Monitoring system for filled steel pipe concrete columns

Abstract

To improve quality of filled steel pipe concrete columns.SOLUTION: A monitoring system 100 for filled steel pipe concrete columns comprises: an image capturing unit 20 capable of capturing images inside steel pipes; an identification unit 31 that identifies objects in an image P captured by the image capturing unit 20 based on learning results of machine learning; and a condition determination unit 32 that determines a state inside a steel pipe based on objects in the image P identified by the identification unit 31. The learning results stored in the identification unit 31 include learning results of machine learning conducted in advance using image data of the inside of the steel pipe taken when the concrete is normally filled into the steel pipe as teacher data.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a monitoring system for filled steel tube concrete columns.

Background Art

[0002] Patent Document 1 discloses a management device for a filled steel tube concrete column, which includes an imaging unit capable of imaging the top end of the concrete filled in the steel tube of the filled steel tube concrete column.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When filling the steel tube with concrete using the management device for the filled steel tube concrete column described in Patent Document 1, in order to confirm the quality of the filled steel tube concrete column, for example, whether the steel tube is filled with concrete without gaps and whether foreign substances are mixed in, the operator needs to constantly monitor the image captured by the imaging unit.

[0005] However, monitoring relatively monotonous images over a long period of time requires excessive labor even if multiple operators take turns, and since the confirmation work is based on personal visual inspection, there may be differences in the judgment of the presence or absence of abnormalities due to differences in the experience of the operators. For this reason, it is difficult to sufficiently ensure the quality of the filled steel tube concrete column.

[0006] An object of the present invention is to improve the quality of filled steel tube concrete columns.

Means for Solving the Problems

[0007] The present invention is a monitoring system for a filled steel tube concrete column, comprising an imaging unit capable of imaging the inside of the steel tube, an identification unit that stores the learning results of machine learning performed in advance and identifies an object in the image captured by the imaging unit based on the learning results, and a state determination unit that determines the state inside the steel tube based on the object in the image identified by the identification unit. The learning results stored in the identification unit include the learning results of machine learning performed in advance using the image data inside the steel tube captured when the steel tube is normally filled with concrete as teacher data.

Advantages of the Invention

[0008] According to the present invention, the quality of the filled steel tube concrete column can be improved.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, with reference to the drawings, a monitoring system for a concrete-filled steel tube column according to an embodiment of the present invention will be described.

[0011] A monitoring system 100 for a concrete-filled steel tube column according to an embodiment of the present invention is a system that monitors the filling state of fresh concrete (fresh concrete in an uncured state; hereinafter referred to as "fresh concrete") into the concrete-filled steel tube column 10 as shown in FIG. 1, and automatically determines whether the filling of the fresh concrete is being performed normally.

[0012] The concrete-filled steel tube column 10 is a column member of a CFT (Concrete Filled Steel Tube) structure in which fresh concrete is poured into a square or circular steel tube to form a column. For example, as shown in FIG. 1, it is composed of a steel tube portion 11 formed of a square steel tube and a diaphragm 12 provided between the steel tube portions 11 in the vertical direction and integrated with the steel tube portion 11 by butt welding.

[0013] The diaphragm 12 is, for example, a so-called continuous diaphragm formed of a substantially square steel plate having a side length larger than that of the steel tube portion 11, and is installed in the concrete-filled steel tube column 10 as a reinforcing member for stress transmission from the beam 15 to the column 10. A placing hole 12a through which the fresh concrete filled in the steel tube can pass is formed through the substantially central portion of each diaphragm 12, and a plurality of air vent holes 12b described later penetrate around the placing hole 12a.

[0014] Since the flange portions of the beam 15 formed of H-shaped steel members are welded to the diaphragm 12, the diaphragm 12 is arranged at a predetermined interval in the vertical direction in accordance with the interval between the pair of flange portions of the beam 15. That is, the installation interval of the diaphragm 12 provided in the filled steel tube concrete column 10 is preset according to the installation interval of the beam 15 in the vertical direction and the interval between the flange portions of the beam 15.

[0015] Note that the diaphragm 12 is not limited to a through diaphragm, and may be an inner diaphragm welded and fixed inside the steel pipe portion 11.

[0016] The upper end of the filled steel tube concrete column 10 is closed by a closing plate 14. In the closing plate 14, similar to the diaphragm 12 shown in FIG. 3, a placing hole 14a is formed penetrating substantially at the center for checking the filling condition of the fresh concrete, and a plurality of air vent holes (not shown) are formed around the placing hole 14a. The placing hole 14a is used as a through hole for suspending the imaging unit 20 described later into the steel pipe. Note that the through hole for suspending the imaging unit 20 into the steel pipe may be provided by penetrating the steel pipe portion 11 in a substantially horizontal direction.

