WO2016060139A1

WO2016060139A1 - Bridge-inspecting robot system

Info

Publication number: WO2016060139A1
Application number: PCT/JP2015/078968
Authority: WO
Inventors: 野中　俊一郎; 渡辺　幹夫; 直孝式田; 文敬山崎
Original assignee: 富士フイルム株式会社; 一般財団法人首都高速道路技術センター; 株式会社イクシスリサーチ
Priority date: 2014-10-14
Filing date: 2015-10-13
Publication date: 2016-04-21
Also published as: JP6355517B2; JP2016079614A

Abstract

Provided is a bridge-inspecting robot system that reduces the movement restrictions that are imposed by the structure of a bridge and that can reliably inspect a lower surface section and both side sections of the bridge. A first suspended dolly (11) and a second suspended dolly (12) have four wheels (41-44) and first through third displacement parts (46-48) and travel while suspended from a lower flange (30c). When an obstruction such as a splice plate (31) is present, the first wheel (41) through the third wheel (43) are withdrawn from an upper surface of the lower flange (30c) by means of the first through third displacement parts (46-48). The first suspended dolly (11) and the second suspended dolly (12) are coupled by a rail (13) that has a total length (L4) that is longer than the interval (L3) between main girders (25). A camera dolly (73) travels the rail (13) and passes through a lower part of the suspended dollies (11, 12). The present invention makes it possible to capture images of not only an inside surface of two of the main girders (25) of an inspection target but also an outside surface of the girders (25).

Description

Bridge inspection robot system

The present invention relates to a bridge inspection robot system suitable for, for example, inspecting the lower surface of a bridge floor slab and steel girder, etc., and has applied for a patent related to the results of commissioned research by the national government (2014-2018, a new administrative agency). Energy and Industrial Technology Development Organization “Development project for social issues such as infrastructure maintenance and renewal, etc./Development of robot technology for infrastructure maintenance and non-destructive inspection equipment / Study on a robot approach to close proximity visual inspection equipped with a compound eye imaging device "Development" contract research, patent application subject to Article 19 of the Industrial Technology Strengthening Act).

河川 Various bridges such as river bridges and viaducts are regularly inspected as part of maintenance management after erection to confirm the deterioration and damage of components. The inspection of the bridge is generally a so-called proximity inspection in which an inspector approaches the member and visually observes the member.

There are some parts of the bridge that are difficult for the inspector to access, such as the lower surface of the floor slab and the part between the main girders of the steel girder arranged on the lower surface of the floor slab. In order to perform close inspection of such parts, temporary scaffolds are placed along the girders on the lower surface of the floor slab to bring the inspector closer to the main girder and floor slab, or the inspector is used using an aerial work vehicle. It is approaching the lower surface of the floor slab.

However, when a temporary scaffold is arranged on the lower surface of the floor slab, there is a disadvantage that it takes a lot of labor and cost. In addition, when the lower surface of the floor slab is inspected with a work vehicle, it is necessary to occupy a part of the road by stopping the vehicle main body, which is a large vehicle, on the floor slab, which disadvantageously affects road traffic.

In order to solve such an inconvenience, there has been proposed a work trolley equipped with a camera and configured to run while being engaged with the lower end of a bridge girder. For example, in the work cart described in Japanese Patent Application Laid-Open No. 2003-119721, drum-shaped wheels are respectively provided at the tips of a plurality of arms that are swingably attached to the work cart, and these drum-shaped wheels are used for bridge I. The both ends of the lower flange of the girders made of steel molds are sandwiched and the wheels are driven to drive the work carriage along the girders. The work cart is remotely controlled for traveling by the wheel and the operation of the camera, and approaches a predetermined part of the bridge to take a picture with the camera. As a result, a remote inspector visually inspects the camera image and inspects the bridge.

In the bridge inspection device described in Japanese Patent Application Laid-Open No. 2013-194457, two wire ropes constructed along a bridge girder, a carriage traveling on the wire rope, and an inspection device mounted on the carriage are used. The bridge is being inspected. In addition to the traveling wheels, the carriage includes reaction force wheels that contact the lower flanges of the girders. The reaction force wheel is pressed against the lower flange by the damper device, and the traveling wheel is pressed against the wire rope by the reaction force accompanying this, and the carriage is reliably moved along the girder.

For example, some highway bridges are composed of piers, steel girder, floor slabs, etc. The pier is a foundation member that is spaced apart in the longitudinal direction of the bridge and receives a steel girder, floor slab, and the like. Steel girder is arranged on the upper surface of the pier. The steel girder is composed of a plurality of main girders, a horizontal girder connecting them, a counter-tilt structure, and the like, and a floor slab made of, for example, reinforced concrete is fixed to the upper part. The main girder is constituted by connecting a plurality of I-type steel materials with an attachment plate. The I-type steel material is formed in an I-type from a vertical plate and flanges arranged on upper and lower ends thereof. The attachment plate is attached to the I-type steel material by fastening with bolts and nuts, welding, or the like. The anti-tilt structure is a structure in which a plurality of members are combined in, for example, a triangle.

