Human detection system for crane operations

The person detection system addresses the limitations of human visual inspection by using automated zoom and helmet color differentiation to reliably detect workers in dangerous crane zones, enhancing safety through precise monitoring and alerts.

JP7591385B2Active Publication Date: 2024-11-28TODA CORP
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
JP2020194176
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-11-24
Publication Date
2024-11-28
Estimated Expiration
2040-11-24

AI Technical Summary

Technical Problem

Existing crane safety systems rely on human visual inspection of distant images, which are prone to errors and limited by image quality, making it difficult to accurately detect workers in dangerous areas under suspended loads.

Method used

A person detection system with a camera, rangefinder, camera control unit, and image processing unit that automatically adjusts zoom based on distance, distinguishes worker helmets, and issues alarms for non-involved workers, ensuring consistent display of a dangerous area on a monitor.

Benefits of technology

Enhances crane safety by accurately detecting workers in dangerous areas, reducing human error, and preventing accidents through automated alerts.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007591385000002
    Figure 0007591385000002
  • Figure 0007591385000003
    Figure 0007591385000003
  • Figure 0007591385000004
    Figure 0007591385000004
Patent Text Reader

Abstract

To provide a human-detection system for crane works which prevents crane accidents by detecting people in a dangerous area under a suspended load in crane works.SOLUTION: A human-detection system consists of a human-detection camera 5 to take a picture of a lower part of a suspended load 4 attached to a crane 2, a distance meter 7 attached near the human-detection camera 5 and to measure the distance to a working surface 6, a camera control part 8 to control the human-detection camera 5, and a human-detection monitor 9 installed in a driver's seat and to display the image taken by the human-detection camera 5. The camera control part 8 is equipped with an automatic zooming function automatically zooming in conjunction with the distance measured by the distance meter 7 so that the dangerous area 10 corresponding to the area on the working surface 6 centered on the suspended load is displayed in a fixed size on the human-detection monitor 9.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a method for operating a crane. Lower area of ​​the load The present invention relates to a system for detecting a person present within a danger zone. [Background technology]

[0002] The load suspended by the crane lower area It is extremely dangerous for workers to enter these areas. For this reason, when performing crane work, a supervisor or signal person is stationed to ensure the safety of workers on the work surface, and the operator operates the crane according to their signals.

[0003] In order to improve safety in crane operations, measures have been taken in the past, such as those disclosed in Patent Documents 1 and 2 below, in which the area below the suspended load is photographed with a camera fixed to the tip of the crane's boom, the image taken by the camera is displayed on a monitor installed in the operator's room, and an alarm signal is sounded if, for example, a worker enters a danger area below the suspended load. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2010-241548 A [Patent Document 2] JP 2020-7134 A Summary of the Invention [Problem to be solved by the invention]

[0005] The images taken by the camera fixed to the tip of the boom are also used to check the lifting center during crane operations, so they are usually long-distance images. When there is something to check, the image can be zoomed in manually (by pressing a button) as appropriate to switch to a close-up image.

[0006] Since the images displayed on the monitor are checked by the operator's eyes, human error such as oversight is likely to occur. Also, since the images are often taken from a long distance, there is a limit in terms of image quality to the extent to which a person or computer can recognize whether or not a worker is present in the captured image.

[0007] Therefore, a main object of the present invention is to contribute to preventing crane accidents by detecting a person present in a danger area below a suspended load during crane operation. [Means for solving the problem]