[0017] Further, below the steel pipe portion 11 located at the lowermost end side of the filled steel tube concrete column 10, a filling port 13 for filling the fresh concrete into the steel pipe is provided. A pumping pump device 18 capable of pumping the fresh concrete or a fresh concrete supply pipe extending from a pumping pump truck (not shown) is connected to the filling port 13. When the filled steel tube concrete column 10 is high, since the fresh concrete cannot be filled up to the top of the column at once, the filling port 13 may be appropriately provided on an intermediate floor. Note that the filling port 13 provided on the intermediate floor in this way can also be used as a through hole for suspending the imaging unit 20 into the steel pipe.

[0018] Note that the filled steel tube concrete column 10 is not limited to being erected along the vertical direction as shown in FIG. 1, and may be erected with an inclination at a predetermined angle or may be inclined from the middle.

[0019] As shown in FIG. 2, a monitoring system 100 for monitoring the filling state of fresh concrete into the filled steel tube concrete column 10 with the above-described configuration mainly includes an imaging unit 20 capable of imaging the inside of the steel tube of the filled steel tube concrete column 10, an identification unit 31 for identifying an object in an image P captured by the imaging unit 20, a state determination unit 32 for determining the state inside the steel tube based on the object in the image P identified by the identification unit 31, a display unit 40 for displaying the result etc. determined by the state determination unit 32, and a communication unit 42 for transmitting the result etc. determined by the state determination unit 32 to the outside.

[0020] The imaging unit 20 is a so-called digital camera, and as shown in FIG. 1, it is suspended into the steel tube of the filled steel tube concrete column 10 through a through hole formed in the closing plate 14 via a cable 21.

[0021] As shown in FIG. 1, when fresh concrete 16 is filled into the steel tube of the filled steel tube concrete column 10, the imaging unit 20 is suspended via a cable 21 in a posture capable of imaging the region including the top end 16a of the fresh concrete 16, that is, in a state where the imaging direction faces downward, and captures an image P as shown in FIG. 3, for example.

[0022] In the image P shown in FIG. 3, at approximately the center, the top end 16a of the fresh concrete 16 imaged through the placing hole 12a of the diaphragm 12 is shown. Further, in the image P, the entire diaphragm 12 and the inner wall surfaces 11a of the four inner surfaces of the steel tube portion 11 extending upward from the four sides of the diaphragm 12 are shown. Also, a plurality of the above-described air vent holes 12b are shown around the placing hole 12a of the diaphragm 12.

[0023] The image P captured by the imaging unit 20 is transmitted to a control unit 30 described later by wireless communication or wired communication. Note that the cable 21 may be used as a transmission cable to transmit the image P through the cable 21.

[0024] The cable 21 is wound up and unwound by a reel device (not shown), and the ascending speed and descending speed of the imaging unit 20 are controlled by the reel device. Note that the control of the reel device may be performed by the control unit 30 described later, or may be performed by an operator. Further, a laser distance meter (not shown) capable of measuring the distance between the imaging unit 20 and the top end 16a of the raw pipe 16 may be suspended together with the imaging unit 20 via the cable 21, and the reel device may be controlled so that the distance measured by this laser distance meter becomes a constant value.

[0025] In addition, in order to illuminate the inside of the steel pipe into which light is not inserted, an illumination device such as an LED light (not shown) is suspended via the cable 21 together with the imaging unit 20 in a posture capable of irradiating light toward the top end 16a of the raw pipe 16.

[0026] The identification unit 31 and the state determination unit 32 are shown as virtual units of a part of the functions of a general personal computer used as the control unit 30, and do not physically exist.

[0027] The control unit 30 having the identification unit 31 and the state determination unit 32 as part of its functions is composed of a microcomputer including a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and an I / O interface (Input / Output interface). The RAM stores data in the processing of the CPU, the ROM stores in advance the control program of the CPU, etc., and the I / O interface is used for input / output of information with devices connected to the control unit 30. Note that the control unit 30 may be composed of a plurality of microcomputers. Further, it is preferable that the control unit 30 includes a GPU (Graphics Processing Unit) specialized for image processing.

[0028] The learning results of machine learning performed in advance are stored in the identification unit 31, and the identification unit 31 identifies what the object shown in the image P captured by the imaging unit 20 and transmitted to the control unit 30 is based on the stored learning results.

[0029] The learning results stored in the identification unit 31 include the learning results of machine learning previously performed using a large number of image data inside the steel pipe captured when the fresh concrete 16 is normally filled into the steel pipe as teacher data. In addition, there are the learning results of machine learning previously performed using the image data in which the top end 16a of the fresh concrete 16 has abnormally started to condense or has condensed when the fresh concrete 16 is filled, the learning results of machine learning previously performed using the image data of the inner wall surface 11a of the steel pipe portions 11 of various shapes into which the fresh concrete 16 is filled as teacher data, the learning results of machine learning previously performed using the image data of the diaphragms 12 of various shapes installed in the filled steel pipe concrete column 10 as teacher data, the learning results of machine learning previously performed using the image data of various foreign objects F1 to F3 mixed into the steel pipe as teacher data, and the like.