In this way, the steel girder is provided with a horizontal girder and various reinforcing members such as an inclined structure between the main girders, and a connecting plate and its fixing bolts are arranged at regular intervals in the longitudinal direction. For this reason, even if an attempt is made to photograph the lower surface of a bridge using a cart such as that disclosed in Japanese Patent Application Laid-Open No. 2003-119721, there is a problem that these various reinforcing members and attachment plates obstruct the progress of the cart. For example, a vertical stiffener may be provided on the side surface of the vertical plate of the main girder. This vertical stiffener may extend in a direction perpendicular to the vertical plate and have a width that reaches the edge of the lower flange. In such a case, in the work cart disclosed in Japanese Patent Application Laid-Open No. 2003-119721, it is difficult for the wheels sandwiching the both end edges of the lower flange to get over the vertical stiffener, and as a result, the range of traveling along the main girder There is a problem that is limited. Further, when the camera is provided with a support member and the camera is projected upward from the carriage to capture an image of the inside of the steel girder, the support member hits various reinforcing members, which hinders the progress of the work carriage.

Furthermore, the work carriage disclosed in Japanese Patent Laid-Open No. 2003-119721 has a problem that the engagement of the wheel is easily released when the thickness of the lower flange with which the drum-shaped wheel is engaged is large. For this reason, there exists a possibility that a wheel may drop | omit at the junction part by the attachment board which becomes larger than the thickness of a lower flange.

As disclosed in Japanese Patent Application Laid-Open No. 2013-194457, when a work carriage is run using a wire rope, it is necessary to provide the wire rope in parallel with the main girder so as to avoid various reinforcing members. For this reason, it is necessary for the inspector to go near the steel girder, visually check the steel girder to be inspected, and stretch the wire rope while avoiding these various reinforcing members. Thus, there is a problem that time and labor are required for the wire rope tensioning operation which is a preliminary operation before the inspection operation.

Further, in the inspection apparatus described in Japanese Patent Application Laid-Open No. 2003-119721 or Japanese Patent Application Laid-Open No. 2013-194457, the inside of adjacent main girders to be inspected can be inspected, but the outside of the main girder to be inspected In the case of inspecting, it is necessary to reattach the work carriage to the inspection object side, and work efficiency is lowered.

An object of the present invention is to provide a bridge inspection robot system capable of reliably inspecting a lower surface portion of a bridge while reducing restrictions on movement due to the structure of the bridge. It is another object of the present invention to provide a bridge inspection robot system that can easily inspect the main girder and floor slab on the opposite side of the inspection object without re-installing the camera or carriage.

The bridge inspection robot system of the present invention inspects a bridge in which a plurality of main girders having a lower flange are arranged apart from each other in a direction perpendicular to the longitudinal direction of the main girder. The bridge inspection robot system includes a first suspension carriage and a second suspension carriage, a rail, a camera, and an image processing unit. The first suspension carriage and the second suspension carriage are attached to a plurality of main girders and travel in the longitudinal direction of the main girders. The rail is attached to the first suspension carriage and the second suspension carriage, is formed longer than the interval between the main girders to which the first suspension carriage and the second suspension carriage are attached, and protrudes from at least one suspension carriage. The camera travels attached to a rail and moves beyond one suspended carriage. The image processing unit identifies a damaged portion of the bridge based on the camera data.

The first suspension carriage and the second suspension carriage are in contact with the upper surface of the lower flange of the main girder and are spaced apart in the longitudinal direction of the main girder, and the first wheel and the second wheel. It is preferable to have a first displacement portion and a second displacement portion that are displaced between a travel position that individually contacts the upper surface of the lower flange and a retracted position that is retracted from the upper surface of the lower flange. In this case, when there is a protrusion in the traveling direction of the suspension carriage, each wheel can be retracted to pass a portion where the protrusion is present. The retreat position refers to a position where the wheel does not hit an obstacle when the suspended carriage travels and does not become an obstacle to travel.

The first displacement unit and the second displacement unit have an obstacle sensor that detects an obstacle that is an obstacle to travel of the first wheel and the second wheel, and the first wheel is based on an obstacle detection signal from the obstacle sensor. Alternatively, it is preferable to move the second wheel from the traveling position to the retracted position. In this case, each wheel can be reliably retracted from the obstacle, and the portion where the obstacle is present can be reliably passed.

It is preferable that the offset length in the longitudinal direction of the main girder of the first wheel and the second wheel, which are arranged apart from the longitudinal direction of the main girder in the longitudinal direction of the main girder, is longer than the longitudinal length of the obstacle. . In this case, in addition to the wheel that retreats from the obstacle, a wheel that is in contact with the lower flange and maintains the suspended state is always secured, so that the suspended carriage does not fall off the lower flange.

The first suspension carriage and the second suspension carriage are in contact with the upper surface of the lower flange of the main girder, and are arranged with a first wheel and a second wheel that are spaced apart in the longitudinal direction of the main girder, and a first wheel and a second wheel. It is preferable to have the 1st displacement part and the 2nd displacement part which displace between the running position which contacts the upper surface of a lower flange separately, and the climbing position which gets over the protrusion from the upper surface of a lower flange. In this case, if there is a protrusion in the traveling direction of the suspension carriage, it can pass over it.

