Safety device and pick-up lift having the same

Abstract

To estimate a mounting state of a mounting tool of a safety belt using a learned model with a new configuration in a safety device of a picking lift.SOLUTION: In the safety device 3 for a picking lift, the first to third marks 31, 32, 33 are attached to the attachment tool 40 of the safety belt 4. The safety device 3 includes the imaging device 5 that images the driver's cab of the picking lift, the target detection unit 70 that inputs the image data acquired by the imaging device 5 to the learned model 730 and detects the first to third detection targets included in the image data, and the determination unit 71 that determines whether to output a warning based on the detection result of the target detection unit 70. The first, second, and third detection targets are portions including a part of the work clothes of the operator, a part of the mounting fixture 40 on the part of the work clothes, and the first, second, and third marks 31, 32, and 33 attached on the part of the mounting fixture 40, respectively.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present application relates to a safety device for a picking lift that prevents an operator from falling from the cab, and a picking lift equipped with the same. [Background technology]

[0002] A picking lift is equipped with a driver's cab that can be raised and lowered along a mast and a pair of forks that protrude rearward from the cab. Handling work using a picking lift involves supporting a pallet with the forks, and the operator moving the cab to a higher position to load and unload cargo between the shelf and the pallet supported by the forks.

[0003] For example, as in Patent Document 1, some picking lifts are equipped with a safety device to prevent the operator from falling from the cab to the floor. The safety device has a safety belt that connects the cab to the operator riding in it. As in Patent Document 1, for example, the safety belt is a waist belt type safety belt that has a belt worn around the operator's waist.

[0004] In such picking lifts, operators may find it troublesome to wear a belt or dislike being restrained by a belt, so they may perform loading and unloading work in the high-altitude cab without wearing a belt. This work is extremely dangerous.

[0005] Therefore, the safety device of Patent Document 1 is equipped with an image processing device. The image processing device captures an image of the driver's cab with a camera, processes the acquired image data, and determines whether the operator is wearing the belt. Specifically, the image processing device determines whether the operator is wearing the belt by detecting a first mark on the front side of the belt or a second mark on the back side of the belt in the image data through image processing.

[0006] If the picking lift determines that the belt is not fastened as a result of image processing, it outputs a warning in the form of sound and / or light to the operator to urge them to fasten the belt. In addition, if necessary, the picking lift disables travel operation and hydraulic operation.

[0007] However, with the recent development of machine learning technology, it may be possible to estimate (determine) the wearing status more precisely by detecting the state of the belt and the operator in the cab using a trained model.

[0008] However, to achieve this, a huge amount of training data including the operator and the entire belt must be prepared, which requires a large-capacity image processing device. Furthermore, to detect such wearing conditions on a real-time basis, an image processing device with extremely high processing speed is required. As a result, the image processing device will inevitably become larger and more expensive. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-144865 Summary of the Invention [Problem to be solved by the invention]

[0010] This application provides a novel configuration for estimating the attachment state of attachment devices using a trained model in a picking lift safety device. [Means for solving the problem]

[0011] The present application provides a safety device for a picking lift to prevent an operator from falling from a cab, the safety device comprising: a safety belt having a harness attached to the operator; A first mark attached to the front side of the wearing tool; a second mark attached to the front side of the wearing tool, the second mark having a different characteristic from the first mark; a third mark attached to the front side of the wearing tool, the third mark having characteristics different from the first mark and the second mark; an imaging device that images the driver's cab; an object detection unit that inputs image data acquired by the imaging device into a trained model and detects a first detection object, a second detection object, and a third detection object included in the image data; and a decision unit that decides whether to output a warning based on a detection result of the object detection unit. wherein the first detection target is a portion consisting of a part of the operator's work clothes, a part of the wearing equipment on the part of the work clothes, and the first mark attached to the part of the wearing equipment, the second detection target is a portion consisting of a part of the operator's work clothes, a part of the wearing equipment on the part of the work clothes, and the second mark attached to the part of the wearing equipment, The third detection target includes a part of the operator's work clothes, a part of the wearing equipment on the part of the work clothes, and the third mark attached to the part of the wearing equipment.