[0008] In order to solve the above problems, the present invention relates to a crane. Lower area of ​​the load a distance meter attached near the person detection camera and measuring the distance to a work surface; a camera control unit that controls the person detection camera; and a person detection monitor that is provided in the crane driver's seat and displays an image captured by the person detection camera. an image processing unit that processes the image captured by the human detection camera, determines whether or not a worker is present in the danger zone, and issues an alarm signal if a worker is present; It consists of: the image processing unit adds a danger area to the image data transmitted from the human detection camera, and transmits the image data to which the danger area has been added to the human detection monitor; The camera control unit has an auto-zoom function that automatically zooms in conjunction with the distance measured by the rangefinder so that a danger area, which is an area on the work surface centered on the suspended load, is displayed at a constant size on the human detection monitor. of Be prepared And, The colors of the helmets of the workers involved in the lifting work are differentiated from the colors of the helmets of the other workers, and in the image processing unit, when only the helmets of the workers involved in the lifting work are detected in the danger zone, no alarm is issued, but when the helmets of the workers other than the workers involved in the lifting work are detected in the danger zone, an alarm signal is issued. The present invention provides a human detection system for crane operation.

[0009] In the invention described in claim 1, a person detection camera and a range finder are attached to the crane, and an image captured by the person detection camera is displayed on a person detection monitor installed in the crane driver's seat. At this time, the camera control unit that controls the person detection camera is provided with an auto-zoom function that automatically zooms in conjunction with the distance measured by the range finder so that a danger area, which is an area on the work surface centered on a preset suspended load, is displayed at a constant size on the person detection monitor. In this way, even if the distance to the work surface changes due to crane work, the danger area is always displayed at a constant size on the person detection monitor, making it easier to detect people on the work surface, and work actions that may cause accidents can be detected in advance, which contributes to the prevention of crane accidents.

[0010] In the present invention, an image processing unit is provided which processes the image captured by the human detection camera, determines whether a worker is present in the danger zone, and issues a warning signal if a worker is present, and the image processing unit adds a danger zone to the image data transmitted from the human detection camera, and transmits the image data with the danger zone added to the human detection monitor. The colors of the helmets of the workers involved in the hoisting work are differentiated from the colors of the helmets of the other workers, and the image processing unit does not issue a warning if only the helmets of the workers involved in the hoisting work are detected in the danger zone, but issues a warning signal if the helmets of workers other than those involved in the hoisting work are detected in the danger zone. Therefore, This will prevent crane operator oversights and more reliably prevent crane accidents.

[0011] As the present invention relating to claim 2, there is provided a person detection system for crane operation as described in claim 1, in which a camera for photographing a suspended load is provided in a position where the crane's suspension core can be photographed, and a monitor for displaying a suspended load is provided in the crane driver's seat, which displays the image photographed by the camera for photographing the suspended load.

[0012] In the invention described in claim 2 above, a camera for photographing a suspended load is provided in addition to the person detection camera, making it easy to check the suspension center during crane operation.

[0013] Claim 3 In the present invention, the size of the danger area displayed on the human detection monitor is Depending on the length of the load Claim 1 that can be set arbitrarily 、2 Any one of the above-described human detection systems for crane operation is provided.

[0014] The above claims 3 In the described invention, the size of the danger area displayed on the human detection monitor is Depending on the length of the loadThis allows the user to set the range as desired, making it possible to respond to situations such as when the user wants to check the inside of the danger zone in more detail, or when the user wants to check a wider area around the danger zone. Effect of the Invention

[0015] INDUSTRIAL APPLICABILITY As described above in detail, the present invention makes it possible to detect a person present in a danger area below a suspended load during crane operation, thereby contributing to the prevention of crane accidents. [Brief description of the drawings]

[0016] [Figure 1] 1A is a schematic diagram of the human detection system 1 when the suspension height is low, and FIG. 1B is a schematic diagram of the human detection system 1 when the suspension height is high. [Diagram 2] 1 is a block diagram of a human detection system 1 according to the present invention. [Diagram 3] 1A is a schematic diagram of the crane when the suspended load is moved from a low position to a high position, and FIG. 1B is a display image on the person detection monitor 9. FIG. [Figure 4] 13 is an example of a display image on the person detection monitor 9. [Diagram 5] 13 is an example of a display image on the person detection monitor 9. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0018] The human detection system 1 according to the present invention is Lower area of ​​the load This system aims to prevent crane accidents by detecting whether or not a person (worker) is within the danger zone from image data captured by a camera.

