resets the resharpening counter. Then, the module controls the wheel motors 704 to move the ROV 100 forward (step 2402). The above is the explanation of the control flow 2400.

[0069] Next, the defect detection process flow will be described. FIG. 25 is an activity diagram showing an example of a defect detection processing flow. In a flow 2500 shown in the figure, the image correction module 1531 of the ROV 100 corrects the camera image acquired from the measurement computer 1309A or 1309B (step 2501). Specifically, the module performs lens distortion correction, perspective distortion correction, and contrast correction.

[0070] Next, the image analysis module 1532 analyzes the corrected image generated by the image correction module 1531 to detect defects (step 2502). The defects referred to here are marine biological fouling (in other words, marine loss). Specifically, the module inputs the corrected image into a trained model and obtains an output value indicating the amount of defects. If the obtained output value exceeds a predetermined threshold, the module determines that there is a defect, and if the output value does not exceed the predetermined threshold, the module determines that there is no defect. The module outputs the determination result as defect information to the movement control module 1533 and the cleaning control module 1534. The defect detection process flow 2500 has been described above.

[0071] According to the ROV100 described above, firstly, marine loss on the hull surface can be removed. As a result, the decrease in fuel efficiency of the ship can be suppressed, and the bringing of organisms from other regions attached to the bottom of the ship into the port can be suppressed. Secondly, if there are any areas that have been missed after cleaning, the missed areas can be suppressed by cleaning repeatedly. In addition, redundant cleaning can be avoided to protect the painted surface of the hull. Thirdly, images before and after cleaning can be recorded using two cameras.

[0072] 2. Second Example The ROV 100A according to this embodiment is thinner than the ROV 100 according to the first embodiment. Hereinafter, the thinner ROV 100A will be described with reference to the drawings.

[0073] 2-1.Configuration Figures 26 to 32 show examples of the appearance of the ROV100A. Figure 26 is a perspective view seen from above, Figure 27 is a perspective view seen from below, Figure 28 is a plan view, Figure 29 is a front view, Figure 30 is a side view, Figure 31 is a rear view, and Figure 32 is a bottom view. In these figures, the same components as those in the first embodiment are denoted by the same reference numerals.

[0074] The ROV 100A mainly comprises a frame 101A, ten thrusters 102A to 102J, a front camera unit 2601A, a rear camera unit 2601B, three rotating brushes 104, a control unit 105, two drive wheels 106, and a caster 107. Each component will be described below.

[0075] First, frame 101A will be described. Figures 33 to 37 show examples of the appearance of frame 101A. Figure 33 is a perspective view seen from above, Figure 34 is a perspective view seen from below, Figure 35 is a plan view, Figure 36 is a side view, and Figure 37 is a bottom view. In these figures, the same components as in the first embodiment are given the same reference numerals.

[0076] Frame 101A has a substantially rectangular frame body 801, a front beam section 802 and a rear beam section 803 bridged over the left and right frame sides of frame body 801, a pair of rectangular parallelepiped wheel attachment sections 3301 extending from the lower surface of front beam section 802, a connecting plate 3302 connecting the centers of front beam section 802 and rear beam section 803, a mounting plate 809 curved in an arc attached to the upper surface of connecting plate 3302, and a hollow brush attachment section 810 extending in the left-right direction attached to the center of the lower surface of connecting plate 3302. The connection between brush attachment section 810 and connecting plate 805 may have elastic mobility in the pressing direction of rotating brush 104.

[0077] Three cylindrical openings 811 are formed in the bottom surface of the last brush attachment part 810 (see FIG. 37). The three cylindrical openings 811 are formed at the left and right ends and in the center of the brush attachment part 810. In addition, two cylindrical openings 812 are formed in the top surface of the brush attachment part 810 (see FIG. 35). The two cylindrical openings 812 are formed at positions that do not overlap with the three cylindrical openings 811 in a plan view.

[0078] Next, the ten thrusters 102A to 102J will be described with reference again to FIGS. Of the ten thrusters 102A-102J, eight thrusters 102A-102H are movement means for moving the ROV 100A in water. Of these, the thrusters 102A-102D are arranged at approximately equal intervals on the left frame side that constitutes the frame 801 (see FIG. 28). The thrusters 102E-102H are arranged at approximately equal intervals on the right frame side that constitutes the frame 801 (see FIG. 28). The thrusters 102B, 102C, 102F, and 102G may be arranged on the brush attachment portion 810.