[0012] For example, when the operator is in a driving position in which he or she operates the control panel of the picking lift with the attachment properly attached to the driver's seat, the imaging device may be oriented so that the first detection target is imaged by the imaging device and the second detection target and the third detection target are not imaged, and the first mark, the second mark, and the third mark may be attached to the front side of the attachment.

[0013] Furthermore, for example, when the operator is in a position facing the opposite direction to the driving position with the attachment properly attached in the driver's seat, the imaging device may be oriented and the first mark, the second mark, and the third mark may be attached to the front side of the attachment so that the second detection target is imaged by the imaging device and the first detection target and the third detection target are not imaged.

[0014] For example, the determination unit may determine to output the warning when the object detection unit continuously detects the same detection object for a set period of time.

[0015] For example, the safety device may further include: a mounting device detection device provided in the cab, the mounting device including a fastener for hanging the mounting device, and a sensor arranged to detect whether the mounting device is hung on the fastener; The determination unit It may be determined whether to output the warning based on the output of the sensor and the detection result.

[0016] For example, the wearing device is a belt worn around the waist of the operator.

[0017] Further, the present application provides a picking lift, The picking lift is The above safety devices, and a vehicle control unit that controls the operation of the picking lift.

[0018] For example, the picking lift may be Further comprising a warning device that outputs the warning in the form of sound and / or light, The vehicle control unit may activate the warning device when the decision unit decides to output the warning.

[0019] For example, the vehicle control unit may control the picking lift so as to disable operation of the picking lift when the decision unit decides to output the warning. [Brief explanation of the drawings]

[0020] [Figure 1] FIG. 1 is a schematic side view of an exemplary picking lift. [Figure 2] FIG. 2 is a block diagram showing the configuration of the picking lift. [Figure 3] FIG. 3 is a block diagram showing the configuration of the image processing device. [Figure 4] FIG. 4 is a side view showing the state of the harness and / or the operator. [Figure 5] FIG. 5 is a side view showing the state of the harness and / or the operator. [Figure 6] FIG. 6 is a diagram illustrating an example detection target. [Figure 7] FIG. 7 is a flowchart illustrating an example of control of the picking lift. DETAILED DESCRIPTION OF THE INVENTION

[0021] Embodiments of the present invention will now be described with reference to the accompanying drawings, which are merely exemplary of the invention. The drawings are schematic only and may not be drawn to scale. It should also be understood that the same reference numerals are used throughout the figures to represent the same or similar elements.

[0022] As shown in Figure 1, the picking lift 1 comprises a vehicle body 10, a mast 11 erected at the rear of the vehicle body 10, a driver's cab 2 that moves up and down along the mast 11, and a pair of left and right forks 12 that protrude rearward from the driver's cab 2.

[0023] The cab 2 includes a floor frame 20, a front frame 21, a head guard 22, and an operation panel 23 provided on the front frame 21. The operation panel 23 is provided with a key switch for starting the picking lift 1, as well as handles and levers for instructing the picking lift 1 to travel and the cab 2 to rise and fall.

[0024] As shown in FIG. 2, the picking lift 1 includes a vehicle control unit 13, a traveling system 14, a hydraulic system 15, a warning device 16, and a safety device 3.

[0025] The traveling system 14 includes a traveling motor for rotating the wheels provided on the vehicle body 10, a turning motor for turning the wheels, etc. The hydraulic system 15 includes a hydraulic tank, a hydraulic pump, a hydraulic cylinder, etc. for raising and lowering the cab 2 along the mast 11.

[0026] The warning device 16 outputs a warning in the form of sound and / or light. In this embodiment, the warning device 16 is composed of a lamp, a buzzer, and / or a speaker, etc., provided on the front frame 21 of the cab 2, and issues a warning to the operator O by lighting or flashing a lamp, sounding a buzzer, and / or outputting a voice from the speaker, etc., as described below.