[0030] As a machine learning method, for example, a method such as semantic segmentation using deep learning is used. This is a type of segmentation that identifies the type of object displayed in each pixel for each pixel and detects the pixel region where the object to be identified exists. In order to be able to identify, as the types of objects, mainly the fresh concrete 16, the diaphragm 12, the inner wall surface 11a of the steel pipe portion 11, and unknown objects not corresponding to these from the image data, learning is performed to detect the pixel region where the fresh concrete 16 exists, the pixel region where the diaphragm 12 exists, the pixel region where the inner wall surface 11a of the steel pipe portion 11 exists, and the pixel region where unknown objects and foreign objects F1 to F3 exist.

[0031] The machine learning method is not limited to semantic segmentation, and any method can be used as long as it enables the identification of the type of object and region classification in the image data. For example, other known methods using deep learning such as instance segmentation or panoptic segmentation may be used.

[0032] As teacher data in machine learning, various image data are used, such as the image P inside the steel pipe captured when the fresh concrete 16 is normally filled inside the steel pipe as shown in FIGS. 3 to 7, the image P when the shape of the inner wall surface 11a of the steel pipe portion 11 and the shape of the diaphragm 12 are different from those shown in FIG. 3 as shown in FIG. 8, and the image P when foreign objects F1 to F3 exist inside the steel pipe as shown in FIG. 9. Note that these are examples, and a large number of image data corresponding to these are used as teacher data.

[0033] Specifically, FIG. 3 shows the state before the top end 16a of the fresh concrete 16 passes through the placing hole 12a of the diaphragm 12, and FIG. 4 shows the state where the top end 16a of the fresh concrete 16 passes through the placing hole 12a of the diaphragm 12, and a first boundary line B1, which is the boundary between the fresh concrete 16 that has passed through the placing hole 12a and the diaphragm 12, is formed.

[0034] Further, FIG. 5 shows the state where the fresh concrete 16 passes through a plurality of air vent holes 12b in addition to the placing hole 12a, and a second boundary line B2, which is the boundary between the fresh concrete 16 that has passed through the air vent holes 12b and the diaphragm 12, is formed. FIG. 6 shows the state where the fresh concrete 16 that has passed through the placing hole 12a and the fresh concrete 16 that has passed through the plurality of air vent holes 12b merge, the first boundary line B1 and the second boundary line B2 overlap, a part of the fresh concrete 16 reaches the inner wall surface 11a of the steel pipe portion 11, and a third boundary line B3, which is the boundary between the fresh concrete 16 and the inner wall surface 11a, is formed.

[0035] Further, FIG. 7 shows the state where the first boundary line B1 and the second boundary line B2 completely overlap, the fresh concrete 16 reaches the inner wall surface 11a over the entire circumference of the steel pipe portion 11, and only the third boundary line B3 exists.

[0036] From a series of images P captured when the fresh concrete 16 is normally filled in the steel pipe as shown in FIGS. 3 to 7, the positional relationship among the fresh concrete 16, the diaphragm 12, and the inner wall surface 11a of the steel pipe portion 11 is learned. Also, when the fresh concrete 16 passes through the diaphragm 12, how the fresh concrete 16 appears from the placing holes 12a and the air vent holes 12b in what order is learned. Further, the positions of the boundaries (the first boundary line B1 and the second boundary line B2) between the fresh concrete 16 and the diaphragm 12 and the position of the boundary (the third boundary line B3) between the fresh concrete 16 and the inner wall surface 11a during normal times are learned.

[0037] In addition, FIG. 8 shows an example of a filled steel pipe concrete column 10 in which the steel pipe portion 11 is a circular steel pipe instead of a square steel pipe, the placing holes 12a of the diaphragm 12 are octagonal (polygonal) instead of circular, and the air vent holes 12b are provided at eight locations instead of four locations.

[0038] From the images P such as those shown in FIG. 8 where the shape of the inner wall surface 11a of the steel pipe portion 11 is different or the shape of the diaphragm 12 is different, the general positional relationship between the placing holes 12a and the air vent holes 12b of the diaphragm 12 and the general positional relationship between the diaphragm 12 and the inner wall surface 11a are learned. By learning in advance the general positional relationship of the internal structure of the filled steel pipe concrete column 10, it becomes possible to perform the determination by the state determination unit 32 described later for filled steel pipe concrete columns 10 having various cross-sectional shapes.

[0039] Also, FIG. 9 shows an example in which there is accumulated water such as rainwater or dew condensation as the foreign object F1, a plastic bottle as the foreign object F2, and a tool as the foreign object F3 in the steel pipe.

[0040] From the image P in which foreign objects F1 to F3 are present in the steel pipe as shown in FIG. 9, various foreign objects are learned. In the image P shown in FIG. 9, a plurality of foreign objects F1 to F3 are shown together, but any one of the foreign objects F1 to F3 may be shown. Further, the foreign objects are not limited to the above-described foreign objects F1 to F3, and may be building materials such as wire ropes and reinforcing bars, or living organisms such as work gloves and insects or small animals that have entered the steel pipe.