It is preferable that the camera has a camera body, an elevating part that moves the camera body up and down, and a camera carriage that moves the elevating part on a rail. In this case, the camera can be set at an arbitrary position on the lower surface of the bridge, and an image without a blind spot can be obtained.

The camera body is preferably a binocular camera that obtains a plurality of parallax images. In this case, the size and position of the damaged portion of the bridge can be known using parallax.

The image processing unit detects a position of a member between a plurality of main beams based on a plurality of parallaxes as a distance from the camera body, specifies a three-dimensional coordinate position of the member based on a moving distance of the suspension carriage and the camera carriage, It is preferable to specify the photographing position of the camera body that avoids the member based on the three-dimensional coordinate position. In this case, it is possible to obtain a fluoroscopic image without various reinforcing members.

It is preferable that the first suspension carriage and the second suspension carriage have a lower surface contact wheel that contacts the lower surface of the lower flange. In this case, the first wheel and the second wheel can be reliably brought into contact with the upper surface of the lower flange, and the posture of the suspension carriage is maintained when the first wheel or the second wheel is in the retracted position. Therefore, stable running is possible.

It is preferable to have a controller for remotely operating the first and second suspension carts and the camera. In this case, a desired area can be photographed by setting the camera at an arbitrary position.

According to the present invention, it is possible to reliably inspect the lower surface portion of the bridge with less restriction on movement due to the structure of the bridge. In addition, the main girder and the bottom of the floor slab on the opposite side of the inspection object can be easily inspected without re-installing the camera or the carriage.

It is a front view which shows the bridge inspection state of the bridge inspection robot system of this invention. FIG. It is a front view which shows the movement of a suspension cart and a camera. It is a front view which notches and shows a part of suspension cart. It is a top view which shows arrangement | positioning of the wheel of a suspension cart, and a camera cart. It is a perspective view which shows a camera. It is a top view which shows arrangement | positioning of the wheel of the suspension trolley | bogie in another embodiment, and a camera trolley | bogie. It is front sectional drawing which shows the displacement part of the suspension trolley | bogie in other embodiment. It is a side view which shows the movement of rolling contact of the suspension trolley | bogie in other embodiment.

As shown in FIG. 1, the bridge inspection robot system 10 of the present invention includes a first suspension carriage 11 and a second suspension carriage 12, a rail 13, and an obstacle sensor (obstacle detection unit) 14 (see FIG. 4). The camera 15 and the image processing unit 16 (see FIG. 3) are provided.

The bridge 20 to be inspected is composed of, for example, a bridge pier 21, a steel slab girder 22, and a floor slab 23 made of reinforced concrete, and the floor slab 23 is fixed on the pier 21 by a steel spar 22. Hereinafter, for convenience of explanation, the longitudinal direction of the bridge is defined as the X-axis direction, the width direction orthogonal to the longitudinal direction of the bridge is defined as the Y-axis direction, and the vertical direction is described as the Z-axis direction.

As shown in FIG. 2 and FIG. 3, the steel girder 22 includes a plurality of main girders 25 arranged in the longitudinal direction of the bridge 20, and reinforcement such as a horizontal girder (not shown) that connects these main girders 25 and a tilting structure 26. It is comprised from the member. The main girder 25 is configured by connecting a plurality of I-type steel materials 30 by means of an attachment plate 31. The I-type steel material 30 has an I-shaped cross section in which an upper flange 30b is arranged at the upper end of the vertical plate 30a and a lower flange 30c is arranged at the lower end. The attachment plate 31 is attached to the I-type steel material 30 by fastening with bolts and nuts (not shown), welding, or the like, and connects the I-type steel materials 30 to each other in the X-axis direction.

The anti-tilt structure 26 is configured as a reinforcing member based on a triangle using an L-shaped steel material 27, a mounting plate 28, and the like. Various members such as a load distribution cross beam and a reinforcing cross beam are arranged between the I-type steel members 30 in addition to the counter tilting structure 26, but these are not shown.

As shown in FIG. 4, the first suspension carriage 11 includes a rectangular base 40, first to third wheels 41 to 43 attached to the base 40, a lower surface contact wheel 44, and a connecting block 45 (see FIG. 4). 2). The first to third wheels 41 to 43 are displaced by the first to third displacement portions 46 to 48 between a traveling position indicated by a solid line in FIGS. 2 and 3 and a retracted position indicated by a two-dot chain line. In the present embodiment, a certain position that is the maximum opening angle is set as the retreat position so that all obstacles predicted in advance can be retreated. However, the position may be any position where the wheels 41 to 43 do not hit the obstacle, and the retraction position (retraction angle) may be changed according to the size of the obstacle.