[0027] The vehicle control unit 13 is made up of a microprocessor (MPU) and is configured to control the traveling system 14 and hydraulic system 15 in response to inputs from the operator O via the operation panel 23. This allows the operator O to operate the picking lift 1. The vehicle control unit 13 is also configured to control the traveling system 14, hydraulic system 15, and warning device 16 based on commands from the safety device 3, as will be described later. The vehicle control unit 13 is realized by executing a program stored in a storage medium.

[0028] 1, the safety device 3 includes a safety belt 4. In this embodiment, the safety belt 4 is a reel-type waist belt-type safety belt, and includes a wearing device 40, a strap 41, and a reel 42. In this example, the wearing device 40 is a belt worn around the waist of the operator O. One end of the strap 41 is attached to the wearing device 40, and the other end is attached to the reel 42. Alternatively, the safety belt 4 may be a harness-type safety belt, and the wearing device 40 may be configured as a harness.

[0029] The reel 42 is fixed to the head guard 22 of the cab 2. A portion of the strap 41 up to 1 m from one end is always pulled out from the reel 42, and when tension is applied, the remaining portion is pulled out in accordance with the tension. When a large tension is applied to the strap 41, the reel 42 prevents the strap 41 from being pulled out.

[0030] Therefore, if the operator O is wearing the attachment 40 properly, even if he falls from the cab 2 at a high place, he will be suspended near the cab 2 and will not collide with the floor.

[0031] Loading and unloading operations using the picking lift 1 are performed by supporting the pallet P on the forks 12, having the operator O board the cab 2 and move to a high position, and loading and unloading cargo W between a shelf (not shown) and the pallet P.

[0032] The safety device 3 is equipped with an imaging device 5 for capturing images of the riding space of the cab 2. The imaging device 5 includes a CCD camera or the like. In this embodiment, the imaging device 5 is provided at the left or right end of the front frame 21 below the operation panel 23, and is oriented so that the torso of the operator O riding in the cab 2 is included in the field of view. The imaging device 5 captures images at predetermined intervals, such as 30 images per second, to sequentially acquire image data.

[0033] The safety device 3 includes a wearing device detection device 6 (FIG. 2) for detecting whether the wearing device 40 is not being used by the operator O and is being held in a predetermined position on the cab 2. The wearing device detection device 6 includes a locking device 60 (FIG. 1) provided on the upper part of the front frame 21 of the cab 2 so as to be able to hold the wearing device 40. The locking device 60 is, for example, a hook.

[0034] The harness detection device 6 also includes a sensor 61 (FIG. 3) that detects whether the harness 40 is fastened to and held by the fastening device 60. The sensor 61 is, for example, a switch configured to turn on when the harness 40 is fastened to the fastening device 60, an optical sensor, or a weight sensor provided on the fastening device 60. This allows the harness detection device 6 to detect whether the harness 40 is not being used by the operator O and is hanging on the fastening device 60 (FIG. 4(A)) or whether it has been removed from the fastening device 60 (FIGS. 4(B), 4(C), etc.).

[0035] 2, the safety device 3 includes an image processing device 7. Although not shown in FIG. 1, the image processing device 7 is disposed in the cab 2 and is connected to the imaging device 5, the harness detection device 6, and the vehicle control unit 13. The image processing device 7 is used to estimate, on a real-time basis, the state of the harness 40 worn by the operator O and the posture of the operator O wearing the harness 40, based on image data acquired by the imaging device 5. Furthermore, for this purpose, at least three marks 31, 32, and 33 are attached to the front side of the harness 40 in the present application.

[0036] In this embodiment, three marks are used: a first mark 31, a second mark 32, and a third mark 33. These marks 31 to 33 have different characteristics. In this embodiment, the first mark 31 is an arrow mark, the second mark is a circle mark, and the third mark is a star mark. However, the marks 31 to 33 are not limited to these, and may be of any shape as long as they can be distinguishably detected by image processing, such as a one-dimensional code (barcode), a two-dimensional code, or an AR code.