[0041] When identifying what the object shown in the image P captured by the imaging unit 20 is by the identification unit 31 in which the above-described learning result is stored, in order to improve the identification accuracy of the object with respect to the image P, for example, known image processing such as a binarization process or a process of three-dimensionally processing image data using SfM (Structure from Motion) may be performed in advance.

[0042] The state determination unit 32 determines the state of the green concrete 16 filled in the steel pipe and the filling speed of the green concrete 16 filled in the steel pipe based on the objects in the image P identified by the identification unit 31, mainly the green concrete 16, the diaphragm 12, and the inner wall surface 11a of the steel pipe portion 11, and determines whether the green concrete 16 is normally filled in the steel pipe. Further, the state determination unit 32 determines the presence or absence of an abnormality in the steel pipe based on whether the object in the image P identified by the identification unit 31 is, for example, a foreign object F1 to F3 or an unknown object.

[0043] Specifically, when the state determination unit 32 identifies the green concrete 16 and the inner wall surface 11a of the steel pipe portion 11 as objects by the identification unit 31 and also identifies the boundary (third boundary line B3) between the inner wall surface 11a and the green concrete 16, for example, when there is a difference of a predetermined value or more in the positions of the third boundary lines B3 formed on each inner wall surface 11a and they are not uniform, that is, when the top end 16a of the green concrete 16 does not spread evenly toward each inner wall surface 11a but spreads unevenly toward any one of the inner wall surfaces 11a, it is determined that there is an abnormality in the green concrete 16 because there is a variation in the properties of the green concrete 16 and local hardening may be progressing.

[0044] Also, when the shape of the third boundary line B3 formed on each inner wall surface 11a is an arc shape with a convex central portion and its curvature is larger than a predetermined reference value, that is, when it is assumed that the bulge of the top end 16a of the fresh concrete 16 is relatively large near the center of the filled steel pipe concrete column 10 (near the center of the placing hole 12a), it is also determined that there is an abnormality in the fresh concrete 16 because there are variations in the properties of the fresh concrete 16 and there is a possibility that local hardening is progressing.

[0045] Further, when the state determination unit 32 identifies the fresh concrete 16 and the diaphragm 12 as objects by the identification unit 31, and also identifies the boundary (the first boundary line B1) between the fresh concrete 16 that has passed through the placing hole 12a and the diaphragm 12 and the boundary (the second boundary line B2) between the fresh concrete 16 that has passed through the air vent hole 12b and the diaphragm 12, if there is a flow of fresh concrete 16 that deviates from the normal appearance order of the fresh concrete 16, it is determined that there is an abnormality in the fresh concrete 16 because there are variations in the properties of the fresh concrete 16 and there is a possibility that voids where the fresh concrete 16 is not filled locally are generated.

[0046] Specifically, when the fresh concrete 16 normally passes through the diaphragm 12, first, the fresh concrete 16 appears from the central placing hole 12a and spreads on the diaphragm 12 in a substantially circular shape (see FIG. 4). Then, before the fresh concrete 16 (the first boundary line B1) that has appeared from the placing hole 12a reaches each air vent hole 12b, the fresh concrete 16 appears from each air vent hole 12b (see FIG. 5). After that, the fresh concrete 16 (the first boundary line B1) that has appeared from the placing hole 12a merges with the fresh concrete 16 (the second boundary line B2) that has appeared from each air vent hole 12b, and finally reaches the inner wall surface 11a of the steel pipe portion 11 (see FIGS. 6 and 7).

[0047] On the other hand, for example, if the first boundary line B1 reaches the air vent hole 12b before the second boundary line B2 is identified in any of the air vent holes 12b, or if there is a variation exceeding a predetermined allowable range in the spread of the second boundary line B2 of the raw material 16 emerging from each air vent hole 12b, it is determined that there is an abnormality in the raw material 16 on the assumption that there is a flow of the raw material 16 deviating from the normal emergence order of the raw material 16.

[0048] Also, when the first boundary line B1 does not spread evenly toward each air vent hole 12b and there is a variation exceeding a predetermined allowable range at the time when the first boundary line B1 overlaps with each second boundary line B2, it is also determined that there is an abnormality in the raw material 16 on the assumption that there is a variation in the properties of the raw material 16 and there may be a gap where the raw material 16 is not filled locally.

[0049] In making such a determination, the state determination unit 32 stores threshold values such as reference values and allowable ranges for determining the presence or absence of an abnormality in the raw material 16 based on the positions and shapes of the boundary lines B1, B2, and B3. The magnitudes of these threshold values may be updated at any time according to the learning results stored in the identification unit 31.

[0050] Note that, as teacher data in the above machine learning, an image P inside the steel pipe captured when the raw material 16 is not normally filled in the steel pipe is used, and the learning result when the raw material 16 is not normally filled is stored in advance in the identification unit 31. Based on whether the positional relationship of the objects in the image P identified by the identification unit 31 is closer to the learning result when the raw material 16 is normally filled or the learning result when the raw material 16 is not normally filled, the presence or absence of an abnormality in the raw material 16 may be determined.

