Work analysis device and method

Abstract

This task analysis device for generating information on a plurality of workers performing a plurality of tasks in a workplace comprises: an acquisition unit that acquires image data representing images in which the workplace is captured; a control unit that, on the basis of the image data, generates task history information indicating a task performed in the workplace by an individual worker among the plurality of workers; and a storage unit that stores the task history information. The control unit sequentially recognizes the positions and tasks of the plurality of workers on the basis of the image data at each time in the workplace, detects entanglement between a plurality of flow lines which include the positions of the plurality of workers at each time, and when no entanglement is detected, generates task history information by mapping the tasks recognized at each time to individual workers on the basis of the plurality of flow lines, and when entanglement is detected, maps the recognized tasks to individual workers on the basis of the recognized tasks and previous task history information.

Description

[Technical Field]

[0001] The present disclosure relates to a work analysis apparatus and method. [Background technology]

[0002] Patent Document 1 discloses a video surveillance system that identifies a person and tracks their movements. The video surveillance system detects people and abandoned objects captured in images taken by any of multiple imaging devices and identifies the person who left the abandoned object. The video surveillance system searches for images capturing the person from among the images captured by each imaging device based on the person's facial features and clothing features such as the color and shape of their clothing. The video surveillance system outputs a display showing the movement of the person on a screen based on the imaging device that captured the image capturing the person and the time of capture. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2018 / 198373 Summary of the Invention [Problem to be solved by the invention]

[0004] The present disclosure provides an activity analysis device that can estimate the worker who will perform each activity when multiple activities are performed by multiple workers. [Means for solving the problem]

[0005] A work analysis device according to one aspect of the present disclosure generates information about multiple workers performing multiple tasks in a workplace. The work analysis device includes an acquisition unit, a control unit, and a storage unit. The acquisition unit acquires image data representing captured images of the workplace. The control unit generates work history information representing tasks performed in the workplace by individual workers from the multiple workers based on the image data. The storage unit stores the work history information. The control unit sequentially recognizes the positions and tasks of the multiple workers based on the image data for each hour in the workplace. The control unit detects congestion among multiple flow lines including the positions of the multiple workers at each hour. When congestion is not detected, the control unit generates work history information by associating tasks recognized at each hour with individual workers based on the multiple flow lines. When congestion is detected, the control unit associates the recognized tasks with individual workers based on the recognized tasks and past work history information.

[0006] These general and specific aspects may be realized by a system, a method, and a computer program, as well as combinations thereof. [Effects of the Invention]

[0007] According to the work analysis device and method disclosed herein, when multiple tasks are performed by multiple workers, it is possible to estimate the worker who will perform each task. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram showing an overview of a work analysis system according to a first embodiment. [Figure 2] A block diagram illustrating the configuration of a work analysis device in a work analysis system. [Figure 3] FIG. 1 is a diagram illustrating map data in a work analysis device. [Figure 4] FIG. 10 is a diagram for explaining work sequence information in a work analysis device. [Figure 5] FIG. 1 is a diagram showing a first example for explaining a problem related to a work analysis device. [Figure 6]FIG. 2 shows a second example to explain the problem related to the work analysis device. [Figure 7] A flowchart for explaining the overall operation of the work analysis device [Figure 8] 1 is a flowchart illustrating a worker discrimination process in the work analysis device of the first embodiment; [Figure 9] FIG. 10 is a diagram for explaining task combinations in worker identification processing; [Figure 10] FIG. 10 is a diagram for explaining work plan information in the work analysis device of the second embodiment. [Figure 11] 10 is a flowchart illustrating a worker discrimination process according to a second embodiment. [Figure 12] FIG. 10 is a diagram for explaining the worker discrimination process according to the second embodiment. [Figure 13] FIG. 10 is a diagram for explaining task combinations in the worker discrimination process of the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments will be described in detail with reference to the drawings as appropriate. However, more detailed description than necessary may be omitted. For example, detailed description of already well-known matters or redundant description of substantially identical configurations may be omitted. This is to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art. Note that the inventor(s) provide the accompanying drawings and the following description to enable those skilled in the art to fully understand the present disclosure, and are not intended to limit the subject matter described in the claims.

[0010] (Embodiment 1) 1. Configuration The work analysis system according to the first embodiment will be described with reference to Fig. 1. Fig. 1 is a diagram showing an overview of the work analysis system 1 according to the present embodiment.

[0011] 1-1. System Overview As shown in FIG. 1, the present system 1 includes a camera 2 and a work analysis device 5. The present system 1 is applied to applications in which the efficiency of workers W1, W2, and W3 performing multiple tasks in a workplace 6 such as a logistics warehouse is analyzed. Hereinafter, workers W1 to W3 will also be referred to as worker W. The present system 1 may also include a monitor 4 for presenting an analysis chart 7 relating to a predetermined analysis period to a user 3, such as a manager or analyst of the workplace 6. The analysis period is a period to be analyzed by image recognition or the like using the camera 2 in the present system 1, and is set in advance, for example, from one day to several months.

[0012] 1, a conveyor line 61 and shelves 62 are installed in a work area 6. In this example, the multiple tasks that each worker W1 to W3 performs while moving around the work area 6 include "collection" of taking products from shelves 62, "boxing" of packing products into boxes on the conveyor line 61 and discharging them, and "box preparation" of preparing the boxes.

[0013] The analysis chart 7 of this system 1 shows the ratio of each item during the analysis period for each worker W1 to W3, with each task classified into "main task," "secondary task," and "non-task" according to the added value of the task. In the example of Figure 1, packing is the main task, and auxiliary tasks related to the main task, such as collection and box preparation, and movement toward the conveyor line 61 or shelf 62, are secondary tasks. Waiting states unrelated to the main task are classified as non-task. In this way, the tasks to be analyzed by the task analysis device 5 are not limited to main tasks and secondary tasks, but also include non-tasks.

[0014] According to the work analysis system 1 of this embodiment, by presenting the analysis chart 7, the user 3 can analyze the work content of each worker W1 to W3, for example, to consider improving the work efficiency of the workplace 6.

[0015] The camera 2 of the present system 1 is positioned, for example, in the workplace 6 so as to capture the entire area in which the workers W1 to W3 move. The camera 2 repeats image capturing operations, for example, at a predetermined cycle in the workplace 6, and generates image data representing the captured images. The camera 2 is connected to the work analysis device 5, for example, so that the image data is transmitted to the work analysis device 5. Note that while one camera 2 is shown as an example in FIG. 1, the camera 2 included in the present system 1 is not limited to one camera, and may be two or more cameras.

[0016] The work analysis device 5 is configured as an information processing device such as a server device. The work analysis device 5 is communicably connected to an external information processing device such as a PC including a monitor 4. The configuration of the work analysis device 5 will be described with reference to FIG. 2.