As shown in FIG. 4, the first wheel 41 and the second wheel 42 are provided near one side edge of the base 40 and near the corners on the front end side and the rear end side. The third wheel 43 is provided at an intermediate position in the X direction near the other side edge of the base 40. The left and right wheels are arranged so as to be shifted in the X-axis direction between the first wheel 41 and the third wheel 43 and between the third wheel 43 and the second wheel 42, for example. The shift length (offset length) L2 is set longer than the obstacle length L1 in the X-axis direction. By having this offset length L2, when one side edge wheel, for example, the first wheel 41 is in the retracted position to avoid the obstacle plate 31 that is an obstacle, one remaining wheel For example, the second wheel 42 and the wheel on the other side edge, for example, the third wheel 43 are surely in contact with the upper surface of the lower flange 30c. Thereby, since it is always suspended at two points, the

suspension carts

11 and 12 are prevented from falling off from the lower flange 30c. Further, since the lower surface contact wheel 44 is in contact with the lower surface of the lower flange 30c, the horizontal position is maintained without the

suspension carts

11 and 12 being inclined.

As shown in FIG. 3, the lower surface contact wheel 44 is attached to the center of the base 40 by a support bracket 50. The support bracket 50 rotatably holds the lower surface contact wheel 44 and presses the lower surface contact wheel 44 toward the lower surface of the lower flange 30c by a coil spring (not shown). The support bracket 50 is not displaced like the first to third displacement portions 46 to 48.

The wheels 41 to 44 have a built-in motor, and the wheels 41 to 44 rotate in synchronization with the control of the first controller 51. Accordingly, the first suspension carriage 11 travels in the X direction while being suspended downward from the lower flange 30c. Each wheel 41 to 44 may be rotated by an external motor instead of the motor built-in type wheel.

The first displacement portion 46 includes a wheel holding bracket 55 and a swinging portion 56 that swingably holds the wheel holding bracket 55. The wheel holding bracket 55 is bent in an L shape so that the first wheel 41 is perpendicular to the upper surface of the lower flange 30c. The swinging portion 56 of the first displacement portion 46 swings the wheel holding bracket 55 by a motor (not shown) to bring the first wheel 41 into contact with the upper surface of the lower flange 30c (indicated by a solid line), and the lower flange It is displaced between the retracted position retracted from the upper surface of 30c (indicated by a two-dot chain line). The other second displacement portion 47 (see FIG. 4) and the third displacement portion 48 are configured in the same manner as the first displacement portion 46. In the following description, the same constituent members are denoted by the same reference numerals, and redundant description is omitted. The second suspension carriage 12 is also configured in the same manner as the first suspension carriage 11. The arrangement of the wheels 41 to 43 is opposite to that of the first suspension carriage 11.

As shown in FIG. 4, the first suspension carriage 11 and the second suspension carriage 12 are provided with an obstacle sensor 14 that detects an obstacle ahead of each of the wheels 41 to 43 in the traveling direction. The obstacle sensor 14 detects an obstacle having such a height that the wheels 41 to 43 cannot get over, such as the attachment plate 31 and the mounting bolt nut. As the obstacle sensor 14, a sensor that detects an obstacle mechanically by an actuator or a sensor that can detect an obstacle by optical, magnetic, or other methods is used.

As shown in FIG. 2, a rail 13 is detachably attached to the first suspension carriage 11 and the second suspension carriage 12 via a connection block 45. The rail 13 connects the first suspension carriage 11 and the second suspension carriage 12 and is spanned between the two main girders 25. As shown in FIG. 1, the rail 13 has a total length L4 that is longer than the interval L3 of the main girder 25, and the width direction (Y direction) of the bridge 20 from one or both of the first suspension carriage 11 and the second suspension carriage 12. ).

As shown in FIG. 2, the camera 15 includes a camera body 70, a holding unit 71, an elevating unit 72, and a camera carriage 73. As shown in FIG. 5, the camera body 70 is constituted by a twin-lens camera, and a plurality of photographing lenses 70a are arranged in the Y direction. The camera body 70 is attached to the upper end of the elevating part 72 by a holding part 71. The holding unit 71 rotates the camera body 70 around the Y axis in an angle range of 270 °, for example, and rotates the camera body 70 around the Z axis in an angle range of 270 °, for example. Note that these rotation angle ranges may be changed as appropriate. In short, it is only necessary to change the posture of the camera body 70 and capture the entire area of the lower surface of the bridge 20 without blind spots.

The camera body 70 has higher functionality than a normal camera. For example, the influence of camera shake can be suppressed by having a high brightness illuminator 70b such as an LED and performing shutter control at a high speed of, for example, 1/500 second or more. In addition, it has a shake correction control function such as an optical shift method, and the influence of camera shake can be suppressed. Furthermore, in addition to having a high-sensitivity shooting function and reducing noise, the high-precision exposure control function can cope with brightness fluctuations in outdoor dark areas by improving the photometry area and control program. Image processing for obtaining depth information and removing obstacles by the binocular function is reliably performed. In addition, the length of the damaged portion can be accurately measured by selecting an appropriate baseline length for the binocular function and simplifying calibration before inspection.

As shown in FIG. 2, the elevating unit 72 is attached to a camera carriage 73, and holds the camera body 70 so as to be movable in the vertical direction (Z-axis direction). The raising / lowering stroke of the raising / lowering part 72 is 2 m, for example. At the lowest position, the lower part of the camera body 70 is housed in the recess 73a (see FIG. 4) of the camera carriage 73. At the time of storage, the camera body 70 is held at a position where it does not come into contact with various members of the steel girder 22. The camera main body 70 has a viewing angle that allows the entire lower surface portion of the steel girder 22 to be in the field of view in the lowest lowered state, and further has a zoom function. Further, when it is at its highest position, it can approach the lower surface of the floor slab 23.