[0051] Further, when the state determination unit 32 determines that the fresh concrete 16 and the diaphragm 12 are identified as objects by the identification unit 31 and the boundary (first boundary line B1) between the fresh concrete 16 that has passed through the placement hole 12a and the diaphragm 12 is identified, based on the time elapsed from when the first boundary line B1 is identified by the identification unit 31 until the fresh concrete 16 passes through the placement hole 12a of another diaphragm 12 disposed above the diaphragm 12, that is, until the first boundary line B1 is identified by the identification unit 31 on another diaphragm 12, it determines the filling rate of the fresh concrete 16 filled in the steel pipe. Since the installation interval of the diaphragms 12 in the vertical direction is determined in advance as described above, the filling rate of the fresh concrete 16 can be easily obtained by measuring the time required for the fresh concrete 16 to pass through the placement holes 12a of the two diaphragms 12 arranged in the vertical direction.

[0052] When the filling rate of the fresh concrete 16 obtained in this way exceeds a preset reference rate (for example, 1 m / min), it is determined that there is an abnormality in the filling rate of the fresh concrete 16, assuming that there may be a gap where the fresh concrete 16 is not filled on the lower surface of the diaphragm 12 due to the high filling rate. Also, when the filling rate of the fresh concrete 16 is less than the reference rate, it is determined that there is an abnormality in the filling rate of the fresh concrete 16, assuming that local hardening may be progressing.

[0053] The state determination unit 32 stores threshold values such as reference values and allowable ranges for determining the presence or absence of an abnormality in the filling rate of the fresh concrete 16. The magnitudes of these threshold values may be updated at any time according to the learning results stored in the identification unit 31.

[0054] In addition, when the foreign objects F1 to F3 are identified as objects by the identification unit 31, for example, when the foreign object F1 is a liquid such as standing water, the water-cement ratio of the fresh concrete 16 may change and the strength of the hardened concrete may decrease. Thus, it is determined that there is an abnormality in the steel pipe and the state inside the steel pipe is not suitable for filling the fresh concrete 16. Also, when the foreign objects F2 and F3 are solids such as plastic bottles or tools, it is assumed that the strength of the filled steel pipe concrete column 10 after hardening may decrease due to cross-sectional defects. Thus, it is determined that there is an abnormality in the steel pipe and the state inside the steel pipe is not suitable for filling the fresh concrete 16.

[0055] Also, when an unknown object that cannot be identified using the learning results stored in the identification unit 31 is identified as an object by the identification unit 31, the state determination unit 32 determines that there may be some problem and that there is an abnormality in the steel pipe and the state inside the steel pipe is not suitable for filling the fresh concrete 16.

[0056] The determination result determined by the state determination unit 32, the identification result by the identification unit 31, the image P inside the steel pipe captured by the imaging unit 20, etc. are displayed on the display unit 40 and transmitted to the outside through the communication unit 42.

[0057] The display unit 40 is a monitor device installed at the work site. The worker checks the determination result displayed on the display unit 40 and considers countermeasures. Note that a plurality of display units 40 may be provided. For example, they may be installed near the operation panel for operating the reel device that unwinds the cable 21 or near the operation panel for operating the pumping pump device 18. Also, the display unit 40 may be provided with a speaker capable of notifying an abnormality by sound or a rotating light capable of notifying an abnormality by light to quickly notify the worker of the abnormality.

[0058] The communication unit 42 is a general wired or wireless communication device capable of transmitting data via the Internet line, and transmits the determination result etc. determined by the state determination unit 32 to a tablet terminal or computer held by a supervisor who monitors the work situation at a remote location away from the work site.

[0059] Accordingly, not only at the work site where the filling operation of the fresh concrete 16 is performed, but also at a remote location far from the work site, it is possible to quickly consider a countermeasure method based on the determination result determined by the state determination unit 32.

[0060] Note that the identification unit 31 and the state determination unit 32 do not necessarily need to be provided within one control unit 30, that is, within one personal computer. For example, they may be provided in different servers respectively and configured to transmit and receive data wired or wirelessly.

[0061] Subsequently, a monitoring method performed using the monitoring system 100 having the above configuration will be described with reference to the flowchart of FIG. 10.

[0062] First, before filling the fresh concrete 16, in step S11, determination of the internal state of the filled steel pipe concrete column 10 without the fresh concrete 16 is started.

[0063] The determination before filling is started in subsequent step S12 by gradually lowering the imaging unit 20 into the steel pipe of the filled steel pipe concrete column 10, and is continued while the imaging unit 20 is being lowered.

[0064] Specifically, the imaging unit 20 is inserted into the steel pipe through the through hole provided in the closing plate 14, and the cable 21 is unwound from the reel device at a relatively slow predetermined speed (for example, 1 m / min), so that the imaging unit 20 is lowered to a predetermined position, for example, a position higher than the filling port 13 where the fresh concrete 16 is filled by a predetermined height.