[0017] 1-2.Configuration of the work analysis device Fig. 2 is a block diagram illustrating the configuration of the work analysis device 5. The work analysis device 5 illustrated in Fig. 2 includes a control unit 50, a storage unit 52, an operation unit 53, a device interface 54, and an output interface 55. Hereinafter, the interface will be abbreviated as "I / F."

[0018] The control unit 50 includes, for example, a CPU or MPU that works in cooperation with software to realize predetermined functions, and controls the overall operation of the work analysis device 5. The control unit 50 reads data and programs stored in the memory unit 52 and performs various arithmetic processes to realize various functions. For example, the control unit 50 includes an image recognition unit 51 as a functional component.

[0019] The image recognition unit 51 applies various image recognition technologies to the image data to recognize the position of a predetermined processing target in the image represented by the image data and output the recognition result. In the image recognition unit 51 of this embodiment, a person such as worker W is set as the processing target. The recognition result may include, for example, information indicating the time when the position of the processing target was recognized. The image recognition unit 51 performs image recognition processing using a trained model based on a neural network such as a convolutional neural network. The image recognition processing may be performed using various image recognition algorithms.

[0020] The control unit 50 executes, for example, a program including a set of instructions for realizing the functions of the work analysis device 5. The program may be provided via a communication network such as the Internet, or may be stored on a portable recording medium. The control unit 50 may also include an internal memory as a temporary storage area for storing various data and programs.

[0021] The control unit 50 may be a dedicated electronic circuit designed to realize a predetermined function or a hardware circuit such as a reconfigurable electronic circuit. The control unit 50 may be configured with various semiconductor integrated circuits such as a CPU, an MPU, a GPU, a GPGPU, a TPU, a microcomputer, a DSP, an FPGA, and an ASIC.

[0022] The storage unit 52 is a storage medium that stores programs and data necessary to realize the functions of the work analysis device 5. The storage unit 52 is configured, for example, by a hard disk drive (HDD) or a semiconductor storage device (SSD). For example, the storage unit 52 stores the above-mentioned programs as well as various types of information such as flow line data D0, map data D1, work sequence information D2, and authentication information D3.

[0023] The flow line data D0 indicates the flow line of the worker W moving in the workplace 6. The flow line data D0 is generated based on the recognition results obtained by inputting image data acquired from the camera 2 into the image recognition unit 51, for example. The map data D1 indicates the layout of various pieces of equipment such as the conveyor line 61 and shelves 62 in the workplace 6 in a predetermined coordinate system. The work sequence information D2 is information indicating the temporal execution sequence of a combination of tasks. The authentication information D3 is information for identifying individuals such as each worker W1 to W3. Details of each piece of information will be described later.

[0024] The storage unit 52 may include a temporary storage element configured by, for example, a DRAM or an SRAM, and may function as a work area for the control unit 50. For example, the storage unit 52 may temporarily store image data received from the camera 2 and the recognition results by the image recognition unit 51.

[0025] The operation unit 53 is a general term for operation members that accept user operations. The operation unit 53 is configured by, for example, any one of a keyboard, a mouse, a trackpad, a touchpad, buttons, switches, etc., or a combination thereof. The operation unit 53 acquires various information input by user operations.

[0026] The device I / F 54 is a circuit for connecting an external device such as the camera 2 to the work analysis device 5. The device I / F 54 performs communication in accordance with a predetermined communication standard. Predetermined standards include USB, HDMI (registered trademark), IEEE1395, IEEE802.11, Bluetooth (registered trademark), etc. The device I / F 54 is an example of an acquisition unit in the work analysis device 5 that receives various information from external devices. In the work analysis system 1, the work analysis device 5 acquires, for example, image data representing a video captured by the camera 2 via the device I / F 54.

[0027] The output I / F 55 is a circuit for outputting information. The output I / F 55 outputs video signals and the like to an external display device such as a monitor or projector for displaying various types of information, in accordance with, for example, the HDMI standard.

[0028] The above configuration of the work analysis device 5 is one example, and the configuration of the work analysis device 5 is not limited to this. The work analysis device 5 may be configured with various types of computers, including a PC (personal computer). In addition to or instead of the output I / F 55, the work analysis device 5 may also include a display unit configured, for example, as a built-in display device, such as a liquid crystal display or organic EL display. Furthermore, the work analysis method of this embodiment may be executed using distributed computing.

[0029] In addition to or instead of the above configuration, the work analysis device 5 may have a configuration for communicating with an external information processing device via a communications network. For example, the operation unit 53 may be configured to accept operations from an external information processing device connected via the communications network. Furthermore, the output I / F 55 may transmit various types of information to the external information processing device via the communications network.

[0030] The acquisition unit in the work analysis apparatus 5 may also be realized in cooperation with various software in the control unit 50, etc. The acquisition unit in the work analysis apparatus 5 may acquire various pieces of information by reading out the information stored in various storage media (for example, the storage unit 52) ​​into the work area of ​​the control unit 50.

[0031] 1-3. Various data structures

[0032] As described above, the work analysis device 5 of this embodiment stores the flow line data D0, map data D1, work sequence information D2, and authentication information D3 in the storage unit 52. An example of the structure of the various data D0 to D3 will be described below.

[0033] The flow line data D0 manages, for example, time, a flow line ID that identifies the flow line of the worker W, and the position of the worker W in the workplace 6 recognized at that time by the image recognition unit 51 in association with each other. The flow line data D0 associates, for example, map data D1 with the flow line based on the position of the worker W at each time.

[0034] 3 is a diagram illustrating the map data D1. The map data D1 manages the layout of sections and work areas (described later) by associating them with data showing a map coordinate system, such as the layout of various facilities viewed from above the workplace 6. Hereinafter, two directions that are perpendicular to each other on a horizontal plane in the workplace 6 are referred to as the X direction and the Y direction. A position in the workplace 6 is defined by, for example, an X coordinate indicating a position in the X direction and a Y coordinate indicating a position in the Y direction.

[0035] The map data D1 in FIG. 3 shows conveyor lines 61 and shelves 62 spaced apart in the X direction, corresponding to the workplace 6 shown in FIG. 1. In the example of FIG. 3, the conveyor lines 61 extend in the Y direction, and boxes are conveyed in a direction from positive to negative in the Y direction. The map data D1 in this example manages the workplace 6 by dividing it into multiple sections in the Y direction. FIG. 3 shows an example in which the workplace 6 is divided into section Z1 and section Z2. Each of sections Z1 and Z2 is set in advance as a unit section where, for example, a worker W performs a main task in the workplace 6.