[0079] The remaining thrusters 102I and 102J are means for discharging dirt removed by the rotating brush 104 above the ROV 100A. These thrusters 102I and 102J are fitted into openings 812 (see FIG. 35) of the frame 101A, respectively.

[0080] Next, the front camera unit 2601A and the rear camera unit 2601B will be described. The front camera unit 2601A and the rear camera unit 2601B have a substantially rectangular parallelepiped shape. Of the two camera units, the front camera unit 2601A is attached to the frame 101A so that its upper portion is sandwiched between the front frame side of the frame body 801 and the front beam portion 802 (see FIG. 28). On the other hand, the rear camera unit 2601B is attached to the frame 101A so that its upper portion is sandwiched between the rear frame side of the frame body 801 and the rear beam portion 803 (see FIG. 28). In the following description, unless there is a particular need to distinguish between the front camera unit 2601A and the rear camera unit 2601B, they will be collectively referred to as the "camera unit 2601."

[0081] Fig. 38 is a horizontal cross-sectional view of camera unit 2601. Fig. 39 is a vertical cross-sectional view of camera unit 2601. In these figures, the same components as those in the first embodiment are denoted by the same reference numerals.

[0082] The camera unit 2601 roughly comprises a water box 3801, a glass plate 1302, a camera cover 3803, two image sensor boards 1307, a measurement computer 1309, and two rows of illumination lamps 1310. Each of the components will be described below.

[0083] Water box 3801 is formed by arranging two small water boxes 3804 in the left-right direction. Each small water box 3804 has a hollow shape of a roughly square truncated pyramid, an open top, and a bottom sealed with a rectangular glass plate 1302.

[0084] The camera cover 3803 has two rectangular parallelepiped camera housing sections 3805 and a rectangular parallelepiped computer housing section 3806 sandwiched between the two camera housing sections 3805 .

[0085] The lower opening edge of each of the two camera housings 3805 is connected to the upper opening edge of the small water box 3804. Each camera housing 3805 has a partition plate 1304 that separates it into upper and lower sections. A circular opening 1305 is formed in the center of the partition plate 1304. A protective dome glass 1306 is attached to the underside of the partition plate 1304 so as to cover the opening 1305.

[0086] In the upper space of camera housing portion 3805 formed by partition plate 1304, image sensor board 1307 is housed so that lens 1308 looks downward through opening 1305.

[0087] The computer housing section 3806 houses the measurement computer 1309 . Two rows of lamps 1310 are arranged along the leading and trailing edges of the glass plate 1302, respectively.

[0088] The space between the glass plate 1302 and the protective dome glass 1306 of this camera unit 2601 is filled with a highly transparent liquid (e.g., water). Therefore, the image sensor board 1307 can capture an image of the hull surface through the highly transparent liquid. The image sensor board 1307 referred to here is a means for acquiring an image of the hull surface.

[0089] Next, the three rotating brushes 104 will be described. Each of the three rotating brushes 104 has a disk portion 701 and a spiral brush portion 702 extending downward from the bottom surface of the disk portion 701 (see FIG. 32).

[0090] Each of the rotating brushes 104 is rotatably attached to an opening 811 of the frame 101A. Each of the rotating brushes 104 is driven by a brush motor (not shown).

[0091] Each rotating brush 104 attached to the frame 101A has a circular opening 703 (see FIG. 32) formed in the center of the disk portion 701 thereof, which communicates with an opening 811 (see FIG. 37) of the frame 101A.

[0092] The three rotating brushes 104 described above are cleaning means for cleaning the hull surface. These three rotating brushes 104 and the above-mentioned front camera unit 2601A and rear camera unit 2601B are arranged in a line along the traveling direction of the ROV 100A (see FIG. 32). In addition, the three rotating brushes 104 are arranged so as to be sandwiched between the front camera unit 2601A and the rear camera unit 2601B along the traveling direction of the ROV 100A.

[0093] The cleaning ranges of the three rotating brushes 104 (more specifically, the cleaning ranges extending in a direction substantially perpendicular to the traveling direction) correspond to the image acquisition ranges of the front camera unit 2601A and the rear camera unit 2601B (more specifically, the image acquisition ranges extending in a direction substantially perpendicular to the traveling direction) (see FIG. 32). In other words, the cleaning ranges and the image acquisition ranges appear to overlap when viewed from the front.

[0094] Next, the control unit 105 will be described. The control unit 105 has a substantially cylindrical case, and houses various devices such as a control computer and a position measuring device inside this case. The control unit 105 is attached to a mounting plate 809 of the frame 101A (see FIG. 29).