[0037] In this embodiment, the wearing device 40 is a belt, and is shown in an unfolded state in Figures 2 and 3. When the belt is worn and forms a loop, the marks 31 to 33 are placed at approximately equal angular intervals (approximately 120° intervals).

[0038] Furthermore, the orientation of the imaging device 5 and the positions of the marks 31 to 33 on the mounting fixture 40 are determined as follows. As shown in FIG. 5(D), when the operator O is in a posture (driving posture) in which he or she properly wears the mounting fixture 40 and operates the operation panel 23 to operate the picking lift 1, of the marks 31 to 33, only the first mark 31 is imaged by the imaging device 5, and the second mark 32 and the third mark 33 are not imaged. On the other hand, as shown in FIG. 5(E), when the operator O is in a posture facing in the opposite direction to the driving posture (a posture that appears during picking work), of the marks 31 to 33, only the second mark 32 is imaged by the imaging device 5, and the first mark 31 and the third mark 33 are not imaged.

[0039] In the case of other standing postures when wearing the equipment, the imaging device 5 captures an image of only the third mark 33 or any two of the marks 31 to 33. When the operator is in a crouching posture or when the wearing equipment 40 is hung on the fastening tool 60 as shown in FIG. 4(A), none of the marks 31 to 33 are captured.

[0040] 3, the image processing device 7 includes an object detection unit 70, a determination unit 71, a model generation unit 72, and a storage medium 73. The object detection unit 70, the determination unit 71, and the model generation unit 72 are configured by a microprocessor (MPU), and are realized by executing a program stored in the storage medium 73. Note that, unlike the embodiment, these functional units 70 to 72 and the vehicle control unit 13 may be configured integrally by a single microprocessor.

[0041] The target detection unit 70 uses the trained model 730 stored in the storage medium 73 to detect whether the image data acquired by the imaging device 5 contains the first detection target, the second detection target, and the third detection target described below.

[0042] The first to third detection targets and the trained model 730 will be described. The first to third detection targets are parts of the operator O wearing the garment 40. More specifically, the first detection target is a portion consisting of a part of the work clothes C of the operator O, a part of the garment 40 on that part of the work clothes C, and a first mark 31 attached to that part of the garment 40, as shown in FIG. 6(A). The second detection target is a portion consisting of a part of the work clothes C, a part of the garment 40 on that part of the work clothes C, and a second mark 32 attached to that part of the garment 40, as shown in FIG. 6(B). The third detection target is a portion consisting of a part of the work clothes C, a part of the garment 40 on that part of the work clothes C, and a third mark 33 attached to that part of the garment 40, as shown in FIG. 6(C). That is, the first, second, and third detection targets respectively consist of the entire mark 31, 32, and 33 that appears when the operator O actually puts on the harness 40, and the surrounding area including the harness 40 and part of the work clothes C.

[0043] The storage medium 73 stores a large number of pieces of training data, each of which is image data including a first detection target (e.g., image data such as that shown in Figure 6(A)) labeled to indicate that it is the first detection target, image data including a second detection target (e.g., image data such as that shown in Figure 6(B)) labeled to indicate that it is the second detection target, and image data including a third detection target (e.g., image data such as that shown in Figure 6(C)) labeled to indicate that it is the third detection target.

[0044] Then, the model generation unit 72 generates a trained model 730 by having the neural network perform machine learning on the large amount of training data stored in the storage medium 73, and stores this in the storage medium 73.

[0045] Therefore, when image data is input to trained model 730, trained model 730 outputs output data indicating whether the first detection target is detected or not, whether the second detection target is detected or not, and whether the third detection target is detected or not. In other words, trained model 730 is a detector that distinguishes and detects the first to third detection targets in the image data.

[0046] Therefore, the target detection unit 70 inputs image data showing the state of the cab 2, which is sequentially acquired by the imaging device 5, into the trained model 730, and uses the output data to detect the first detection target, the second detection target, and the third detection target within the image data on a real-time basis.