[0036] Each of the sections Z1 and Z2 includes a work area indicating the region where the worker W works in the workplace 6. The section Z1 shown in FIG. 3 includes a work area A1 near the conveyor line 61 and a work area A2 near the shelf 62. Each of the work areas A1 and A2 is set in advance as an area indicating the range of positions in the workplace 6 where work related to the conveyor line 61 or the shelf 62, respectively, is expected to be performed.

[0037] Furthermore, in the work analysis device 5 of this embodiment, work area information that associates tasks with positions in the work area 6 is stored in the memory unit 52. The work area information is managed, for example, by associating work areas with the tasks performed in each work area for each section in the work area 6. For example, in work area A1 near the conveyor line 61 in section Z1, boxing and box preparation work is performed, and in work area A2 near the shelves 62 in section Z1, collection work is performed. For work area A1 with which multiple tasks are associated, the work area information may include information indicating, for example, the positional relationship in the Y direction and the correspondence between each task.

[0038] FIG. 4 is a diagram for explaining the work sequence information D2 in the work analysis device 5 of this embodiment.

[0039] As shown in FIG. 4, for example, the work sequence information D2 manages "sections" in the work area 6 in association with "abnormal sequences" that indicate sequences that are assumed to be abnormal in the execution sequence of work by one worker W. In the example of FIG. 4, the abnormal sequence of "packing, moving, packing" is associated with section Z1 in chronological order. Such an abnormal sequence is set in advance, for example, on the assumption that the sequence in which a worker W who has packed items in section Z1 moves and then packs items again without carrying them is abnormal. The abnormal sequence is an example of a predetermined sequence in this embodiment.

[0040] Although not shown in the figures, the work analysis device 5 of this embodiment stores, for example, in the memory unit 52, work trend information related to the workplace 6 in addition to the above-mentioned work sequence information D2. The work trend information includes, for example, information such as the standard work period set for each task performed by each worker W1 to W3. The work trend information may also include information indicating various tendencies in the way various tasks are performed by the worker W to be analyzed in the workplace 6. The work trend information may also include information indicating classifications such as main tasks and sub-tasks in the analysis chart 7.

[0041] Furthermore, the work analysis device 5 of this embodiment stores authentication information D3 for identifying individuals such as each of the workers W1-W3 in the memory unit 52. The authentication information D3 is obtained in advance, for example, by a card reader or the like installed in the workplace 6 through an authentication operation performed by each of the workers W1-W3 upon entering the workplace 6, and is then transmitted to the work analysis device 5. The authentication information D3 includes, for example, information indicating the time at which the authentication operation by each of the workers W1-W3 was accepted. The operation of the work analysis device 5 using this various information will be described later.

[0042] 2.Operation The operation of the work analysis system 1 and work analysis device 5 configured as above will be described below.

[0043] The work analysis system 1 shown in Figure 1 uses image recognition processing to recognize the position (i.e., movement line) of a worker W in a workplace 6 at each time, and recognizes the performed work, which is the work performed at each position at each time. The system 1 accumulates information indicating the recognition results of the performed work, and based on the accumulated information, generates an analysis chart 7 that visualizes the performed work for each worker W1 to W3 during the analysis period.

[0044] The work analysis device 5 of this embodiment associates the flow line ID with each worker W1-W3 based on, for example, the time when the position of each flow line ID in the workplace 6 was first recognized in the flow line data D0 and authentication information D3 including the time when each worker W1-W3 entered the workplace 6. The work analysis device 5, for example, uses the image recognition unit 51 to perform image recognition on images of the workplace 6 captured by the camera 2, and recognizes the position and performed work of the worker W. The work analysis device 5 updates the flow line data D0 to associate the recognized position with the past positions of each worker W1-W3, thereby identifying the worker W1-W3 who performed the performed work at each recognized position.

[0045] Here, even if the tasks being performed at each time can be recognized by image recognition processing based on images captured by camera 2, if congestion occurs, such as multiple traffic lines intersecting, it may be difficult to identify the traffic lines corresponding to each performed task or the worker W. In response to this, this embodiment provides a work analysis device 5 in the work analysis system 1 that can estimate the worker W who performed each recognized performed task, even if congestion such as the one described above occurs.

[0046] 2-1.About the issues A situation where a problem arises in identifying the worker W for each performed task in the task analysis system 1 of this embodiment will be described with reference to FIGS. 5 and 6. FIG.

[0047] Fig. 5 is a diagram showing a first example for explaining a problem related to the work analysis device 5. Fig. 6 is a diagram showing a second example for explaining a problem related to the work analysis device 5. Figs. 5 and 6 are views of a worker W in a workplace 6 seen from above, showing workers W1 and W2 working in section Z1.

[0048] Figure 5(A) shows a scene in which worker W1 is "collecting" and worker W2 is "packing." Figure 5(B) shows a scene in which workers W1 and W2 are "moving" from the scene in Figure 5(A). Figure 5(C) shows a scene in which worker W1, who has moved from the scene in Figure 5(B), is "packing" and worker W2 is "collecting."

[0049] In the examples of Figures 5(A) to 5(C), the position of worker W is recognized by image recognition from the image captured by camera 2, and the work performed at each position is recognized according to the work areas A1, A2, etc. in Figure 3. Here, in the scene of Figure 5(B), occlusion occurs, in which worker W2 is blocked by worker W1 in the line of sight of camera 2 and does not appear in the captured image. In this case, if the movement lines of workers W1 and W2 become confused, even if work performed at two positions is recognized in the example of Figure 5(C), it is difficult to determine from image recognition of each position to which each worker W1 and W2 moved and performed the corresponding work.

[0050] Figure 6(A) shows a scene in which worker W1 is "moving" and worker W2 is "packing." Figure 6(B) shows a scene from the scene in Figure 6(A) in which worker W1 is "moving" and worker W2 is continuing "packing." Figure 6(C) shows a scene from the scene in Figure 6(B) in which worker W1, who has moved, is "packing" and worker W2 is moving on to "preparing boxes."

[0051] In the examples of Figures 6(A) to 6(C), in the scene of Figure 6(B), occlusion occurs due to the movement of worker W1, similar to Figure 5(B). In this case, if the movement lines of workers W1 and W2 become mixed, even if the work performed at two positions is recognized in the example of Figure 6(C), it is difficult to determine from image recognition of each position to which position worker W1 moved to perform the corresponding work.

[0052] As described above, even if the positions and tasks performed by the workers W at each time point are recognized in the workplace 6, if the lines of movement of the workers W become mixed up, it may be difficult to distinguish the workers W performing each task. In particular, in a situation in the workplace 6 where, for example, the workers W1 to W3 are wearing uniforms of similar color and shape, it may be difficult to distinguish the workers W1 to W3 performing each task by image recognition or the like when the lines of movement become mixed up.