[0095] Next, the two drive wheels 106 will be described. Each drive wheel 106 is rotatably attached to a wheel mounting portion 804 of the frame 101 (see FIG. 32), and is driven by a wheel motor 704. The drive wheels 106 may also have a suspension mechanism. Each drive wheel 106 is a means of movement for the ROV 100A to move underwater. The drive wheels 106 allow the ROV 100A to move linearly on the surface of the hull, performing reciprocating motion. The direction of movement can be changed by providing a rotation difference between the two drive wheels.

[0096] Finally, the caster 107 will be described. The caster 107 is attached to the rear end of the lower surface of the connecting plate 3302 of the frame 101A (see FIG. 27). The caster 107 is a driven wheel. The caster 107 may have a suspension mechanism or a rotation angle detection function.

[0097] 40 is a diagram comparing the thickness of camera unit 2601 and camera unit 103 according to the first embodiment. As shown in the figure, camera unit 2601 is thinner than camera unit 103. This is because the number of image sensor boards 1307 has been increased from one to two. As a result, the volume of the water box has been reduced, and the weight on the water has also been reduced.

[0098] Fig. 41 is a diagram comparing the thickness of the ROV100A and the ROV100 according to the first embodiment. As shown in the figure, the ROV100A is thinner than the ROV100. This is because the reduced thickness of the camera unit 2601 allows the control unit 105 to be positioned lower. By making it thinner, the influence of the current can be reduced and operation can be performed even if there is a small gap between the hull and the quay.

[0099] Next, the electrical configuration of the ROV 100A will be described. Figure 42 shows an example of the electrical configuration of the ROV 100A. In this figure, the same components as those in the first embodiment are given the same reference numerals.

[0100] Compared to the ROV100, the ROV100A further includes cameras 1307C and 1307D. The camera 1307C is an image sensor board 1307 included in the front camera unit 2601A. The measurement computer 1309A acquires a camera image from this camera 1307C. On the other hand, the camera 1307D is an image sensor board 1307 included in the rear camera unit 2601B. The measurement computer 1309B acquires a camera image from this camera 1307D. ROV100A is equipped with two cameras per camera unit 2601.

[0101] Next, the functional configuration of the ROV100A will be described. The control computer 1501 of the ROV 100A includes a main memory device, an auxiliary memory device, and a processor. The main memory device stores programs such as an image correction module 1531, an image analysis module 1532, a movement control module 1533A, and a cleaning control module 1534A (all not shown). The processor executes these programs to realize various functions of the ROV 100A.

[0102] Each module may be implemented in hardware by integration, etc. Each module may be an independent program or application, or may be implemented as a subprogram or function in one integrated program or application. The image correction module 1531 and the image analysis module 1532 may be provided on the measurement computer.

[0103] Among the functions of the ROV100A, the image correction module 1531 and the image analysis module 1532 are the same as those in the first embodiment, and therefore their explanation will be omitted. Below, only the movement control module 1533A and the cleaning control module 1534A, which are the differences from the first embodiment, will be explained. These modules execute different processes from those in the first embodiment because each camera unit 2601 has two cameras.

[0104] First, the movement control module 1533A controls the wheel motor 704 based on the camera image (i.e., the camera image of the rear camera unit 2601B) acquired from the measurement computer 1309B. Specifically, the module causes the ROV 100A to move forward when the image analysis module 1532 analyzes two rear camera images (i.e., the camera image of the camera 1307B and the camera image of the camera 1307D) and determines that cleaning is complete for both camera images. On the other hand, the module causes the ROV 100A to move backward when the image analysis module 1532 determines that cleaning is not complete for either camera image. This allows the target location to be cleaned again.

[0105] The module also determines that cleaning is complete when the number of cleaning attempts exceeds a certain threshold, and moves the ROV100A forward to avoid redundant cleaning to protect the hull's paintwork.

[0106] The cleaning control module 1534A controls the brush motor 1503 based on the camera image (i.e., the camera image of the front camera unit 2601A) acquired from the measurement calculator 1309A. Specifically, the module analyzes two front camera images (i.e., the camera image of the camera 1307A and the camera image of the camera 1307C) by the image analysis module 1532, controls the number of rotations of the rotating brush 104 according to the value of the defect amount output for each camera image, and improves the cleaning ability. At that time, the module controls the number of rotations of the rotating brush 104 according to the larger value of the defect amount output. Also, the larger the value of the defect amount, the greater the cleaning ability is, for example, by increasing the number of rotations of the rotating brush 104. This makes it possible to realize cleaning according to the defect amount.