[0047] In this embodiment, if one or two detection targets are detected, it is considered that the operator O is in a standing position wearing the attachment 40. Furthermore, as described above, if only the first detection target is detected, it is considered that the operator O is properly wearing the attachment 40 and is in a driving position (FIG. 5(D)). If only the second detection target is detected, it is considered that the operator O is in the process of picking (FIG. 5(E)). If none of the detection targets are detected, it is considered that the attachment 40 is not being worn. It is unlikely that three detection targets will be detected simultaneously in normal use of the attachment 40.

[0048] To improve detection accuracy, the model generation unit 72 may perform supervised machine learning by including image data that should not be erroneously detected as incorrect data in the training data, and then generate the trained model 730. Such incorrect data includes image data that includes only the marks 31, 32, and 33 but does not include the harness 40 or the work clothes C, or image data that includes only the portion of the harness 40 with the marks 31, 32, and 33 attached but does not include the work clothes C. This further reduces the possibility of incorrect detection in situations where the marks 31, 32, and 33 are captured by the imaging device 5 when only the harness 40 is present or when the harness 40 and the operator O are located apart, as shown in FIGS. 4(B) and 4(C).

[0049] The decision unit 71 decides whether to output a warning based on the detection result from the object detection unit 70, the output from the garment detection device 6 (its sensor 61), and vehicle information from the vehicle control unit 13, and when it has decided to output a warning, it sends an instruction to that effect to the vehicle control unit 13. Note that the image processing device 7 is equipped with a time counter (not shown), and the decision unit 71 uses the time counter to decide whether to output a warning.

[0050] As shown in Figure 7, the picking lift 1 equipped with the safety device 3 executes the following control. In the explanation of the flowchart in Figure 7, the first detection target is designated as (X), the second detection target is designated as (Y), and the third detection target is designated as (Z). The flowchart in Figure 7 is merely one example of the present application.

[0051] When the operator O turns on the key switch on the operation panel 23, the vehicle control unit 13 starts the picking lift 1 (S1). In response to this, the imaging device 5 captures an image of the cab 2 and acquires image data at predetermined intervals, and the object detection unit 70 detects the first to third detection objects (X) to (Z) in the image data.

[0052] The determination unit 71 first determines whether the wearing tool 40 is hooked on the fastening tool 60 using the output of the sensor 61 of the wearing tool detection device 6 (S2). Then, when the wearing tool detection device 6 does not detect the wearing tool 40 (NO in S2), the determination unit 71 determines whether the first detection target (X) is detected using the detection result from the target detection unit 70 (S3).

[0053] When only the first detection target (X) is detected in S3, the decision unit 71 decides to enable operation (travel and hydraulic operation) of the picking lift 1 and sends an instruction indicating this to the vehicle control unit 13, which then enables operation of the travel system 14 and the hydraulic system 15 (S4). As described above, when only the first detection target (X) is detected, it can be assumed that the operator O is properly wearing the attachment 40 and is in a driving position.

[0054] On the other hand, if the answer is YES in S2, it means that the attachment 40 is not in use, so the vehicle control unit 13 automatically turns off the power to the picking lift 1 even though the key switch is on (S15).

[0055] Following S4, the determination unit 71 counts time using a time counter (S5), and determines whether only the same detection target has been continuously detected for a predetermined first set time or longer (S6). For example, in the case of S3 → S4 → S5 → S6, the determination unit 71 determines whether only the first detection target (X) has been continuously detected for a first set time or longer.

[0056] If the answer is NO in S6, it is determined whether the key switch is still on (S7). If the answer is YES in S7, the process returns to S2. On the other hand, if the answer is NO in S7, this means that the key switch has been turned off, and the vehicle control unit 13 accordingly turns off the power to the picking lift 1 (S17).

[0057] If the answer is YES in S6, the decision unit 17 decides to output a warning and sends a command indicating this to the vehicle control unit 13, and the vehicle control unit 13 responds by activating the warning device 16 to output a warning (by sounding a warning sound and / or turning on a warning light), and also controls the driving system 14 and the hydraulic system 15 so that they cannot be operated (S11).