[0053] Therefore, the work analysis device 5 of this embodiment executes processing to estimate the worker of each work based on work tendency information such as work sequence information D2, in addition to the position of the work performed based on the captured images. As a result, it is possible to identify the worker W of each work performed even in situations such as those shown in Figures 5(C) and 6(C), where the movement lines of multiple workers are mixed up and it is difficult to identify the worker W from image recognition of each position.

[0054] For example, in the situation shown in Figure 5(C), based on the abnormal sequence of "packing, moving, packing" in the work sequence information D2, worker W2, who was packing, does not start packing after moving, so it can be assumed that worker W2 is collecting. In other words, it is possible to assume that the workers W who collected and packed were workers W2 and W1, respectively. Furthermore, in the situation shown in Figure 6(C), by using information about the period elapsed since the start time of packing for worker W2, it is possible to assume that the workers W who packed and prepared the boxes were workers W1 and W2, respectively.

[0055] 2-2. Overall operation The overall operation of the work analysis device 5 in the work analysis system 1 will be described with reference to FIG.

[0056] 7 is a flowchart for explaining the overall operation of the work analysis device 5. The processing shown in this flowchart is executed by the control unit 50 of the work analysis device 5, for example.

[0057] First, the control unit 50 acquires image data for an analysis period from the camera 2, for example, via the device I / F 54 (S1). For example, while workers W1 to W3 are working in the workplace 6, the camera 2 captures video and generates image data representing the captured images at each time at a predetermined cycle, such as the frame cycle of the video, and records the image data in its internal memory. The camera 2 transmits the image data recorded during the analysis period to the work analysis device 5. The control unit 50 stores the acquired image data in, for example, the memory unit 52.

[0058] Next, the control unit 50 selects one frame of image data representing an image captured at each time from the acquired image data for the analysis period, for example, in chronological order (S2). The control unit 50 records the time at which the selected one frame was captured as, for example, the time in the flow line data D0.

[0059] The control unit 50 functions as the image recognition unit 51 to recognize the position and work of the worker W in the image represented by the selected frame of image data (S3). In step S3, the control unit 50 converts, for example, the position recognized in the image into a coordinate system indicating the position in the workplace 6 based on the map data D1. The control unit 50 recognizes the work being performed at each recognized position based on, for example, work area information, depending on whether the recognized position is in work area A1, A2, or another area.

[0060] 6(C), the positions of two workers W are recognized in the work area A1 corresponding to the two tasks of box packing and box preparation. In this case, the control unit 50 recognizes the work being performed at the position of each worker W, for example, based on the relationship that box preparation is performed on the upstream side of the conveyor line 61 (the +Y direction in FIG. 3).

[0061] The control unit 50 detects a state of congestion of the flow lines, for example, based on the recognition result of step S3 (S4). For example, the control unit 50 detects whether or not occlusion due to overlapping of the positions of multiple workers W has occurred in the captured image of the selected frame. For example, the control unit 50 determines that congestion of the flow lines has occurred when it determines that occlusion has occurred and the recognized position in the workplace 6 is within a predetermined range from the positions of the multiple flow lines at the most recent time in the flow line data D0. The predetermined range is set in advance as a range small enough to be considered as the range in which the workers W move in the workplace 6 at the time interval of the frame period, for example.

[0062] If a line congestion state is not detected (NO in S4), the control unit 50 updates the flow line data D0 so as to add the position recognized in step S3 this time as the position of the corresponding flow line ID (S6). At this time, the control unit 50 associates the performed work for each position recognized in step S3 with each position in the flow line data D0 and the corresponding flow line ID, thereby determining the worker W1 to W3 associated with each performed work (S6). The information associating the performed work and worker W with each position in the flow line data D0 is an example of work history information in this embodiment.

[0063] On the other hand, when a line congestion state is detected (YES in S4), the control unit 50 of this embodiment determines the worker W associated with each recognized task based on the tasks recognized in the line congestion state and the past tasks performed by each worker associated with the flow line data D0 (S5). This line congestion state worker determination process (S5) makes it possible to estimate the worker W even in a line congestion state where it is not possible to determine the worker W for each task by associating the position recognized in step S3 with the past flow line. The control unit 50 of this embodiment performs the line congestion state worker determination process (S5) by referring to task tendency information such as the task order information D2. Details of the line congestion state worker determination process (S5) will be described later. Hereinafter, the line congestion state worker determination process will also be simply referred to as the line congestion state worker determination process.

[0064] After identifying the worker W for each task (S5, S6), the control unit 50 proceeds to step S7. If all frames in the image data for the analysis period have not yet been selected (NO in S7), the control unit 50 repeats the processing of steps S2 to S6 for the image data for the next time. This results in the acquisition of flow line data D0 based on the image data for each time during the analysis period. Note that the processing of steps S3 to S6 may be performed for each section of the workplace 6 as shown in FIG. 2, or the process may proceed to step S7 after steps S3 to S6 have been performed for all sections per frame.

[0065] When all frames in the analysis period have been selected (YES in S7), the control unit 50 performs a visualization process (S8) to generate an analysis chart 7. The control unit 50 counts the number of tasks determined for each time interval, such as the period of one frame, for each worker W1 to W3 in the workplace 6, for example. Once the total number of tasks in the analysis period for each worker has been calculated in this way, the control unit 50 calculates the proportion of each task for each worker and generates the analysis chart 7. In the analysis chart 7, the proportion of each task is shown, for example, as the proportion of time spent on each task relative to the analysis period.

[0066] The control unit 50 stores the analysis chart 7 generated by the visualization process (S8) in the storage unit 52, for example, and ends the process shown in this flowchart.

[0067] According to the above process, the positions and tasks of workers W in the workplace 6 are recognized based on the image data (S3), and the recognized positions are associated with past movements in the movement line data D0 to identify the workers W performing tasks at each position (S6). If the movements are mixed (YES in S4), the workers are identified through worker identification processing (S5). This provides information relating the tasks performed by each worker in the movement line data D0, and an analysis chart 7 is generated based on the tasks performed by each worker at all time intervals during the analysis period (S8).

[0068] In step S1 above, image data generated by camera 2 may be sequentially acquired. For example, instead of step S7, the control unit 50 may repeat the processes from step S1 onwards until traffic line data D0 based on image data of the number of frames in the analysis period is obtained. In addition, when detecting a traffic line congestion state (S4), the control unit 50 may detect a traffic line congestion state according to, for example, either an occlusion in the captured image or the position of the worker W in the workplace 6.

[0069] 2-3. Operator identification process The details of the worker discrimination process in step S5 of FIG. 7 will be described with reference to FIGS.