[0019] As shown in FIG. 1 and FIG. 2, the human detection system 1 detects a load 4 attached to a crane 2. lower areaThe crane is equipped with a person detection camera 5 that captures an image of a person, a rangefinder 7 that is attached near the person detection camera 5 and measures distances H1 and H2 to a work surface 6, a camera control unit 8 that controls the person detection camera 5, and a person detection monitor 9 that is provided in the crane driver's seat and displays images captured by the person detection camera 5.

[0020] The crane 2 may be a jib crane equipped with a boom 3 (jib) as in the illustrated example, or may be an overhead crane, a bridge crane, etc. Although a mobile crane that can travel on wheels or the like is illustrated in Fig. 1, a fixed type may also be used.

[0021] The person detection camera 5 is generally called a crane camera, and is a camera that photographs the area around the lower part of the suspended load 4. The person detection camera 5 is attached to the crane body near the base end of the wire that hangs down to suspend the suspended load 4, and in the case of the jib crane shown in the figure, it is attached to the tip of the boom 3. The person detection camera 5 used has a zoom function with an optical zoom of about 30x. Images photographed by the person detection camera 5 are displayed in real time on a person detection monitor 9 installed in the crane driver's seat.

[0022] The work surface 6 is the surface of the suspended load 4. lower area In the case of lifting a load placed on the ground to the top of a building, the work surface 6 refers to the ground at the beginning of the lifting and the top of the building when the load is lowered. In this way, the level height of the work surface 6 may change as the load 4 moves horizontally.

[0023] As the range finder 7, any known non-contact range finder such as an optical type, an ultrasonic type, or a laser type can be used without any restrictions, and among these, a highly accurate and inexpensive laser type is particularly suitable.

[0024] In this human detection system 1, the camera control unit 8 has an auto-zoom function that automatically zooms in conjunction with the distance measured by the range finder 7 so that a danger area 10, which is an area on the work surface 6 centered on the suspended load 4, is displayed at a constant size on the human detection monitor 9. That is, the camera control unit 8 performs control such that, when the distance H1 measured by the range finder 7 is small as shown in FIG. 1(A), the camera control unit 8 automatically lowers the zoom magnification of the human detection camera 5, and when the distance H2 measured by the range finder 7 is large as shown in FIG. 1(B), the camera control unit 8 automatically increases the zoom magnification of the human detection camera 5. As a result, the danger area 10 is always displayed at a constant size on the human detection monitor 9 as shown in each of the (b) diagrams of FIGS. 1(A) and (B). The relationship between the distance H measured by the rangefinder 7 and the zoom magnification of the human detection camera 5 controlled by the camera control unit 8 can be set to a proportional relationship, for example, such that when the distance H is 10 m, the zoom magnification is 1.0x, when the distance H is 20 m, the zoom magnification is 2.0x, and when the distance H is 35 m, the zoom magnification is 3.5x.

[0025] The danger area 10 is an area on the work surface 6 centered on the center of the suspended load 4 (suspension core position). The planar shape is preferably circular as in the illustrated example, but may be elliptical, oval, polygonal, etc. The size (radius) of the danger area 10 can be set arbitrarily in advance according to the suspended load length, which is the longitudinal length of the suspended load 4. The size of the danger area 10 can be set arbitrarily for each crane operation of the day or for each type of suspended load, taking into account the suspended load length. To give a specific example, when the suspended load length is 5 m, the radius R of the danger area 10 can be set to 10 m, and when the suspended load length is 12 m, the radius R of the danger area 10 can be set to 18 m.

[0026] In this way, in this person detection system 1, even if the distance to the work surface 6 changes, the danger area 10 is always displayed at a constant size on the person detection monitor 9, making it easier for the crane operator to detect the worker W on the work surface 6 without having to manually zoom in.This makes it possible to detect work actions that could lead to accidents, such as the entry of unauthorized persons during crane operation or workers entering under the suspended load, in advance, helping to prevent crane accidents.