[0058] The loading and unloading work of a picking lift involves repeatedly operating the picking lift and picking operations. Therefore, if the operator O is properly wearing the attachment 40 while performing the loading and unloading work, it is generally impossible for the same detection target to be detected continuously for a long period of time. This may be because the attachment 40 is not worn correctly, or the operator O dislikes the attachment 40 and has tampered with the image processing device 7 to cause false detection so that the operator can operate without wearing it. Therefore, as in S6 → S11, if the same detection target is detected continuously for more than a predetermined set time, a warning is output to the operator O to properly wear the attachment 40, and control is implemented to disable travel and hydraulic operation.

[0059] This first set time is determined appropriately depending on the work environment of the picking lift and the properties of the cargo W in the cargo handling work, etc. As a non-limiting example, the first set time is 5 minutes.

[0060] If the answer is NO in S3, the decision unit 71 determines whether only the second detection target (Y) or only the third detection target (Z) is detected (S8). If the answer is YES in S8, the process proceeds to S5 described above. The steps following S5 are as described above. If the answer is NO in S8, the decision unit 71 determines whether the operator O is performing a travel operation or a hydraulic operation (S9).

[0061] In S9, if the decision unit 71 determines that hydraulic or traveling operation is being performed based on information from the vehicle control unit 13 (YES in S9), the process proceeds to S11, and the vehicle control unit 13 outputs a warning and performs operation disable control, as described above. This is because it is considered that the operator O is not wearing the attachment 40 or is operating the picking lift 1 in an inappropriate state or posture.

[0062] In S9, when the decision unit 71 determines based on the information from the vehicle control unit 13 that the hydraulic pressure or driving operation is not being performed (NO in S9), it determines whether two or more detection targets are detected (S10).

[0063] If the determination result in S10 is YES, the decision unit 71 proceeds to S5. The steps subsequent to S5 are as described above.

[0064] If the determination result in S10 is NO, the decision unit 71 proceeds to S11, and the vehicle control unit 13 outputs a warning and performs the operation disable control, as described above. In the flow of S3 and S8, the determination result in S10 is NO because it is clear that no detection target has been detected and the operator O is not in the wearing state and posture shown in FIG. 5(D).

[0065] After passing through S6 → S11 or S10 → S11, the decision unit 71 then counts the time and determines whether the counted time is equal to or greater than the second set time (S12). If the counted time is less than the second set time (NO in S12), the process returns to S11 and loops between S11 and S12. Therefore, the warning continues to be output for the second set time. A non-limiting example of the second set time is 1 to 2 minutes.

[0066] When the second set time has elapsed (YES in S12), the determination unit 71 determines whether the detection result is unnatural (S13). In this embodiment, an unnatural detection result is when no detection target is detected at the time of S13, or when the first detection target (X) is continuously detected over the first set time (S6) + the second set time (S12). This is because such a detection result is considered to indicate a very high possibility that the operator O is still not wearing the harness 40 properly.

[0067] If the answer is YES in S13, the decision unit 71 sends a command to the vehicle control unit 13 to stop the warning and displays an error message on a display not shown (S14), and then automatically turns off the power to the picking lift 1 (S15).

[0068] If the answer is NO in S13, the decision unit 71 causes the vehicle control unit 13 to stop the warning (S16), and then proceeds to S2. Therefore, after that, by going through S2, S3, and S4 in order, the operator O becomes able to operate (pilot) the picking lift 1.

[0069] In this manner, the safety device 3 of this embodiment affixes three marks 31 to 33 to the front side of the harness 40, and detects first, second, and third detection targets, each consisting of the marks 31 / 32 / 33 and parts of the harness 40 and work clothes C around them, using the trained model 730. The detection results (the type and number of detected detection targets) relate to the wearing state and posture of the operator O of the harness 40. Using this, the determination unit 71 determines whether to output a warning or prohibit operation, thereby urging the operator O to wear the harness 40 appropriately.

[0070] As explained in the background art, if machine learning is performed using training data that includes the operator and the entire attachment to generate a trained model and then attempt to detect the attachment status and posture on a real-time basis, an image processing device with a large capacity and high processing speed is required, which inevitably increases the size and cost of the device.