[0107] 2-2. Operation method The operation method of the ROV100A is the same as in the first embodiment, so the explanation will be omitted.

[0108] 2-3.Operation Next, we will explain the operation of the ROV100A. Specifically, we will explain the control flow during cleaning and the defect detection process flow. First, the control flow during cleaning will be described.

[0109] The control flow during cleaning of the ROV 100A is generally the same as in Example 1. Therefore, in the following description, reference will be made again to FIG.

[0110] In the flow 2400 shown in the figure, the movement control module 1533A of the ROV 100A first resets the resharpening counter (step 2401). Next, the module controls the wheel motor 704 to move the ROV 100A forward (step 2402). Next, steps 2403 to 2405 are executed in parallel.

[0111] First, in step 2403, the image analysis module 1532 analyzes the camera image acquired from the measurement computer 1309A (i.e., the camera image of the front camera unit 2601A) to detect defects. Note that the defects referred to here are marine biological fouling (in other words, marine loss). Incidentally, the camera image analyzed here is a camera image corrected by the image correction module 1531 .

[0112] In step 2404, the cleaning control module 1534A controls the rotation speed of the rotating brush 104 in accordance with the amount of defects detected in step 2403. In this case, the module controls the rotation speed of the rotating brush 104 in accordance with the larger of the defect amount values ​​output for the two camera images. Also, the module increases the rotation speed of the rotating brush 104 as the defect amount value increases.

[0113] In step 2405, the image analysis module 1532 analyzes the camera image acquired from the measurement computer 1309B (i.e., the camera image of the rear camera unit 2601B) to detect defects. Incidentally, the camera image analyzed here is the camera image corrected by the image correction module 1531.

[0114] If the analysis result of step 2405 indicates that there is no defect in either of the two camera images ("no marine life remaining"), the movement control module 1533A returns to step 2402 and controls the wheel motor 704 to move the ROV 100A forward. On the other hand, if the analysis result of step 2405 indicates that there is a defect in either of the camera images ("marine life remaining"), the movement control module 1533A increments the re-scrubbing counter (step 2406). Then, if the incremented value is equal to or less than the upper limit value ("upper limit or less"), the movement control module 1533A controls the wheel motor 704 to move the ROV 100A backward (step 2407). This allows the target area to be cleaned again. After the ROV 100A retreats, steps 2403 to 2405 are executed in parallel again.

[0115] On the other hand, if the incremented value exceeds the upper limit value ("Exceeds Upper Limit Value"), the module returns to step 2401 and resets the resharpening counter. Then, the module controls the wheel motors 704 to move the ROV 100A forward (step 2402). The above is the explanation of the control flow 2400.

[0116] Next, the defect detection process flow will be described. The defect detection process flow of the ROV100A is generally the same as that of the first embodiment, so in the following description, reference will be made again to FIG.

[0117] In the flow 2500 shown in the figure, the image correction module 1531 of the ROV 100A corrects the camera image acquired from the measurement computer 1309A or 1309B (step 2501). Specifically, the module performs lens distortion correction, perspective distortion correction, and contrast correction.

[0118] Next, the image analysis module 1532 analyzes the corrected image generated by the image correction module 1531 to detect defects (step 2502). The defects referred to here are marine biological adhesion (in other words, marine loss). Specifically, the module inputs the corrected image to a trained model and obtains an output value indicating the amount of defects. If the obtained output value exceeds a predetermined threshold, the module determines that there is a defect, and if the obtained output value does not exceed the predetermined threshold, the module determines that there is no defect. The module outputs the determination result as defect information to the movement control module 1533A and the cleaning control module 1534A. The defect detection process flow 2500 has been described above.

[0119] The ROV 100A described above has an advantage of being thinner than the ROV 100 according to the first embodiment.

[0120] 3. Third Example The ROV 100B according to this embodiment lacks the front camera unit 2601A compared to the ROV 100A according to the second embodiment. That is, the ROV 100B has only one camera unit 2601. Hereinafter, the ROV 100B having only the single camera unit 2601 will be described with reference to the drawings.

[0121] 3-1.Configuration Figures 43 to 45 show an example of the appearance of the ROV 100B. Figure 43 is a perspective view seen from above, Figure 44 is a perspective view seen from below, and Figure 45 is a plan view. In these figures, the same components as those in the second embodiment are denoted by the same reference numerals.