[0027] As shown in Fig. 1 and Fig. 2, in this person detection system 1, a camera 11 for photographing a suspended load is provided at a position where the hoisting core of the crane 2 can be photographed, and a monitor 12 for displaying a suspended load photographed by the camera 11 for photographing a suspended load is provided in the crane driver's seat. In the case of the illustrated jib crane, the camera 11 for photographing a suspended load is preferably provided at the tip of the boom 3 so that the hoisting core of the crane can be photographed. As shown in each of Figs. 1(A) and (B)(c), the camera 11 for photographing a suspended load is preferably always fixed in a long-distance image, i.e., with a reduced zoom magnification, so that the hoisting core can be confirmed. Therefore, a camera without a zoom function may be used as the camera 11 for photographing a suspended load.

[0028] As shown in each of Figures 1(A) and (B) (c), the image captured by the suspended load photographing camera 11 shows the worker W large when the height H1 to the work surface 6 is low, but when the height H2 to the work surface 6 is high, the worker W appears small and it becomes difficult to recognize him on the screen.

[0029] As shown in FIG. 2, the human detection system 1 includes an image processor 13 that processes the image captured by the human detection camera 5, determines whether or not a worker W is present in the danger area 10, and issues an alarm signal if the worker W is present. The image processor 13 adds the danger area 10 to the image data transmitted from the human detection camera 5, and transmits the image data to which the danger area 10 has been added to the human detection monitor 9. The image processor 13 also determines whether or not a worker W is present in the danger area 10, and sends a command to an alarm generator 14 to issue an alarm signal if the worker W is present. When the alarm generator 14 receives a command to issue an alarm signal, the alarm generator 14 can take one of the following measures, or two or more measures at the same time: issue an alarm sound, display an alarm on the human detection monitor 9, or stop driving the crane.

[0030] In order for the image processing unit 13 to determine whether or not the worker W is present in the danger area 10, it is preferable for the image processing unit 13 to detect the worker W by identifying the color of the worker's helmet. The image processing unit 13 can accurately identify the color of the helmet of the worker W from the image data by performing appropriate learning using a technology such as artificial intelligence (AI).

[0031] Also, the color of the helmets of the slingers, slingers, signallers, supervisors and other workers may be differentiated from the color of the helmets of other workers, and the image processing unit 13 may be configured not to issue an alarm signal when only the helmets of the slingers are detected in the danger area 10, but to issue an alarm signal when the helmets of workers other than those involved in the slingers are detected in the danger area 10. The helmet colors can be differentiated by the color of the helmet band or helmet cover.

[0032] Next, the operation of the human detection system 1 during actual crane work will be described with reference to Fig. 3. Assume that the crane work involves lifting a load placed on the ground to the top of a building, as shown in Fig. 3(A). Fig. 3(A) shows the movement process of a suspended load 4 moving in sequence from (a) to (d) superimposed on one drawing, and Fig. 3(B) shows images displayed on the human detection monitor 9 at each of the suspended load positions (a) to (d) in sequence. The worker Ws involved in the lifting work wears a black helmet, and the other workers Wt wear white helmets.

[0033] Although the distance between the tip of the boom 3 and the work surface 6 differs at each of the lifting positions (a) to (d), the zoom magnification of the person detection camera 5 is appropriately controlled by the camera control unit 8, so that the danger area 10 is displayed at approximately the same size in the image displayed on the person detection monitor 9 at each of the lifting positions. This makes it easy to recognize the workers Ws and Wt on the work surface 6, and makes it easy to distinguish between Ws and Wt by the color of their helmets.

[0034] At the hanging positions (a) and (d), a person Ws involved in the hanging work is present within the danger zone 10, but since no workers unrelated to the hanging work are present, no alarm is issued. On the other hand, at the hanging position (b), since a worker Wt not involved in the hanging work is present within the danger zone 10, the image processing unit 13 issues a command to the alarm generator 14 to generate an alarm signal, and an alarm is issued.