[0071] In contrast, in the embodiment of the present application, the detection target is only a portion consisting of the marks 31 / 32 / 33 and their surrounding areas on the attachment 40 worn by the operator O, and the detection target is detected using the trained model 730. Therefore, in the embodiment of the present application, the amount of training data required and the capacity and processing speed required of the image processing device 7 can be smaller or lower than those of the background art. Therefore, a safety device 3 for a picking lift 1 using a trained model can be realized while avoiding an increase in the size and cost of the image processing device 7.

[0072] In another embodiment, four or more marks may be attached to the attachment 40, four or more detection targets may be detected by the trained model 730, and the decision unit 71 may decide whether to issue a warning based at least on the type and number of the detected detection targets. [Explanation of symbols]

[0073] 1 Picking lift 13 Vehicle control unit 16 Warning device 2 Driver's cab 23 Control panel 3 Safety equipment 31 First Mark 32 Second Mark 33 Third Mark 4 Safety harness 40 Wearing equipment (e.g., belt) 5. Imaging device 6. Wearing device detection device 60 Locking device 61 Sensors 7 Image Processing Device 70 Target detection unit 730 trained models O Operator C Work clothes

Claims

1. A safety device for a picking lift to prevent an operator from falling from the cab, a safety belt having a harness attached to the operator; A first mark attached to the front side of the wearing tool; a second mark attached to the front side of the wearing tool, the second mark having a different characteristic from the first mark; a third mark attached to the front side of the wearing tool, the third mark having characteristics different from the first mark and the second mark; an imaging device that images the driver's cab; an object detection unit that inputs image data acquired by the imaging device into a trained model and detects a first detection object, a second detection object, and a third detection object included in the image data; a determination unit that determines whether to output a warning based on a detection result of the object detection unit, the first detection target is a portion consisting of a part of the work clothes of the operator, a part of the wearing equipment on the part of the work clothes, and the first mark attached to the part of the wearing equipment, the second detection target is a portion consisting of a part of the operator's work clothes, a part of the wearing equipment on the part of the work clothes, and the second mark attached to the part of the wearing equipment, the third detection target includes a part of the operator's work clothes, a part of the wearing equipment on the part of the work clothes, and the third mark attached to the part of the wearing equipment; Safety equipment.

2. The imaging device is oriented and the first mark, the second mark, and the third mark are attached to the front side of the attachment so that the first detection target is imaged by the imaging device and the second detection target and the third detection target are not imaged when the operator is in a driving posture in which the operator operates the operation panel of the picking lift with the attachment properly attached to the driver's cab.

2. The safety device of claim 1.

3. the imaging device is oriented and the first mark, the second mark, and the third mark are attached to the front side of the wearing tool so that the second detection target is imaged by the imaging device and the first detection target and the third detection target are not imaged when the operator is in a position facing a direction opposite to the driving position with the wearing tool properly attached in the cab; 3. The safety device of claim 2.

4. the determination unit determines to output the warning when the object detection unit continuously detects only the same detection object for a set time or more.

4. The safety device of claim 3.

5. The safety device further comprises: a mounting device detection device provided in the cab, the mounting device including a fastener for hanging the mounting device, and a sensor arranged to detect whether the mounting device is hung on the fastener; The determination unit determining whether to output the warning based on the output of the sensor and the detection result; 2. The safety device of claim 1.

6. The wearing device is a belt worn around the waist of the operator.

2. The safety device of claim 1.

7. A picking lift, A safety device according to any one of claims 1 to 6; A vehicle control unit that controls the operation of the picking lift, Picking lift.

8. Further comprising a warning device that outputs the warning in the form of sound and / or light, the vehicle control unit activates the warning device when the decision unit decides to output the warning; The picking lift according to claim 7.

9. the vehicle control unit controls the picking lift so as to disable operation of the picking lift when the decision unit decides to output the warning. The picking lift according to claim 8.

Images