As shown in FIGS. 4 and 5, the camera carriage 73 is attached between the two rails 13 and moves on the rails 13. A rail groove 73 b into which a part of the rail 13 is inserted is formed in the front end portion and the rear end portion of the camera carriage 73. The camera carriage 73 only needs to have a structure capable of traveling on the rail 13, and the movement method is not particularly limited. For example, the rail 13 is driven by drive wheels, rack and pinion, traction by wire, ball screw, and the like. In the embodiment, the two rails 13 are used, but this may be one rail or three or more rails. The cross-sectional shape of the rail 13 is not limited to a rectangle, but may be a circle, an ellipse, or another polygon. Further, the holding guide method of the camera carriage 73 by the rail 13 is not particularly limited to the illustrated one, and other holding guide methods may be used.

As shown in FIG. 2, in this embodiment, it is possible to photograph not only between (inside) the main beam 25 on which the suspended

carriages

11 and 12 travel, but also on the outside of the main beam 25 on which the suspended

carriages

11 and 12 travel. Thus, it moves on the rail 13 beyond the main girder 25. For this reason, the rail 13 is attached below the

suspension carriages

11 and 12 so that the connecting block 45 of the

suspension carriages

11 and 12 does not obstruct the traveling of the camera carriage 73. The rails 13 may be attached to the

suspension carts

11 and 12 as long as they do not obstruct the traveling of the camera cart 73 and may be a cantilever structure in addition to the both-end support structure. Stoppers 13 a are attached to both ends of the rail 13 so that the camera carriage 73 does not fall off the rail 13.

The camera body 70 can be stored in the recess 73a of the camera carriage 73. When the elevating part 72 is lowered to the lowest end, the camera body 70 enters the recess 73a. In the stored state of the camera main body 70, the camera carriage 73 can pass through the lower part of the

suspension carriages

11 and 12, move from the inside of the main girder 25 to the outside, and the outside part of the main girder 25 can be photographed. The camera body 70 is not necessarily housed in the recess 73a, and the uppermost part of the camera body 70 is the lowermost of various reinforcing members existing between the main beams 25 in a state where the elevating part 72 is lowered to the lowest position. What is necessary is just to be located below the site | part.

As shown in FIG. 3, the first controller 51 is built in the base 40 of the first suspension carriage 11. The first controller 51 controls each part such as the

suspension carts

11 and 12, the camera main body 70, the holding unit 71, the lifting unit 72, the camera cart 73, and the image data from the camera main body 70 is in the hands of the inspector. Transmit to the second controller 52. Image data is transmitted in a wired or wireless manner. The first controller 51 and the second controller 52 constitute a controller.

The second controller 52 is composed of, for example, a notebook personal computer as shown in FIG. The second controller 52 includes a display 52a, a keyboard 52b, an arithmetic processing unit 52c (see FIG. 3), a data communication unit 52d, a USB port, and the like. An operation member such as a joystick is connected to the USB port as necessary. The arithmetic processing unit 52c is installed with a predetermined application so that the image processing unit 16, the

suspension carts

11 and 12, the camera cart 73, the lifting unit 72, and the holding unit 71 are operated. It functions as the switching unit 76. FIG. 3 shows this functional block.

The second controller 52 receives an operation of an operation member such as a keyboard 52b or a joystick, generates a control signal, sends the control signal to the first controller 51, and remotely operates each unit. The image processing unit 16 processes the image data sent from the first controller 51 and displays a through image from the camera body 70 on the display 52a. In addition, based on the image data from the first controller 51, various reinforcing members are detected and their positions are identified, damaged portions are detected, and their sizes and positions are identified.

The mode switching unit 76 switches between an automatic inspection mode and a manual mode according to an input operation. In the automatic mode, the

suspension carts

11 and 12 and the camera cart 73 are moved according to a predetermined procedure, the lower surface portion of the bridge 20 is photographed, various reinforcing members and damaged portions are identified by image processing, and the positions of the various reinforcing members Memorize the size and evaluation of the damaged part.

It should be noted that, by setting in advance, when a damaged part of a certain size or more is detected, an alarm can be issued to notify the inspector. In this case, the inspector interrupts the automatic mode and switches to the manual mode to perform zoom shooting or shooting with a different shooting angle.

In addition, by inputting the approximate position as the map data in advance for the inspection required location, an alarm can be issued to notify the inspector when these inspection required locations are ready for photographing. Furthermore, a 3D display image is created using a parallax image of a binocular camera.

Hereinafter, the operation of the bridge inspection robot system 10 of the present invention will be described. The inspector can go down to the work landing provided on the lower side of the bridge 20 by a ladder or the like. At the work landing,

suspension carts

11 and 12 are attached to the plurality of main girders 25 of the bridge 20 to be inspected. Thereafter, the rail 13 to which the camera carriage 73 is attached is attached to the

suspension carriages

11 and 12. When the automatic mode is selected and inspection is instructed after various connectors are connected, each unit operates based on a predetermined program, and the inspection is automatically performed.