[0122] The ROV 100B generally comprises a frame 101B, ten thrusters 102A to 102J, a rear camera unit 2601B, three rotating brushes 104, a control unit 105, two drive wheels 106, and casters 107.

[0123] This ROV 100B differs from the second embodiment in that it has a frame 101B instead of the frame 101A and does not have a front camera unit 2601. Of these differences, the frame 101B will be described.

[0124] A frame 801A of the frame 101B has a C-shape in plan view. Compared to the frame 801 according to the second embodiment, this frame 801A is missing a front frame side except for both left and right ends.

[0125] Next, the electrical configuration of the ROV 100B will be described. Figure 46 shows an example of the electrical configuration of the ROV 100B. In this figure, the same components as those in the second embodiment are given the same reference numerals.

[0126] Compared to ROV 100A, ROV 100B lacks cameras 1307A, 1307C, measurement calculator 1309A, driver 1507, and lighting 1310A because ROV 100B lacks forward camera unit 2601A.

[0127] Next, the functional configuration of the ROV100B will be described. The control computer 1501 of the ROV 100B includes a main memory device, an auxiliary memory device, and a processor. The main memory device stores programs such as an image correction module 1531, an image analysis module 1532, and a movement control module 1533A (all not shown). The processor executes these programs to realize various functions of the ROV 100B.

[0128] In addition, the main memory of the ROV 100B does not store the cleaning control module 1534A, as compared with the second embodiment. This is because the ROV 100B lacks the front camera unit 2601A and does not control the brush rotation speed based on the front camera image.

[0129] 3-2. Operation method The operation method of ROV100B is the same as in the first embodiment, so the explanation will be omitted.

[0130] 3-3.Operation Next, we will explain the operation of the ROV100B. Specifically, we will explain the control flow during cleaning and the defect detection process flow. First, the control flow during cleaning will be described.

[0131] FIG. 47 is an activity diagram showing an example of a control flow during cleaning. In the flow 4700 shown in the figure, the movement control module 1533A of the ROV 100B first resets the resharpening counter (step 4701). Next, the module controls the wheel motor 704 to move the ROV 100B forward (step 4702).

[0132] Next, the image analysis module 1532 analyzes the camera image acquired from the measurement computer 1309B (i.e., the camera image of the rear camera unit 2601B) to detect defects (step 4703). Incidentally, the camera image analyzed here is the camera image corrected by the image correction module 1531.

[0133] If the analysis result of step 4703 indicates that there is no defect in either of the two camera images ("no marine life remaining"), the movement control module 1533A returns to step 4702 and controls the wheel motor 704 to move the ROV 100B forward. On the other hand, if the analysis result of step 4703 indicates that there is a defect in either of the camera images ("marine life remaining"), the movement control module 1533A increments the re-scrubbing counter (step 4704). Then, if the incremented value is equal to or less than the upper limit value ("upper limit or less"), the movement control module 1533A controls the wheel motor 704 to move the ROV 100B backward (step 4705). This allows the target area to be cleaned again. After ROV100B retreats, step 4703 is executed again.

[0134] On the other hand, if the incremented value exceeds the upper limit value ("Exceeds Upper Limit Value"), the module returns to step 4701 and resets the resharpening counter. Then, the module controls the wheel motor 704 to move the ROV 100B forward (step 4702). The above is the explanation of the control flow 4700.

[0135] Next, the defect detection process flow will be described. The defect detection process flow of the ROV100B is generally the same as that of the first embodiment, so in the following description, reference will be made again to FIG.

[0136] In a flow 2500 shown in the figure, the image correction module 1531 of the ROV 100B corrects the camera image acquired from the measurement computer 1309B (step 2501). Specifically, the module performs lens distortion correction, perspective distortion correction, and contrast correction.

[0137] Next, the image analysis module 1532 analyzes the corrected image generated by the image correction module 1531 to detect defects (step 2502). The defects referred to here are marine biological adhesions (in other words, marine loss). Specifically, the module inputs the corrected image into a trained model and obtains an output value indicating the amount of defects. If the obtained output value exceeds a predetermined threshold, the module determines that there is a defect, and if the output value does not exceed the predetermined threshold, the module determines that there is no defect. The module outputs the determination result as defect information to the movement control module 1533A. The defect detection process flow 2500 has been described above.