[0070] FIG. 8 is a flowchart illustrating the worker identification process (S5 in FIG. 7) in the work analysis device 5 of this embodiment. FIG. 9 is a diagram for explaining task combinations in the worker identification process. FIGS. 9(A) and 9(B) illustrate task combination tables T1 and T2 corresponding to the scenes shown in FIGS. 5(A) to 5(C) and 6(A) to 6(C), respectively. The task combination tables T1 and T2 store multiple task combination candidates indicating combinations of multiple workers W and multiple tasks. Each candidate indicates a task combination in which a task sequence including two or more tasks is associated for each worker. The task sequence indicates a set of tasks in which each worker's past tasks and the task being performed for which the worker is to be identified are arranged in chronological order of recognition.

[0071] In the worker identification process (S5) of this embodiment, the control unit 50 refers to task tendency information such as task sequence information D2, and performs a process of determining one candidate from multiple candidates in the task combination tables T1, T2 as the task combination of the identification result. The control unit 50 identifies the worker W for each task according to the determined task combination.

[0072] 8, first, the control unit 50 calculates task combinations between tasks recognized after the time when the line congestion state was detected and workers W associated with past tasks, and generates task combination tables T1 and T2 (S11). The control unit 50 generates task combinations that include at least two or more types of tasks, for example, by referring to past tasks and workers W associated with the flow line data D0.

[0073] In the example of FIG. 5, the movement lines of workers W1 and W2 become mixed in the scene of FIG. 5(B). In this case, the work performed by both workers W is recognized as "movement" (S3 in FIG. 7), and workers W1 and W2 can be associated with the same type of work without any special distinction. In such a case, in step S11, the control unit 50 selects, for example, a frame at the next time, and recognizes the position and work of worker W based on the image data of that frame, as in step S3 in FIG. 7. As a result, a frame corresponding to the scene of FIG. 5(C) is selected, and the position of worker W and the work of "packing" and "collection" at each position are recognized.

[0074] 5, the control unit 50 generates an operation combination table T1, as shown in FIG. 9A, based on the workers W1 and W2 involved in the line congestion and the operations recognized from the frames corresponding to the scenes in FIGS. 5A-5C. For example, when detecting a line congestion (S4 in FIG. 7), the control unit 50 determines the workers W1 and W2 involved in the line congestion, i.e., the workers W1 and W2 corresponding to the multiple flow lines where the line congestion occurred, based on the position recognized in step S3 and the past position of the worker W in the flow line data D0. The control unit 50 generates candidate operation combinations for the two workers W1 and W2, corresponding to the number of possible patterns, by interchanging the operations that cannot be associated with the worker W based on the flow lines recognized at the time in FIG. 5C (S11).

[0075] In the task combination table T1 of Figure 9(A), candidates C11 and C12 are stored that include the corresponding past tasks "collection, movement" and "packing, movement" for workers W1 and W2, and that swap "collection" and "packing" which cannot be associated with worker W based on their movement line.

[0076] In the example of Fig. 6, it is assumed that a line congestion state is detected in the frame of the scene in Fig. 6(C) (YES in S4 in Fig. 7). In this case, it is difficult to associate the tasks recognized in step S3 with workers W1 and W2 based on their movement lines as in step S6. Therefore, in this case, the control unit 50 generates an task combination table T2, as shown in Fig. 9(B) (S11), from the tasks recognized when the line congestion state was detected and past tasks corresponding to the scenes in Figs. 6(A) and (B).

[0077] In the work combination table T2 of Figure 9(B), candidates C21 and C22 are stored for workers W1 and W2, which include the past work performed "moving, moving" and "packing, packing", and which are obtained by swapping "box preparation" with "packing", which cannot be associated with worker W based on the movement line.

[0078] Next, the control unit 50 removes candidates from the task combination tables T1 and T2 based on the task sequence information D2 (S12). For example, the control unit 50 determines whether each candidate corresponds to an abnormal sequence in the task sequence information D2, and removes the corresponding candidate from the task combination tables T1 and T2.

[0079] For example, in the task combination table T1 of Figure 9(A), the task combination of worker W2 included in candidate C11 corresponds to the abnormal sequence in the task sequence information D2 illustrated in Figure 4. Therefore, candidate C11 is excluded from the task combination table T1 (S12).

[0080] Next, the control unit 50, for example, based on the standard work period information stored in the memory unit 52, excludes candidates whose work periods exceed the standard work period from the work combination tables T1 and T2 (S13). For example, the control unit 50 calculates the work period of the most recent work in each candidate work sequence, and if that work period exceeds a predetermined period indicating a significant excess over the standard work period, excludes the candidate that includes that work sequence. The standard work period is calculated by averaging the periods measured multiple times in advance as the period required for each work by each worker. The increment of the predetermined period is set, for example, to three times the standard deviation of the measured periods.

[0081] For example, in the task combination table T2 of Fig. 9(B), when the task period of the packing task continuing from the scene of Fig. 6(A) for worker W2 of candidate C21 exceeds a predetermined period, candidate C21 is excluded from the task combination table T2. Packing is an example of the first task in this embodiment.

[0082] After excluding candidates as described above (S12, S13), the control unit 50 determines whether there are multiple candidates that have not been excluded in the task combination tables T1, T2, i.e., whether multiple candidates remain (S14). If all candidates have been excluded in steps S12 and S13, the control unit 50 may, for example, relax the conditions for exclusion based on task tendency information and then execute the processes from step S11 onwards again so that at least one candidate remains by step S14. For example, the predetermined period for exclusion based on task period (S13) may be set longer than in the above example.

[0083] If there are no candidates remaining in the task combination tables T1, T2 (NO in S14), the control unit 50 determines the remaining candidate as the task combination of the discrimination result (S16). In the task combination tables T1, T2 of Figures 9(A) and 9(B), the remaining candidates C12, C22 that have not been excluded are determined as the task combination of the discrimination result.

[0084] On the other hand, if multiple candidates remain (YES in S14), the control unit 50 selects from the multiple candidates, for example, the candidate with the smallest difference between the duration of the most recently performed task in each task sequence and the standard task duration (S15).The control unit 50 determines the selected candidate as the task combination of the determination results (S16).

[0085] The control unit 50 determines, from the task combinations resulting from the determination, the worker W of each task that cannot be associated with the worker W based on the flow line recognized at the latest time (S17). The control unit 50 updates the flow line data D0 so as to add the position of each determined worker W at the latest time as the position of the corresponding flow line ID (S17).

[0086] In the example of Figure 5(C), the workers W performing the collection and boxing tasks are determined to be workers W2 and W1 from candidate C12 determined for the task combination in the task combination table T1 as a result of the discrimination. The flow line data D0 is then updated with the positions corresponding to collection and boxing as the positions of workers W2 and W1. In the example of Figure 6(C), the workers W performing the packing and box preparation tasks are determined to be workers W1 and W2 from candidate C22 determined for the task combination in the task combination table T2 as a result of the discrimination. The flow line data D0 is then updated with the positions corresponding to packing and box preparation as the positions of workers W1 and W2.