[0065] While the imaging unit 20 is descending inside the steel pipe, the identification unit 31 identifies the object in the image P captured by the imaging unit 20 based on the learning result, and the state determination unit 32 determines the state inside the steel pipe based on the object in the image P identified by the identification unit 31 (step S13).

[0066] Since it is before the fresh concrete 16 is filled, the objects identified by the identification unit 31 are the diaphragm 12, the inner wall surface 11a of the steel pipe portion 11, foreign matters F1 to F3, and unknown objects that cannot be identified using the learning results.

[0067] Therefore, in step S13, when foreign matters F1 to F3 or an unknown object is identified by the identification unit 31, it is determined that there is an abnormality in the steel pipe and the state inside the steel pipe is in a state where the fresh concrete 16 cannot be filled. Also, when the placing holes 12a and the air vent holes 12b provided in the diaphragm 12 are not identified, it is determined that there is an abnormality in the arrangement relationship between the steel pipe portion 11 and the diaphragm 12, that is, there is an abnormality in the fitting of the steel frame, and the state inside the steel pipe is in a state where the fresh concrete 16 cannot be filled, and the process proceeds to step S21.

[0068] When it is not determined that there is an abnormality in the steel pipe, the lowering of the imaging unit 20 is continued until the imaging unit 20 reaches a predetermined position, and the identification by the identification unit 31 and the determination by the state determination unit 32 are continued (step S13 and step S14).

[0069] In step S14, when it is determined that the imaging unit 20 has reached a predetermined position, the process proceeds to step S15, and the determination of the state inside the filled steel pipe concrete column 10 filled with the fresh concrete 16 is started.

[0070] Note that if there is time until the filling of the fresh concrete 16 starts, the identification by the identification unit 31 and the determination by the state determination unit 32 may be temporarily stopped. Also, before the filling of the fresh concrete 16 starts, the imaging unit 20 may be reciprocated a plurality of times between a predetermined position and the closing plate 14 to further determine the presence or absence of abnormalities inside the filled steel pipe concrete column 10 in which the fresh concrete 16 is not filled.

[0071] The determination during the filling of the fresh concrete 16 starts in the subsequent step S16 when the filling of the fresh concrete 16 into the filled steel pipe concrete column 10 starts and the imaging unit 20 is gradually raised from a predetermined position, and continues while the imaging unit 20 is being raised.

[0072] Specifically, in step S16, according to the delivery speed of the fresh concrete 16 delivered from the pressure delivery pump device 18 into the filled steel pipe concrete column 10, the cable 21 is rewound at a predetermined speed by the reel device, thereby raising the imaging unit 20 toward the closing plate 14.

[0073] While the imaging unit 20 is ascending within the steel pipe, the identification unit 31 identifies the objects in the image P captured by the imaging unit 20 based on the learning result, and the state determination unit 32 determines the state inside the steel pipe based on the objects in the image P identified by the identification unit 31 (steps S17 to S19).

[0074] Specifically, in step S17, as described above, based on the boundary lines B1, B2, and B3 between the objects identified by the identification unit 31, the state determination unit 32 determines whether there is an abnormality in the filling state, that is, whether there is a problem with the properties of the fresh concrete 16 filled into the steel pipe.

[0075] And, as described above, if there is a possibility that the properties of the fresh concrete 16 vary, it is determined by the state determination unit 32 that there is an abnormality in the fresh concrete 16, and the process proceeds to step S21.

[0076] Also, in step S18, the presence or absence of foreign objects inside the steel pipe is determined. As described above, if the objects identified by the identification unit 31 are foreign objects F1 to F3 or unknown objects, it is determined by the state determination unit 32 that there is an abnormality inside the steel pipe, and the process proceeds to step S21.

[0077] Also, in step S19, the presence or absence of an abnormality in the filling speed of the fresh concrete 16 is determined. For example, if there is a difference greater than a predetermined value between the calculated filling speed and the reference speed, as described above, it is determined by the state determination unit 32 that there is an abnormality in the filling speed of the fresh concrete 16, and the process proceeds to step S21.

[0078] Note that the determination by the state determination unit 32 performed from step S17 to step S19 does not need to be performed in the predetermined order as described above. These determinations may be performed in parallel and simultaneously, or may be performed in any order. As long as the control flow proceeds to step S21 when an abnormality is determined in any of the determinations from step S17 to step S19, it is sufficient.

[0079] When it is not determined that there is an abnormality in the steel pipe, the imaging unit 20 continues to rise until it reaches a predetermined position near the upper end of the filled steel pipe concrete column 10, and the identification by the identification unit 31 and the determination by the state determination unit 32 are continued (steps S17 to S20).

[0080] In step S20, when it is determined that the imaging unit 20 has reached the predetermined position, the monitoring of the inside of the steel pipe performed by the monitoring system 100 ends.