[0138] The ROV100B described above lacks the forward camera unit 103A compared to the second embodiment. However, this ROV100B can remove marine loss on the hull surface, as in the second embodiment. As a result, the fuel efficiency of the ship is reduced, and organisms from other regions attached to the bottom of the ship can be prevented from being brought into the port. In addition, this ROV100B can prevent cleaning omissions by repeatedly cleaning if there are any cleaning omissions after cleaning. Furthermore, this ROV100B can avoid redundant cleaning to protect the painted surface of the hull.

[0139] 4. Variations The above embodiment may be modified as follows. The following modifications may be combined and used.

[0140] (1) Means of transportation The ROV according to each of the above embodiments moves on the hull surface using the drive wheels 106. However, the drive wheels 106 are merely one example of a means for moving on the hull surface, and the ROV may be any means capable of moving on or around the surface to be cleaned, or may be any other means for moving. For example, instead of the drive wheels 106, the ROV may move (i.e., move forward and backward) on the hull surface using thrusters 102A-102H.

[0141] (2) Cleaning means The ROV according to each of the above embodiments is equipped with a rotating brush 104 as a cleaning means. However, the rotating brush 104 is merely one example of the cleaning means. Instead of the rotating brush 104, the ROV may be equipped with a roller type, a high-pressure cleaning type, or a laser type cleaning means.

[0142] (3) Cleaning target The ROV according to each of the above embodiments is intended to clean the bottom of a ship. However, the bottom of a ship is merely one example of a cleaning target. The ROV may be used to clean underwater structures such as bridge piers.

[0143] (4) Camera The ROV according to each of the above embodiments is equipped with a camera. This camera may be a visible light image sensor or a sensor that uses the reflection speed of light or sound waves to generate an image of the cleaning target. Specifically, it may be a time-of-flight camera, a LiDAR scanner, a multi-beam sonar, a Doppler speedometer, an infrared camera, an ultraviolet camera, an acoustic camera, or the like. By using these sensors, it is possible to obtain an image that clearly shows the cleaning target even in water with low transparency. The image includes both a still image and a video (including a still image that is part of a video). It can be used to make a decision for the next cleaning. For example, if there is anything remaining even after cleaning, it can be excluded from the decision for the next cleaning, which leads to a reduction in cleaning time.

[0144] (5) Movement control In each of the above embodiments, the movement (forward and backward) of the ROV is controlled based on the camera images of the camera unit. Alternatively, the movement of the ROV may be controlled based on the friction coefficient of the hull. This is because if the friction coefficient of the hull is high, it is believed that marine organisms are attached to the hull. By repolishing the hull when the friction coefficient of the hull is high, cleaning omissions can be prevented. If the images between the two cameras can be matched, the amount of slippage can be estimated from the difference with the estimated amount of movement from the wheels, which leads to an estimation of the friction coefficient.

[0145] When this modification is adopted, the ROV is provided with an acquisition means for acquiring the friction coefficient of the surface to be cleaned. Specifically, this acquisition means is a sensor for measuring the number of rotations of the casters 107 (in other words, slippage) and a sensor for measuring the torque applied to the rotating brush 104. Based on the output values ​​of these sensors, the movement control module may control the wheel motors 704 to move the ROV backward when the friction coefficient exceeds a threshold value, and to move the ROV forward when the friction coefficient does not exceed the threshold value. Additionally, the cleaning control module may also control the brush motor 1503 based on the coefficient of friction of the hull.

[0146] (6) Control of cleaning methods The ROV according to the first or second embodiment controls the rotation speed of the rotating brush 104 based on the camera image of the front view. This enables cleaning according to the level of dirt. However, this control method is merely one example. As another control method, the rotating brush 104 may be held so that it can swing up and down, and the rotating brush 104 may be pressed against the hull by an actuator or thruster. In this case, the rotating brush 104 may be pressed strongly against the hull when the surface is heavily dirty, and may be pressed lightly against the bottom of the vessel when the surface is lightly dirty.

[0147] (7) Correspondence between image acquisition range and cleaning range In the ROVs according to the above embodiments, the cleaning range of the rotating brush 104 corresponds to the image acquisition range of the camera unit. In other words, the cleaning range and the image acquisition range overlap when viewed from the front. However, the two do not necessarily need to correspond to each other. For example, in the first and second embodiments, the image acquisition range of the front camera unit may be narrower in the left-right direction than the cleaning range.

[0148] (8) Control of movement speed In the above first or second embodiment, the movement speed of the ROV may be changed according to the dirt on the hull. Specifically, the movement control module of the ROV may control the wheel motor 704 based on the camera image of the front camera unit. In this case, the module can realize cleaning according to the dirt by slowing down the movement speed of the ROV in proportion to the analysis result of the camera image (i.e., the amount of defects).