In the automatic inspection, the

suspension carriages

11 and 12 are reciprocated in the X-axis direction on the lower flange 30c of the main girder 25 to be inspected between the piers 21, and the inspection target part is photographed by the camera body 70. First, in the going-out process, a moving image and a necessary range (the entire lower surface of the floor slab 23, the lower surface of the upper flange 30b of the main girder 25, and a part of the side surface of the vertical portion) can be captured without changing the posture of the camera body 70. A still image is obtained accordingly. In addition, a plurality of image data obtained during the going process is subjected to image processing, and coordinate positions (three-dimensional coordinate positions) on the XYZ axes of various reinforcing members are specified. In specifying the coordinate position, first, a reinforcing member such as a cross beam, the tilting structure 26, and a vertical member is recognized. Next, for each recognized reinforcing member, the position of each reinforcing member in the XYZ axial directions is obtained as a distance from the camera body 70 based on a plurality of parallaxes. Based on the obtained distance data and the position data of the camera body 70, the position of each reinforcing member is specified as a three-dimensional coordinate position. The position data of the camera main body 70 can be specified based on the moving distance of the suspended

carriages

11 and 12 and the camera carriage 73. The moving distance can be obtained based on the number of rotations of each motor, the number of rotations of the wheels 41 to 43, and the like.

Also, a fluoroscopic image in which various reinforcing members are deleted by image processing is obtained within the shootable range below the bridge 20. The plurality of fluoroscopic images are combined to generate an entire image of the bottom surface of the floor slab in the inspection target area. AdaBoost (Adaptive Boosting) using well-known image processing, for example, Haar-like feature amount (scalar amount obtained as the difference value of the average brightness of the rectangular area, representing the intensity gradient) for this whole image By performing image processing such as detection processing by, a portion suspected of being damaged (damage candidate) is detected. In this case, an image processing system for discriminating “damage / non-damage” by learning from the feature amount of the damage / non-damage image prepared in advance is constructed. Similarly, a perspective image of the upper flange surface of the main girder 25 in the inspection target area and a part of the vertical plane continuous thereto is synthesized within the range obtained by the first imaging of the inspection target area. An entire image of the digit 25 is generated. Damage candidates are detected from this entire image. Similarly, an entire image of each reinforcing member in the inspection target area is generated, and damage candidates are detected from the entire image.

Damage includes corrosion, cracks, and cracks in the coating film. These damages change in color, partially swell, or have cracks compared to other normal sites, and the type of the damaged part is specified based on changes in these colors and shapes. Further, the size of the damaged portion is identified from the size of the area of the damaged portion, and the degree of damage is evaluated according to this size. The determination accuracy of the image recognition and the evaluation of the degree of damage is enhanced by the learning effect of the feature amount of the damage / non-damage image prepared in advance, and the presence / absence, size, and evaluation of the damage are reliably performed.

In the returning process, based on the position of each reinforcing member obtained in the going process, the lower surface of the floor slab 23 in the inspection target area, the inner surface of the main girder 25 (the lower surface of the upper flange 30b, the side surface of the vertical portion, the upper surface of the lower flange 30c) ) To identify multiple shooting points. The

suspension carriages

11 and 12, the camera carriage 73, the elevating part 72, and the holding part 71 are controlled so that the camera body 70 is sequentially positioned when returning to the specific point. As a result, the shooting data is supplemented for the portion where shooting was impossible in the going process, and shooting without a blind spot is performed in the inspection target area. Based on this shooting data, the shooting data is supplemented for the part that was a blind spot in the going process in the same way as the going process. As a result, the entire perspective image of the floor slab 23 in the inspection target area, the entire perspective image of the upper flange surface, vertical surface, and lower flange surface of the main girder 25 in the inspection target area, and the entire image of each reinforcing member in the inspection target area Etc. are obtained. Damage candidates are detected from these whole images to obtain damage data.

Also, if necessary, perform zoom photography for damage candidates, measure the size and shape of the damaged part, specify the type of damage, and increase the determination accuracy. Then, the damage type and the damage image are stored as damage data together with the coordinate position of the damaged portion. Identification of the coordinate position of the damaged portion is performed based on the amount of movement of each of the suspended

carriages

11 and 12, the camera carriage 73, the elevating part 72, and the amount of displacement of the holding part 71.

In the inspection record creation mode, the inspection record is automatically created by inserting the damage data into a predetermined inspection record format document stored in advance. The created inspection record can be checked by an inspector and edited or corrected. Since the inspection data is automatically created by fitting the damage data into the inspection document in a predetermined format based on the automatically acquired damage data, the work time of the inspector can be greatly shortened.

The field of view of the camera body 70 is a range in which the lower surface of the floor slab 23 of the bridge 20 and the vertical part of the main beam 25 are reflected when the camera body 70 is arranged in the center between the plurality of main beams 25. If the length in the Y direction (inter-digit length) between the plurality of main girders 25 is long and the whole image is not captured within one shooting screen range, a plurality of shooting points are provided in the Y direction and divided shooting is performed. Then, the obtained divided photographed images are combined and photographed so that the lower surface of the floor slab 23 of the bridge 20 and the vertical portion of the main girder 25 are reflected.