[0081] In step S21, various abnormalities determined by the state determination unit 32 are notified through the display unit 40 and the communication unit 42. For example, in step S13 or step S18, when foreign objects F1 to F3 or unknown objects are identified as objects, the locations where the foreign objects F1 to F3 or unknown objects appear on the display unit 40 are highlighted more than other locations by blinking or the like. In step S17, when it is determined that the fresh concrete 16 is not filled normally, on the display unit 40, the location where there is a flow of fresh concrete 16 deviating from the normal appearance order of the fresh concrete 16 or the location where the fresh concrete 16 is unevenly spread with respect to the inner wall surface 11a is highlighted more than other locations by blinking or the like. Further, in step S19, when it is determined that there is an abnormality in the filling speed of the fresh concrete 16, on the display unit 40, the calculated filling speed is highlighted by blinking or the like, and how much it deviates from the allowable range of the filling speed is visually displayed by a graph or the like. Note that as a specific method of notifying the abnormality, a pop-up notifying the abnormality may be displayed on the screen of the display unit 40, or the operator or supervisor may be notified by a short message or the like through the communication unit 42.

[0082] Thereafter, countermeasures corresponding to the notified abnormality are considered by the operator or the supervisor. For example, if it is a minor abnormality before filling the green concrete 16, the operation is performed to return to step S14 in order to continue the monitoring by the monitoring system 100. If it is a minor abnormality during filling of the green concrete 16, the operation is performed to return to step S20 in order to continue the monitoring by the monitoring system 100.

[0083] On the other hand, if it is a serious abnormality such that the filling of the green concrete 16 should not be started or a serious abnormality such that the filling of the green concrete 16 should be interrupted, the operation is performed to temporarily end the monitoring performed by the monitoring system 100.

[0084] In this way, by the monitoring inside the steel pipe performed by the monitoring system 100, the presence or absence of an abnormality inside the steel pipe before filling the green concrete 16 and the presence or absence of an abnormality inside the steel pipe during filling of the green concrete 16 are automatically determined and notified.

[0085] According to the above embodiment, the following effects are obtained.

[0086] According to the monitoring system 100 having the above configuration, an object in the image P inside the steel pipe captured by the imaging unit 20 is identified by the identification unit 31 in which the learning result of machine learning is stored. Based on the identified object in the image P, the state inside the steel pipe into which the green concrete 16 is filled, that is, the presence or absence of an abnormality inside the steel pipe is determined by the state determination unit 32.

[0087] In this way, the presence or absence of abnormalities when the fresh concrete 16 is filled into the steel pipe and the presence or absence of abnormalities inside the steel pipe before the fresh concrete 16 is filled are automatically determined based on the image P captured by the imaging unit 20. As a result, it is no longer necessary for an operator or a supervisor to continuously monitor the image P displayed on a display unit 40 such as a monitor. Thereby, the labor in the monitoring work can be significantly reduced, the work load of the operator and the supervisor can be reduced, and since it is not necessary to rely on the experience of the operator and the supervisor by using the learning result of machine learning in the determination of the presence or absence of abnormalities, the quality of the filled steel pipe concrete column 10 can be stably improved.

[0088] In addition, the determination result and the like determined by the state determination unit 32 are automatically notified to a supervisor who monitors the work situation at a remote location away from the work site. Therefore, it is possible to quickly consider a countermeasure method for the abnormality.

[0089] Note that the following modification examples are also within the scope of the present invention, and it is also possible to combine the configurations shown in the modification examples with the configurations described in the above-described embodiments, or to combine the configurations described in the following different modification examples with each other.

[0090] In the above embodiment, the inside of the filled steel pipe concrete column 10 before the fresh concrete 16 is filled is hollow, but the filled steel pipe concrete column 10 may be a steel pipe with built-in reinforcing bars in which a plurality of main reinforcing bars are arranged inside.

[0091] In addition, in the above embodiment, a digital camera is used as the imaging unit 20. However, in order to improve the detection accuracy of an object inside the steel pipe, the imaging unit 20 may be provided with a distance measuring sensor that uses laser light or an infrared sensor, such as a 3D-ToF sensor, that can widely measure the distance or depth to an object inside the steel pipe, in addition to or instead of this.

[0092] In addition, in order to clarify the reference position of the image P captured by the imaging unit 20, characters, symbols, coloring, etc. indicating the reference position may be provided on the inner wall surface 11a of the diaphragm 12 or the steel pipe portion 11.

[0093] In the above embodiment, the filling rate of the fresh concrete 16 is calculated based on the time from when the fresh concrete 16 passes through the placing hole 12a until it passes through the placing hole 12a of another diaphragm 12. However, the method for obtaining the filling rate of the fresh concrete 16 is not limited to this. The filling rate of the fresh concrete 16 may be obtained, for example, by measuring the rising speed of the top end 16a of the fresh concrete 16 with a distance measuring sensor using laser light or the like, or by using a self-position estimation method based on VisualSLAM (Simultaneous Localization and Mapping) and obtaining it from the pulling-up speed of the imaging unit 20.