[0149] (9) Switchback In the above first or second embodiment, when reciprocating vertically or horizontally on the surface of the hull, the ROV may switch back. That is, the ROV may move forward on the outward journey and move backward on the return journey to thoroughly clean the surface of the hull. In this case, the ROV may switch the roles of the front and rear camera units on the outward journey and the return journey. Specifically, the cleaning control module may control the brush motor 150 based on the camera image of the front camera unit on the outward journey, and control the brush motor 150 based on the camera image of the rear camera unit on the return journey. In addition, the movement control module may control the wheel motor 704 based on the camera image of the rear camera unit on the outward journey, and control the wheel motor 704 based on the camera image of the front camera unit on the return journey.

[0150] (10) Method of determining defects The ROV according to each of the above embodiments uses machine learning to determine the presence or absence of defects. However, this determination method is merely one example. As another determination method, the presence or absence of defects may be determined by counting the number of pixels representing the color of the bottom of the ship. For example, if the bottom of the ship is red, the image analysis module may count red pixels in the corrected image, and if the counted number of pixels exceeds a threshold, determine that there is no defect, and if not, determine that there is a defect. The RGB values ​​of the pixels to be counted are stored in advance in the ROV.

[0151] (11) Determining whether cleaning is complete When the number of cleanings exceeds a predetermined value, the ROV according to each of the above embodiments determines that cleaning is complete and moves forward. That is, the number of cleanings is referenced to determine whether cleaning is complete. However, instead of the number of cleanings, the cleaning time may be referenced. Then, when the cleaning time exceeds a predetermined value, it may be determined that cleaning is complete and the ROV may move forward. This method also makes it possible to avoid redundant cleaning in order to protect the painted surface of the hull.

[0152] (12) Anomaly detection In each of the above embodiments, rust, blisters, scratches, or peeling paint on the hull may be detected during cleaning work. In this case, the ROV is equipped with an anomaly detection module for detecting rust, blisters, scratches, or peeling paint. This anomaly detection module analyzes the corrected image generated by the image correction module to detect anomalies. The anomalies referred to here are rust, blisters, scratches, or peeling paint. Specifically, the module inputs the corrected image into a trained model and obtains an output value indicating the degree of anomaly. Then, the module determines that there is an anomaly if the obtained output value exceeds a predetermined threshold, and determines that there is no anomaly if the obtained output value does not exceed the predetermined threshold.

[0153] The trained model to which the corrected image is input is generated in advance by training a machine learning model (e.g., a neural network) with training data. The training data is, for example, a collection of data sets each consisting of anomaly images of the hull surface and the degree of anomaly.

[0154] If the anomaly detection module determines that there is an anomaly, it may output an anomaly signal to the PC 1511 on the ship. In addition, if the module determines that there is an anomaly, it may determine that cleaning is complete. In response to this determination, the movement control module may stop the retreat of the ROV for repolishing and move the ROV forward. This is because there is a risk of damaging the hull if repolishing is repeated when there is rust, blistering, scratches, or peeling paint on the hull.

[0155] (13) Positioning In each of the above embodiments, a rotation speed sensor may be used as one of the sensors for measuring the position of the ROV. The movement distance of the ROV 100 can be determined by measuring the number of rotations of the caster 107 with this rotation speed sensor. As another method, the extension length of the communication cable 1520 may be measured, and the horizontal movement distance of the ROV may be determined based on this extension length.

[0156] (14)Functional layout In each of the above embodiments, the ROV control computer 1501 realizes various functions such as image correction, image analysis, movement control, and cleaning control. However, one or more of these functions may be realized by the ROV measurement computer 1309 or the onboard PC 1511.

[0157] (15)Other The present invention is not limited to the above-described embodiments, and includes various modified examples. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the configurations described. It is also possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. It is also possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

[0158] In addition, the above-mentioned configurations, functions, processing units, processing means, etc. may be realized in part or in whole by hardware, for example, by designing them as integrated circuits. In addition, the above-mentioned configurations, functions, etc. may be realized in software by a processor interpreting and executing a program that realizes each function. Information such as the program, table, file, etc. that realizes each function can be stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD.