The image used for position detection and damage detection of each reinforcing member may be moving image data or still image data. Still images may be obtained from one frame of a moving image or may be obtained from image synthesis of a plurality of frames.

As shown in FIG. 1, since the left main girder 25 is located on the outermost side, it is necessary to re-inspect the outer surface of the outermost main girder by reattaching the suspension carriage or camera in the conventional inspection method. . On the other hand, in this embodiment, the rail 13 of the suspension carriage 11 suspended from the outermost main girder 25 protrudes outward, and the camera 15 can pass below the suspension carriage 11. It has become. Therefore, the outer part of the suspension carriage can be inspected in the same manner when the inner part is inspected. For this reason, it is not necessary to mount the camera again for the outside inspection again, and the outside portion can be easily and efficiently inspected.

As described above, in the present embodiment, the

suspension carts

11 and 12 are moved in the X direction, and a schematic image of the entire lower surface portion of the bridge 20 is obtained by moving the bridge pier 21 from one side to the other. The position information of various reinforcing members is acquired based on the above, damage candidates are detected based on a fluoroscopic image from which various reinforcing members are deleted, and the damage candidates are photographed by zoom photographing or the like in the return process to determine whether there is damage, the size of the damaged portion, etc. The type and the like can be determined with high accuracy.

In the above-described embodiment, the position of the reinforcing member is specified in the going process, the optimal shooting position is specified based on the position of the reinforcing member in the returning process, and specific shooting data such as the presence or absence of damage is obtained at this specific position. However, the position of the reinforcing member and the damaged portion may be specified in the going process. In this case, after acquiring the image data by moving a certain distance, the position of the reinforcing member is specified, the presence / absence of the damage, the size, and the evaluation are performed based on the image data. The going process includes a few returning processes. However, since all the processes in the going process are not returned, the time required for the inspection can be shortened accordingly.

The camera body 70 is preferably a binocular camera that obtains a plurality of parallax images, but the reinforcing member and damage may be specified using parallax due to movement of the single-lens camera without using the binocular camera. In the case of using a single-lens camera, the 360 ° omnidirectional imaging may be performed by fixing the photographic optical axis to, for example, the Z axis and arranging a convex mirror on the photographic optical axis or using a fisheye lens. In this case, it is preferable to specify the reinforcing member and damage after eliminating the distortion of the entire circumference image obtained by distortion by known image processing.

In the above embodiment, the obstacle sensors 14 are provided on the

suspension carts

11 and 12 to configure the obstacle detection unit. Instead of or in addition to this, the

suspension vehicles

11 and 12 are based on image data from the camera body 70. Obstacles in front of them, such as the attachment plate 31 and bolts and nuts, may be detected.

In the above-described embodiment, the first to third wheels 41 to 43 are provided on the left and right sides of the

suspension carriages

11 and 12, respectively. However, since the

suspension carriages

11 and 12 are integrated by the rail 13, the third carriage located inside is provided. The wheel 43 may be omitted, and the

suspension carts

11 and 12 may be supported by the left and right first wheels 41 and the second wheels 42 only. In this case, there is always one wheel that is in the retracted position with respect to the obstacle, and the other three wheels are always held in the traveling position. Further, the first wheel 41 may be provided on one side edge of the base 40 and the second wheel 42 may be provided on the other side edge of the base 40.

In the above embodiment, the first suspension carriage 11 and the second suspension carriage 12 are brought into contact with the first to third wheels 41 to 43 that can be displaced between the travel position and the retracted position, and the lower surface of the lower flange 30c. In addition to the rotating lower surface contact wheels 44, in addition to the first to third wheels 41 to 43, the total number of right and left displaceable wheels is four as in the second embodiment shown in FIG. The fifth wheel 66 and the fourth displacement portion 67 may be provided in the rear to achieve more stable traveling. Also in this case, the left and right wheels are set to have an offset length L2 longer than the length L1 of the obstacle in the X-axis direction. Although illustration is omitted, the left and right wheels are not limited to two, and may be three or more.

In the above embodiment, the wheels 41 to 43 are swung between the retracted position and the traveling position. However, when there is an area for retracting the traveling wheel above the lower flange 30c, the obstacle is overcome. The traveling wheel may be raised to the retracted position. In this case, as shown in FIG. 7, the wheel holding bracket 81 that holds the wheel 80 of the suspension carriage 79 is held in the holding frame 82 so as to be movable up and down. A coil spring 83 that presses the wheel holding bracket 81 downward is mounted in the holding frame 82. Therefore, when the wheel 80 hits the portion where the contact plate 31 protrudes from the lower flange 30c, the wheel holding bracket 81 retreats upward as indicated by a two-dot chain line against the bias of the coil spring 83. 80 can pass over the attachment plate 31.