[0087] The control unit 50 identifies the worker W and updates the flow line data D0 (S17), and then ends the process shown in this flowchart. Then, the process proceeds to step S7 in FIG.

[0088] According to the above-described worker discrimination process for line congestion (S5), candidate task combinations including tasks that cannot be associated with the worker W due to the line congestion are generated (S11), and a task combination for the discrimination result is determined from the candidates according to the task tendency information (S12-S16). As a result, even if a line congestion occurs, the worker W for each task can be discriminated from the determined task combination (S17), and the worker W for each task can be estimated.

[0089] In addition, in the above step S11, an example was described in which the task sequence in the task combination tables T1, T2 includes tasks for three frames. The task sequence is not limited to tasks for each time period, such as three frames, but may include, for example, three different types of tasks. In this case, the control unit 50 generates the task combination tables T1, T2 by referencing past tasks associated with the flow line data D0 until, for example, three types of tasks are obtained. Furthermore, the task sequence is not limited to three types, and a task sequence may be generated in which three tasks are arranged for each predetermined period. Furthermore, the task sequence is not limited to three types, and a task sequence may be generated in which two tasks are arranged.

[0090] Furthermore, in generating the task combination tables T1 and T2 (S11), the candidate workers W may be narrowed down using the coordinate information and the moving distance per unit time in the flow line data D0. For example, the control unit 50 may determine the worker W associated with the congestion state in the task combination based on the moving speed of the worker W based on the past positions in addition to the past positions of the worker W in the flow line data D0.

[0091] 3. Effects etc. As described above, the work analysis device 5 in this embodiment generates information about multiple workers W performing multiple tasks in the workplace 6. The work analysis device 5 includes a device I / F 54 (an example of an acquisition unit), a control unit 50, and a storage unit 52. The device I / F 54 acquires image data representing captured images of the workplace 6 (S1). Based on the image data, the control unit 50 generates information associating tasks with each worker W at each position in the flow line data D0 as an example of work history information representing tasks performed by individual workers W1-W3 in the workplace 6 (S5, S6). The storage unit 52 stores the work history information. The control unit 50 sequentially recognizes the positions and tasks of the multiple workers W based on the image data for each time in the workplace 6 (S2, S3, S7). The control unit 50 detects a congestion state as an example of congestion between multiple flow lines including the positions of the multiple workers W at each time (S4). When a line congestion state is not detected (NO in S4), the control unit 50 generates work history information by associating the work recognized at each time with each of the workers W1 to W3 based on multiple flow lines (S6).When a line congestion state is detected (YES in S4), the control unit 50 associates the recognized work with each of the workers W1 to W3 based on the recognized work and past work history information (S5).

[0092] According to the above-described work analysis device 5, when a cross-traffic state is detected that makes it difficult to associate the work with the workers W1 to W3 based on the flow lines (S6) (YES in S4), the work is associated with the workers W1 to W3 based on the recognized work and past work history information (S5). As a result, when multiple work tasks are performed by multiple workers W in the workplace 6, it is possible to estimate the worker W for each work task.

[0093] In this embodiment, the memory unit 52 stores work tendency information that indicates the tendency of work to be performed in the workplace 6. When a line congestion state is detected (YES in S4), the control unit 50 refers to the work tendency information and associates the recognized work with each of the workers W1 to W3 (S5). As a result, even if it is difficult to associate the work with the workers W1 to W3 from image recognition of the positions of the workers W due to a line congestion state, it is possible to estimate the worker W corresponding to each work based on the work tendency information.

[0094] In this embodiment, when a line congestion state is detected (YES in S4), the control unit 50 generates multiple task combinations (S11), as an example of calculating multiple combinations of multiple workers W corresponding to the multiple traffic lines where the line congestion occurred and the multiple recognized tasks. The control unit 50 determines one task combination from the multiple task combinations based on the task tendency information (S16), and associates the recognized tasks with individual workers W1 to W3 according to the determined task combination (S17). In this way, task combination tables T1 and T2 containing multiple task combinations as candidates C11 to C22 are generated (S11), and the task combinations of the identification results are determined by narrowing down the candidates C11 to C22 based on the task tendency information (S16). As a result, the worker W for each task can be identified from the determined task combination.

[0095] In this embodiment, the task tendency information includes task order information D2 as an example of information indicating the order of a combination of two or more tasks among the plurality of tasks. The control unit 50 excludes task combinations that correspond to an abnormal order, an example of a predetermined order, from the plurality of task combinations according to the order (S12), and determines one task combination (S16). As a result, a task combination that does not correspond to an abnormal order in the task order information D2 can be determined as the determination result.

[0096] In this embodiment, the task trend information includes information indicating a standard task period set for a first task among the multiple tasks. The control unit 50, based on the period during which the task performed by worker W is recognized as the first task, excludes task combinations whose duration exceeds the standard task period from the multiple task combinations (S13) and determines one combination (S16). In the example of FIG. 6, candidate C22's task combination table T2 in FIG. 9(B) shows that worker W2's packing (an example of the first task) continues beyond the standard task period, so candidate C22 is excluded. This allows a task combination whose task period conforms to the standard task period to be determined as a discrimination result.

[0097] In this embodiment, the control unit 50 detects a line congestion state in accordance with occlusion, which is an example of a state in which the positions of workers W1 and W2 (an example of two or more workers) among multiple workers W are superimposed in the image represented by the acquired image data (S4). This makes it possible to detect a line congestion state based on the position of the worker W on the image.

[0098] In this embodiment, the storage unit 52 further stores authentication information D3 that identifies each of the workers W1-W3. The control unit 50 associates the recognized location of a worker W in the workplace 6 with each of the workers W1-W3 at the first time the location is recognized. In this way, the control unit 50 associates the flow line ID of each location in the flow line data D0 with each of the workers W1-W3 and manages them. This allows the flow line data D0 to be updated so that the positions of the workers W that are sequentially recognized (S3) are associated with the past positions of each of the workers W1-W3, and the workers W1-W3 who performed the work at each recognized location can be identified (S6).

[0099] The control unit 50 generates an analysis chart 7 for each of the individual workers W1-W3 based on the work history information for the analysis period (an example of a predetermined period) as an example of information showing the proportions of multiple tasks over the analysis period. This makes it possible to present the analysis chart 7 for multiple workers W performing multiple tasks in the workplace 6 to, for example, a user 3 of the work analysis system 1. The work analysis device 5 may further include an output I / F 55 and / or a monitor 4 as an example of a display unit that displays the generated information, such as the analysis chart 7.