[0159] In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and not all control lines and information lines in the product are necessarily shown. In reality, it can be considered that almost all components are connected to each other. The above-described embodiments disclose at least the configurations described in the claims. [Explanation of symbols]

[0160] 100, 100A, 100B...ROV, 101, 101A, 101B...frame, 102A to 102J...thruster, 103A...front camera unit, 103B...rear camera unit, 104...rotating brush, 105...control unit, 106...drive wheel, 107...caster, 701...disk portion, 702...brush portion, 703...opening, 704...wheel motor, 801, 801A...frame, 802...front beam portion, 803...rear beam portion, 804...wheel mounting portion, 805...connecting plate, 806...front column portion, 807...notch portion, 808...rear column portion, 809...mounting plate, 810...brush mounting portion, 811...opening, 812...opening, 1301...water box, 1302...glass plate, 1303...camera cover, 1304...partition plate, 1305...opening, 1306...protective dome glass, 1307...image sensor board, 1308...lens, 1309...measurement computer, 1310...illumination light 1310, 1520...communication cable, 1601...small crane, 1602...tether winder, 1801...relay roller, 1802...roller suspension cable, 1901...roller restraining rope, 2601A...front camera unit, 2601B...rear camera unit, 3301...wheel mounting section, 3302...connecting plate, 3801...water box, 3803...camera cover, 3804...small water box, 3805...camera housing section, 3806...computer housing section

Claims

1. 1. A submersible cleaning device comprising: A means of transportation for moving through the water; A cleaning means for cleaning an underwater cleaning target; one or more acquisition means for acquiring an image of the cleaning object or a friction coefficient of a surface of the cleaning object; a movement control means for controlling the movement means based on the image or the friction coefficient; An underwater cleaning device comprising:

2. Further comprising a cleaning control means for controlling the cleaning means based on the image or the friction coefficient.

2. The submersible cleaning device according to claim 1 .

3. The one or more acquisition means are means for acquiring an image of the cleaning target, The one or more acquisition means and the cleaning means are arranged side by side along a direction of travel of the submersible cleaning device; an image acquisition range of the one or more acquisition means extending in a direction substantially perpendicular to the traveling direction corresponds to a cleaning range of the cleaning means extending in the direction substantially perpendicular to the traveling direction; 2. The submersible cleaning device according to claim 1 .

4. The one or more acquisition means are means for acquiring an image of the cleaning target, and include a first acquisition means and a second acquisition means; The first acquisition means and the second acquisition means are arranged on either side of the cleaning means along a traveling direction of the submersible cleaning device.

2. The submersible cleaning device according to claim 1 .

5. the first acquisition means is disposed forward of the second acquisition means in the traveling direction, The movement control means controls the movement means based on the image acquired by the second acquisition means, Further comprising a cleaning control means for further controlling the cleaning means based on the image acquired by the first acquisition means.

5. The submersible cleaning device according to claim 4.

6. The movement control means further controls the movement means based on the image acquired by the first acquisition means to change the movement speed of the submersible cleaning device.

6. The submersible cleaning device according to claim 5 .

7. When the submersible cleaning device performs reciprocating motion, the movement control means controls the movement means on the outbound path based on the image acquired by the second acquisition means, and controls the movement means on the return path based on the image acquired by the first acquisition means, and the cleaning control means controls the cleaning means on the outbound path based on the image acquired by the first acquisition means, and controls the cleaning means on the return path based on the image acquired by the second acquisition means.

6. The submersible cleaning device according to claim 5 .

8. The moving means moves linearly and reciprocally on the surface of the cleaning target, 2. The submersible cleaning device according to claim 1 .

9. Further, an image analysis means is provided for determining whether cleaning is complete or not based on the image, The movement control means controls the movement means to move the submersible cleaning device forward when the cleaning is complete, and controls the movement means to move the submersible cleaning device backward when the cleaning is not complete.

2. The submersible cleaning device according to claim 1 .

10. The movement control means controls the movement means to move the submersible cleaning device forward when the number of cleanings or the cleaning time exceeds a predetermined value in addition to when the cleaning is completed.

10. The submersible cleaning device according to claim 9.

11. further comprising an abnormality detection means for detecting rust or blisters on the surface of the object to be cleaned based on the image and outputting an abnormality signal when rust or blisters are detected.

2. The submersible cleaning device according to claim 1 .

12. Further, an abnormality detection means is provided for detecting rust or blisters on the surface of the object to be cleaned based on the image, the movement control means controls the movement means to advance the submersible cleaning device when rust or blister is detected; 2. The submersible cleaning device according to claim 1 .

13. The image analysis means inputs the image into a trained model and determines whether cleaning is complete or not based on the output value.

10. The submersible cleaning device according to claim 9.

Images