[0094] In the above embodiment, the case where only the diaphragm 12 is provided in the filled steel pipe concrete column 10 has been described. However, when the filled steel pipe concrete column 10 is, for example, a welded fabricated angle steel pipe column assembled by welding four flat plates, in the filled steel pipe concrete column 10, a shape retaining plate provided to make the angle between adjacent flat plates a right angle may be appropriately arranged in addition to the diaphragm 12. Further, the shape retaining plate may be appropriately provided in the filled steel pipe concrete column 10 in order to prevent deformation during transportation or construction of the filled steel pipe concrete column 10. Thus, the shape retaining plate provided in the filled steel pipe concrete column 10 together with the diaphragm 12 is provided with a placing hole formed to penetrate substantially in the center and a plurality of through holes or notches formed around the placing hole for air bleeding, similar to the diaphragm 12. Therefore, it is possible to determine whether the fresh concrete 16 has passed through the shape retaining plate normally in the same manner as when the fresh concrete 16 passes through the diaphragm 12. Note that the learning result stored in the identification unit 31 may include the learning result of machine learning previously performed using the image data of the shape retaining plates of various shapes installed in the filled steel pipe concrete column 10 as teacher data.

[0095] The embodiments of the present invention have been described above. However, the above embodiments merely show some application examples of the present invention, and are not intended to limit the technical scope of the present invention to the specific configurations of the above embodiments.

Explanation of Reference Numerals

[0096] 100 ··· Monitoring system 10 ··· Filled steel tube concrete column 11 ··· Steel tube part 11a ··· Inner wall surface 12 ··· Diaphragm 12a ··· Placing hole 12b ··· Air vent hole 16 ··· Fresh concrete 16a ··· Top end 20 ··· Imaging unit 31 ··· Identification unit 32 ··· State determination unit

Claims

1. A monitoring system for a filled steel pipe concrete column, an imaging unit capable of imaging the inside of the steel pipe, an identification unit in which the learning results of machine learning performed in advance are stored, and which identifies an object in the image captured by the imaging unit based on the learning results, and a state determination unit that determines the state inside the steel pipe based on the object in the image identified by the identification unit. The learning results stored in the identification unit include the learning results of machine learning performed in advance using the image data inside the steel pipe captured when the steel pipe is normally filled with concrete as teacher data. Monitoring system.

2. The learning results stored in the identification unit include the learning results of machine learning performed in advance using the image data of the inner wall surface of the steel pipe into which the concrete is filled as teacher data, the identification unit identifies the inner wall surface of the steel pipe and the concrete filled in the steel pipe as the objects from the image in which the region including the top end of the concrete is imaged, and the state determination unit determines the state of the concrete filled in the steel pipe based on the boundary line between the inner wall surface and the concrete identified by the identification unit. The monitoring system according to Claim 1.

3. The learning results stored in the identification unit include the learning results of machine learning performed in advance using the image data of the diaphragm installed in the filled steel pipe concrete column as a reinforcing member as teacher data, the identification unit identifies the diaphragm and the concrete passing through the placing holes and air vent holes formed in the diaphragm as the objects from the image in which the region including the top end of the concrete is imaged, The state determination unit determines the state of the concrete filled in the steel pipe based on the boundary line between the concrete that has passed through the placement hole identified by the identification unit and the diaphragm, and the boundary line between the concrete that has passed through the air vent hole identified by the identification unit and the diaphragm. The monitoring system according to claim 1.

4. The learning result stored in the identification unit includes the learning result of machine learning previously performed using the image data of the diaphragm installed in the filled steel pipe concrete column as teacher data as a reinforcing member. The identification unit identifies the diaphragm and the concrete passing through the placement hole formed in the diaphragm as the object from the image in which the region including the top end of the concrete is imaged. The state determination unit determines the filling rate of the concrete filled in the steel pipe based on the elapsed time from when it is identified by the identification unit that the concrete passes through the placement hole of the diaphragm until it is identified by the identification unit that the concrete passes through the placement hole of another diaphragm disposed above the diaphragm. The monitoring system according to claim 1.

5. The learning result stored in the identification unit includes the learning result of machine learning previously performed using the image data of foreign matter mixed in the steel pipe as teacher data. When the foreign matter is identified as the object by the identification unit from the image in which the region including the top end of the concrete is imaged, the state determination unit determines that there is an abnormality in the steel pipe in which the concrete is filled. The monitoring system according to claim 1.

6. When the identification unit identifies an unknown object that cannot be identified using the learning result stored in the identification unit from the image in which the area including the top end of the concrete is imaged as the object, the state determination unit determines that there is an abnormality in the steel pipe filled with the concrete. The monitoring system according to any one of claims 1 to 5.

7. The learning result stored in the identification unit includes the learning result of machine learning previously performed using the image data of foreign matters mixed in the steel pipe as teacher data. When the identification unit identifies the foreign matter as the object from the image in the steel pipe imaged before the concrete is filled, or when an unknown object that cannot be identified using the learning result stored in the identification unit is identified as the object, the state determination unit determines that the state in the steel pipe is a state in which the concrete cannot be filled. The monitoring system according to claim 1.

Images