[0100] The work analysis method in this embodiment is a method for generating information about multiple workers W performing multiple tasks in a workplace 6. This method includes a step (S1) in which a computer control unit 50 acquires image data showing an image of the workplace 6, and steps (S2-S7) in which, based on the image data, a computer control unit 50 generates work history information showing the tasks performed in the workplace 6 by individual workers W1-W3 of the multiple workers W. In the step (S2-S7) in which the computer control unit 50 generates work history information, the positions and tasks of the multiple workers W are sequentially recognized (S2, S3, S7) based on the image data for each hour in the workplace 6, and detects any cross-talk between multiple traffic lines including the positions of the multiple workers W for each hour (S4). When no line congestion is detected (NO in S4), the control unit 50 generates work history information by associating the work recognized at each time with individual workers W1 to W3 based on multiple movement lines (S6), and when a line congestion is detected (YES in S4), it associates the recognized work with individual workers W1 to W3 based on the recognized work and past work history information (S5).

[0101] In this embodiment, a program is provided for causing a control unit of a computer to execute the above-described task analysis method. According to the task analysis method of this embodiment, when multiple tasks are performed by multiple workers W, it is possible to estimate the worker W for each task.

[0102] (Embodiment 2) In the first embodiment, the work analysis device 5 performs the worker discrimination process based on the work period information and the work sequence information D2. In the second embodiment, the work analysis device 5 performs the worker discrimination process based on the work plan information related to the workplace 6.

[0103] Below, the description of the configuration and operation similar to those of the work analysis device 5 according to the first embodiment will be omitted as appropriate, and the work analysis device 5 according to this embodiment will be described.

[0104] 10 is a diagram for explaining the work plan information D4 in the work analysis device 5 of this embodiment. The work plan information D4 is an example of work plan information in this embodiment, and indicates quotas in the workplace 6, allocation of workers W, etc.

[0105] The work plan information D4 illustrated in FIG. 10 manages in association with the "workers" in the work area 6, the "number of items to be carried out" indicating the quota for the main work of boxing, and the "area in charge" which is the area in which each worker W mainly works. The area in charge indicates the range of positions in the work area 6 where the worker W does not perform auxiliary work such as box preparation. The auxiliary work is an example of the second work in this embodiment. The number of items to be carried out and the area in charge are set in advance by, for example, user 3. For example, in the work plan information D4 in FIG. 10, the area in charge of workers W1 and W2 is set to area Z1, and the area in charge of worker W3 is set to area Z2.

[0106] The operation of the work analysis device 5 of this embodiment, which uses the above-described work plan information D4, will be described with reference to FIGS.

[0107] Fig. 11 is a flowchart illustrating the worker determination process of this embodiment. In the work analysis device 5 of this embodiment, for example, in addition to the processes of steps S11 to S17 in the worker determination process (S5) of embodiment 1, the control unit 50 excludes candidates from the task combination table based on the task plan information D4 (S21). Fig. 12 is a diagram for explaining the worker determination process of this embodiment. Fig. 13 is a diagram for explaining the task combination in the worker determination process of this embodiment.

[0108] Fig. 12 is a view of the workplace 6 seen from above, similar to Fig. 5 and Fig. 6. Fig. 12 shows workers W1 and W2 working in section Z1 of the workplace 6, as well as worker W3, who is in charge of a different section.

[0109] FIG. 12(A) shows a scene in which worker W1 is "collecting," worker W2 is "packing," and worker W3, who has entered zone Z1 from zone Z2, is "moving." FIG. 12(B) shows a scene in which workers W1 to W3 are each "moving" from the scene in FIG. 12(A). FIG. 12(C) shows a scene in which workers W1 and W3, who have moved, are "packing" and "preparing boxes," respectively, from the scene in FIG. 12(B), and worker W2 is "collecting." Below, we will explain an example in which the movement lines of workers W1 to W3 become confused due to occlusion in the scene in FIG. 12(B).

[0110] The flowchart shown in FIG. 11 starts when a crosstalk state is detected (YES in S4 of FIG. 7) based on image data of a frame corresponding to the scene in FIG. 12(B), for example.

[0111] The control unit 50 recognizes the position and tasks of the worker W from the next frame corresponding to the scene in Fig. 12(C) and generates a task combination table (S11), similar to the example in Fig. 5 in the worker discrimination process (S5) of embodiment 1. Fig. 13 illustrates an example of the task combination table T3 generated in step S11 according to the examples in Figs. 12(A) to 12(C).

[0112] Next, the control unit 50 removes candidates from the task combination table T3 based on the work plan information D4 (S21). For example, the control unit 50 references the assigned section in the work plan information D4 and removes from the task combination table T3 candidates for which the most recent task performed by each worker in the assigned section is the auxiliary task of box preparation. For example, since it is assumed that worker W is likely to perform the main task of box packing or collection related to the main task in the assigned section, an exclusion rule based on the work plan information D4 is preset. In the work plan information D4 of FIG. 10, since the assigned section for worker W1 is section Z1, candidates C35 and C36 for worker W1, whose most recent task was box preparation, are removed from the task combination table T3. Similarly, candidates C32 and C34 for worker W2 are removed from the task combination table T3.

[0113] The control unit 50 then removes candidates from the task combination table T3 based on the task sequence information D2 and task periods (S12-S13). In the example of Fig. 12, the task combination of candidate C33 for worker W2 corresponds to an abnormal sequence in the task sequence information D2, so candidate C33 is removed from the task combination table T3 (S12). The control unit 50 determines candidate C31, which remains in the task combination table T3 and was not removed, as the task combination of the determination result (S16).

[0114] As described above, in this embodiment, the task tendency information includes the task plan information D4 as an example of information associating each of the workers W1-W3 with a task section (an example of a task range) indicating a range of locations in the workplace where the individual workers W1-W3 do not perform an auxiliary task, which is an example of a second task among a plurality of tasks. When the location of the worker W is included in the task section for the worker W, the control unit 50 excludes from the multiple task combinations task combinations in which the task performed by the worker W is an auxiliary task (S21), and determines one task combination (S16). As a result, a task combination in which the worker W performs the main task in each task section can be determined as a determination result.

[0115] The exclusion based on the work plan information D4 (S21) may use information on various work plans, not just the section in charge, and may be performed using information on the number of items to be carried out, for example. For example, if a worker W has packed items into boxes a number of times corresponding to a predetermined number of items to be carried out, the worker W may be excluded as a candidate for packing from that point onward, as the candidate may no longer be a candidate for packing.

[0116] (Other embodiments) As described above, embodiments 1 and 2 have been described as examples of the technology disclosed in this application. However, the technology in this disclosure is not limited to these and can be applied to embodiments in which appropriate modifications, substitutions, additions, omissions, etc. are made. Furthermore, it is also possible to combine the components described in each of the above embodiments to create a new embodiment. Therefore, other embodiments will be described below as examples.