In the above embodiment, the suspended

carriages

11 and 12 are caused to travel using the wheels 41 to 44 and 80. However, for example, a rolling wheel 89 may be used as shown in FIG. The rolling wheel 89 includes a wheel body 90, a pulley 91, and an endless belt 92. The wheel body 90 is formed in a substantially triangular shape in which each vertex of the triangle is formed in a round shape. The pulley 91 is disposed near each vertex of the wheel body 90. The endless belt 92 is disposed on the outer peripheral surfaces of the pulley 91 and the wheel main body 90 and is rotated by a motor (not shown). In this case, when the endless belt 92 hits the attachment plate 31 protruding from the lower flange 30c, the rotational load of the endless belt 92 increases. When the wheel main body 90 is rotated by a motor (not shown) according to this load increase, the rolling wheel 89 rolls as indicated by a two-dot chain line, and the endless belt 92 gets over the attachment plate 31. it can. The rolling wheel 89 has a configuration similar to that of the wheel holding bracket 81 shown in FIG. 7 or another known lifting mechanism so that the wheel body 90 is held so as to be retractable upward when it rides on the attachment plate 31. 90 endless belts 92 are urged toward the upper surface of the lower flange 30c.

In the above embodiment, a twin-lens camera is used, but in addition to this, an infrared camera or the like may be further provided for inspection. In addition, a robot hand or the like may be provided on the

suspension carts

11 and 12, and an inspection process such as hitting a site where corrosion is progressing may be added. In this case, the progress of corrosion can be confirmed reliably. Furthermore, foreign matter may be removed or collected using a robot hand. In this case, the collection of foreign matter is prevented by using a collection box or a fall prevention net. Further, a suction tube may be provided at the tip of the robot hand to suck up dust and the like. Furthermore, air blowing and suction may be performed using separate nozzles, dust may be blown off by the blowing nozzle, and the blown-off dust may be collected by the suction nozzle.

The bridge 20 to be inspected according to the present invention corresponds to a bridge arranged at any position in the river, the ocean, and the land. The use of the bridge is not limited, and any use of a humanitarian bridge, a road bridge, and a railway bridge may be used. Moreover, although the steel girder 22 has been described as an example, it may be a bridge having a plurality of girders having the lower flange 30c, and the structure type and material are not particularly limited.

Claims

In a bridge inspection robot system for inspecting a bridge formed by arranging a plurality of main girders having a lower flange apart from each other in a direction perpendicular to the longitudinal direction of the main girders,
A first suspension carriage and a second suspension carriage which are attached to a plurality of the main girders and run in the longitudinal direction of the main girders;
A rail that is attached to the first suspension carriage and the second suspension carriage and that is formed longer than an interval between the main girders to which the first suspension carriage and the second suspension carriage are attached and protrudes from at least one of the suspension carriages When,
A camera that is attached to the rail and travels beyond one of the suspension carts, and
A bridge inspection robot system comprising: an image processing unit that identifies a damaged portion of the bridge based on image data of the camera.
The first suspension carriage and the second suspension carriage are:
A first wheel and a second wheel that are in contact with the upper surface of the lower flange and are spaced apart in the longitudinal direction of the main beam;
A travel position where the first wheel and the second wheel are individually in contact with the upper surface of the lower flange; and a first displacement portion and a second displacement portion which are displaced between a retracted position retracted from the upper surface of the lower flange. The bridge inspection robot system according to claim 1.
The first displacement unit and the second displacement unit have an obstacle sensor that detects an obstacle that is an obstacle to travel of the first wheel and the second wheel, and an obstacle detection signal from the obstacle sensor The bridge inspection robot system according to claim 2, wherein the first wheel or the second wheel is moved from the traveling position to the retracted position based on the above.
The offset length of the first wheel and the second wheel in the longitudinal direction of the main girder that is spaced apart from the obstacle in the longitudinal direction of the main girder in the longitudinal direction of the main girder as compared to the longitudinal length of the main girder The bridge inspection robot system according to claim 3, wherein the length of the bridge inspection robot system is long.
The first suspension carriage and the second suspension carriage are:
A first wheel and a second wheel that are in contact with the upper surface of the lower flange and are spaced apart in the longitudinal direction of the main beam;
A first displacement portion and a second displacement portion that are displaced between a travel position where the first wheel and the second wheel individually contact the upper surface of the lower flange, and a position where the first wheel and the second wheel get over the protrusion from the upper surface of the lower flange. The bridge inspection robot system according to claim 1.
The bridge inspection robot system according to any one of claims 1 to 5, wherein the camera includes a camera body, an elevating unit that moves the camera body up and down, and a camera carriage that moves the elevating unit on the rail.
The bridge inspection robot system according to claim 6, wherein the camera body is a twin-lens camera that obtains a plurality of parallax images.
The image processing unit detects a position of a member between the plurality of main girders based on a plurality of parallaxes as a distance from the camera body, and three-dimensional coordinates of the member based on a moving distance of the suspension carriage and the camera carriage The bridge inspection robot system according to claim 7, wherein a position is specified, and an imaging position of the camera body that avoids the member is specified based on the three-dimensional coordinate position.
The bridge inspection robot system according to any one of claims 1 to 8, wherein the first suspension carriage and the second suspension carriage have a lower surface contact wheel that contacts a lower surface of the lower flange.
The bridge inspection robot system according to any one of claims 1 to 9, further comprising a controller for remotely operating the first suspension carriage, the second suspension carriage, and the camera.