[0035] In the present human detection system 1, as shown in FIG. 4, the size of the danger area 10 displayed on the human detection monitor 9 can be set arbitrarily. In both of FIGS. 4(A) and (B), the radius R of the danger area 10 is set to 20 m, but the size of the danger area 10 displayed on the human detection monitor 9 is larger in FIG. 4(A) and smaller in FIG. 4(B). That is, the zoom magnification of the human detection camera 5 is larger in FIG. 4(A) and smaller in FIG. 4(B). As a result, in FIG. 4(A), the worker W present in the danger area 10 can be more clearly recognized, and in FIG. 4(B), the worker W present outside the danger area 10 but in the vicinity thereof can also be grasped. The size of the danger area 10 on the screen can be arbitrarily changed by a crane operator or the like, and may be set to a predetermined size before the start of crane work, or may be switched between large and small by a simple button operation during crane work.

[0036] 5, the zoom magnification of the human detection camera 5 can be automatically changed in conjunction with a change in the set value of the radius of the danger area 10, so that the size of the danger area 10 displayed on the human detection monitor 9 can be kept constant. This ensures that the danger area 10 is always automatically displayed at a constant size on the human detection monitor 9, regardless of the set value of the radius of the danger area 10. EXAMPLES

[0037] The number of people visually detected using actual image data in long-distance, mid-distance, and short-distance images (P 0 ) and the number of people automatically detected by a computer in the image processing unit 13 (P c ) and the detection rate (D = P c / P 0 × 100%) was calculated.

[0038] [Table 1]

[0039] As shown in Table 1, the accuracy of automatic detection by a computer improves as the images are closer than those from long distances. With close distance images, further improvement in the detection rate can be expected through additional learning by AI, but with long distance images, a significant improvement in the detection rate cannot be expected due to image quality and other factors. [Explanation of symbols]

[0040] 1...person detection system, 2...crane, 3...boom, 4...suspended load, 5...person detection camera, 6...work surface, 7...rangefinder, 8...camera control unit, 9...person detection monitor, 10...danger area, 11...camera for photographing suspended load, 12...monitor for displaying suspended load, 13...image processing unit, 14...alarm generator

Claims

1. The system comprises a person detection camera which photographs an area below a load attached to a crane, a rangefinder which is attached near the person detection camera and which measures the distance to a work surface, a camera control unit which controls the person detection camera, a person detection monitor which is provided in the crane driver's seat and which displays an image photographed by the person detection camera, and an image processing unit which processes the image photographed by the person detection camera, determines whether or not a worker is present in a danger area, and issues a warning signal if a worker is present, the image processing unit adds a danger area to the image data transmitted from the human detection camera, and transmits the image data to which the danger area has been added to the human detection monitor; The camera control unit has an auto-zoom function that automatically zooms in conjunction with the distance measured by the rangefinder so that a danger area, which is an area on the work surface centered on the suspended load, is displayed at a constant size on the human detection monitor, This human detection system for crane operation is characterized in that the colors of the helmets of people involved in the hoisting work are differentiated from the colors of the helmets of other workers, and the image processing unit does not sound an alarm if only the helmet of a person involved in the hoisting work is detected in the danger zone, but sounds an alarm signal if the helmet of a worker other than the person involved in the hoisting work is detected in the danger zone.

2. A human detection system for crane operation as described in claim 1, wherein a camera for photographing a suspended load is provided at a position where the crane's suspension core can be photographed, and a monitor for displaying a suspended load is provided in the crane driver's seat for displaying an image photographed by the camera for photographing the suspended load.

3. 3. A human detection system for crane operation according to claim 1, wherein the size of the danger area displayed on the human detection monitor can be arbitrarily set according to the length of the suspended load.

Citation Information

Patent Citations

  • JP1986197081U

  • Hung load watching device for crane

    JP2001002369A

  • Safety confirmation device of crane

    JP2010241548A

  • Object detection system

    JP2018160153A

  • Hanging load monitoring system for crane

    JP2020007134A