[0117] In each of the above embodiments, the work sequence information D2 has been described, in which an abnormal sequence is associated with a section in the workplace 6. In this embodiment, the work sequence information D2 may include, for example, a standard sequence indicating the sequence in which tasks are to be performed by one worker W for each section in the workplace 6. In this case, in the exclusion process (S12) based on the work sequence information in the worker identification process (FIGS. 8 and 11), for example, candidates that do not fall under the standard sequence may be excluded from the work combination tables T1 to T3.

[0118] In the above embodiments, the image recognition unit 51 outputs the recognition results of the positions of the workers W1 to W3 without distinguishing them from one another. The image recognition unit 51 of the present embodiment may also distinguish between the workers W1 to W3 and recognize their respective positions using, for example, face recognition technology. Even in this case, the operation of the work analysis device 5 of the present embodiment is applicable when, for example, the work performed by each worker W whose face is not captured in the image captured by the camera 2 can be recognized.

[0119] In the above embodiments, examples have been described in which all workers W at positions recognized by the image recognition unit 51 are workers to be analyzed. The work analysis device 5 of this embodiment may exclude the positions of workers not to be analyzed from the processing targets in the work-related processes (S4 to S5, S8). For example, in the workplace 6, workers in positions other than the work areas A1 and A2 and the area between them, such as an area for replenishment work, are excluded. Furthermore, for example, a worker who manages or monitors processes in the workplace 6 may be excluded as a worker not to be analyzed based on the tendency for the worker's movement line to pass through the area between the work areas A1 and A2 for a long period of time.

[0120] In the above embodiments, the work analysis device 5 performs exclusion based on work period information set in advance (S13). In the work analysis device 5 of this embodiment, the collection work period may be set for each worker based on the length of time the worker has spent in the work area A2 on the shelf 62 side in the past.

[0121] In the above embodiments, examples have been described in which the worker identification process is performed using information such as the work period of each task. In the worker identification process, the work analysis device 5 of this embodiment may perform the worker identification process not only using the work period of each task, but also using information such as the period during which the task is estimated to be performed outside the angle of view of the camera 2, or the worker's rest period.

[0122] Furthermore, in each of the above embodiments, an example has been described in which the task analysis system 1 is applied to a workplace 6 such as a logistics warehouse. In this embodiment, the workplace, i.e., the site, to which the task analysis system 1 and task analysis device 5 are applied is not limited to the above-mentioned workplace 6, but may be various other sites, such as a factory or a store sales floor. Furthermore, the tasks determined by the task analysis system 1 are not limited to the above-mentioned example of packing boxes, but may be various tasks appropriate to various sites. Furthermore, the worker analyzed by the task analysis system 1 is not limited to a person such as worker W, but may be any mobile object capable of performing various tasks. For example, the mobile object may be a robot or various manned or unmanned vehicles.

[0123] As described above, the embodiments have been described as examples of the technology in the present disclosure, and for that purpose, the accompanying drawings and detailed description have been provided.

[0124] Therefore, the components shown in the accompanying drawings and detailed description may include not only essential components for solving the problem, but also components that are not essential for solving the problem in order to illustrate the above technology. Therefore, the fact that these non-essential components are shown in the accompanying drawings or detailed description should not be interpreted as immediately indicating that these non-essential components are essential. [Industrial Applicability]

[0125] The present disclosure is applicable to data analysis applications for analyzing the work of workers in various environments such as logistics sites or factories.

Claims

1. A work analysis device that generates information about a plurality of workers performing a plurality of tasks in a workplace, an acquisition unit that acquires image data representing an image of the workplace; a control unit that generates work history information indicating work performed by each of the plurality of workers at the work site based on the image data; a storage unit for storing the work history information; Equipped with The control unit Sequentially recognizing the positions and tasks of the plurality of workers based on image data collected at each time point in the workplace; Detecting crosstalk between a plurality of flow lines including the positions of the plurality of workers at each time point; When the line crossing is not detected, the work history information is generated by associating the work recognized at each time with the individual worker based on the plurality of flow lines; When the crosstalk is detected, the recognized work is associated with the individual worker based on the recognized work and past work history information. Work analysis device.

2. the storage unit stores work tendency information indicating a tendency for the work to be performed in the workplace; When the crosstalk is detected, the control unit refers to the work tendency information and associates the recognized work with the individual worker. The work analysis device according to claim 1 .

3. When the crosstalk is detected, the control unit calculating a plurality of combinations of a plurality of workers corresponding to the plurality of traffic lines where the traffic congestion has occurred and a plurality of recognized operations; determining one combination from the plurality of combinations based on the work tendency information; According to the determined combination, the recognized work is associated with the individual worker. The work analysis device according to claim 2 .

4. the task tendency information includes information indicating an order in a combination of two or more tasks among the plurality of tasks, The control unit determines the one combination by excluding a combination that falls in a predetermined order from the plurality of combinations according to the order. The work analysis device according to claim 3 .

5. the task tendency information includes information indicating a standard task period set for a first task among the plurality of tasks, The control unit determines the one combination by excluding, from the plurality of combinations, combinations whose duration exceeds the standard work duration, in accordance with a duration during which the work of the worker is recognized as the first work. The work analysis device according to claim 3 or 4.

6. the work tendency information includes information associating a range of responsibility indicating a range of a position in the workplace where the individual worker does not perform a second work of the plurality of works with the individual worker, When the location of the worker is within the worker's assigned area, the control unit excludes a combination in which the worker's task is the second task from the plurality of combinations, and determines the one combination. The work analysis device according to any one of claims 3 to 5.

7. The control unit detects the crosstalk in accordance with a state in which positions of two or more of the plurality of workers are superimposed in an image represented by the acquired image data. The work analysis device according to any one of claims 1 to 6.

8. the storage unit further stores authentication information that identifies the individual worker; The control unit associates the recognized position with the individual worker at the first time the position of the worker in the workplace is recognized. The work analysis device according to any one of claims 1 to 7.

9. The control unit generates information indicating a ratio of the plurality of tasks over a predetermined period for each individual worker based on the task history information for the predetermined period. The work analysis device according to any one of claims 1 to 8.

10. 1. A task analysis method for generating information about a plurality of workers performing a plurality of tasks in a workplace, comprising: The control unit of the computer acquiring image data representing an image of the workplace; generating work history information indicating work performed by each of the plurality of workers at the work site based on the image data; Including, In the step of generating the work history information, the control unit of the computer Sequentially recognizing the positions and tasks of the plurality of workers based on image data collected at each time point in the workplace; Detecting crosstalk between a plurality of flow lines including the positions of the plurality of workers at each time point; When the line crossing is not detected, the work history information is generated by associating the work recognized at each time with the individual worker based on the plurality of flow lines; When the crosstalk is detected, the recognized work is associated with the individual worker based on the recognized work and past work history information. Work analysis method.

11. A program for causing a control unit of a computer to execute the task analysis method according to claim 10.

Images