JP2009110239A - System and method for analysis of working operation - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system and method for analysis of working operation efficiently extracting the problem of an operation. <P>SOLUTION: A data acquisition means 11 of an analysis system 1 for working operation picks up the images of both an experienced operator and an inexperienced operator who perform the same type of operation, and acquires an animation and operation signal. Then, the animation and operation signal are stored in a storage means 14 so as to be associated with the validity/invalidity of a product manufactured by the operation. Then, a difference point extraction means 15 compares an operation signal (non-defective operation signal)of when the product is satisfactory and an operation signal (defective operation signal) of when the product is defective to extract a difference point. Then, a display means 16 is made to display an animation associated with the difference point. Thus, it is possible to efficiently extract the problem of the operation of the inexperienced operator. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a work motion analysis system and a work motion analysis method, and more particularly to a work motion analysis system and a work motion analysis method for extracting problems by comparing the work of a plurality of workers.

  Even now, in the manufacturing process of many products, manual work by skilled workers is required. However, since the skills of skilled workers are difficult to grasp objectively and quantitatively and difficult to convey to others by verbal or written documentation, skill transfer is a major problem in many manufacturing industries. Yes.

  Therefore, conventionally, there has been proposed a method for objectively and quantitatively grasping an operator's work by comparative analysis using video images. For example, in Patent Document 1, marks of various colors are attached to each part of a worker's body, the worker is allowed to work in this state, and the work is imaged from a plurality of angles by a plurality of image sensors. A technique is disclosed in which three-dimensional coordinate data of each target point is created by combining the captured results, and the result is printed out together with standard work data. Thereby, an operator's operation | movement can be analyzed quantitatively.

  However, the above-described conventional technique has a problem that analysis work takes a long time. In addition, it is difficult to extract which part of the measured operation affects the quality of the product. For this reason, the above-described conventional technique has little to obtain for the required labor and time, and is not implemented in practice.

JP-A-4-259803

  An object of the present invention is to provide a work motion analysis system and a work motion analysis method capable of efficiently extracting work problems.

  According to one aspect of the present invention, data acquisition means for acquiring a moving image of a worker who is working and a motion signal representing a change in a feature amount of the movement of the worker, and the moving image of a plurality of the workers And storage means for storing the operation signal in association with the quality of the work result, the operation signal when the work result is good, and the operation signal when the work result is bad. There is provided a work motion analysis system comprising: a difference extraction unit that extracts differences by comparing, and a display unit that displays the moving image corresponding to the difference.

  According to another aspect of the present invention, for each of a plurality of workers performing the same type of work, an operation signal representing a change in the moving image of the worker and the feature amount of the operation of the worker is acquired. A process, a step of storing the moving image and the operation signal in association with the quality of the work result, and the operation signal when the work result is good and the result of the work are poor. A work operation analysis comprising: a step of extracting a difference by comparing with the operation signal; and a step of extracting the cause of the defect by displaying the moving image corresponding to the difference. A method is provided.

  According to the present invention, it is possible to realize a work motion analysis system and a work motion analysis method that can efficiently extract work problems.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
First, a first embodiment of the present invention will be described.
FIG. 1 is a block diagram illustrating a work motion analysis system according to this embodiment.
As shown in FIG. 1, in the work motion analysis system 1 according to the present embodiment, a data acquisition unit 11 is provided. The data acquisition unit 11 is provided with a moving image acquisition unit 12 that acquires a moving image of a worker who is working, and an operation signal acquisition unit 13 that acquires an operation signal of the worker.

  In this specification, “operation” means, for example, a group of operations performed to achieve a specific purpose among operations of an operator engaged in a product manufacturing process. For example, a tool is returned to an original place. Such actions include returning, pushing or stepping on a button, opening the toolbox door to remove a screw, and the like. Further, the “motion signal” refers to a signal obtained by extracting a change in a certain feature amount from the motion of the worker accompanying the work. The “feature amount” is an amount that can be expressed by one numerical value, such as the height of each part of the worker's body and the angle of the elbow. For example, a mark is attached to the worker's body. Measured by. The “change in feature amount” is a change in feature amount over time. Therefore, in one example, the “motion signal” is a signal representing a change in height of the mark attached to the right wrist of the worker.

  The moving image acquisition unit 12 may acquire a worker's moving image from a plurality of angles, and the operation signal acquisition unit 13 may acquire a plurality of types of operation signals. In one example, the moving image acquisition means 12 is a video camera. The operation signal acquisition unit 13 is an image analysis device that extracts the coordinates of the target from the moving image acquired by the moving image acquisition unit 12. Thereby, the data acquisition means 11 acquires a moving image and an operation signal about the worker who is working.

  In the work motion analysis system 1, a storage unit 14 is provided. The storage unit 14 stores the moving image and the operation signal acquired by the data acquisition unit 11 in association with the quality of the work result. The “work result” is, for example, a product manufactured by the work to be analyzed, and the “work result is good” is, for example, the quality judgment result of this product. That is, the storage unit 14 stores a moving image and an operation signal of a work for manufacturing a non-defective product, and a moving image and an operation signal of a work for manufacturing a defective product. The storage unit 14 is a storage device such as an HDD (Hard Disk Drive).

  Further, the work motion analysis system 1 is provided with a difference extraction means 15. The difference extraction means 15 has an operation signal when the work result is good among the data stored in the storage means 14 (hereinafter referred to as “good product operation signal”) and the work result is bad. The operation signal (hereinafter referred to as “defective operation signal”) is compared to extract the difference. The difference extraction means 15 is configured by, for example, a CPU (Central Processing Unit).

  Various methods are conceivable as methods for comparing operation signals and extracting differences. For example, a difference between a non-defective product operation signal and a defective product operation signal may be taken, and a portion where the absolute value of the difference is larger than a predetermined reference value may be set as a difference. Further, by multiway PCA (Principal Component Analysis), a portion having high correlation with quality may be extracted as a difference using a plurality of variables having different dimensions.

  Furthermore, the work motion analysis system 1 is provided with a display unit 16 that displays a moving image stored in the storage unit 14 in accordance with an instruction from the difference extraction unit 15. The display means 16 is a display such as an LCD (Liquid Crystal Display). As will be described later, the display unit 16 displays a moving image corresponding to the difference extracted by the difference extraction unit 15.

Next, the operation of the work motion analysis system 1 according to the present embodiment configured as described above, that is, the work motion analysis method according to the present embodiment will be described.
FIG. 2 is a flowchart illustrating the work operation analysis method according to this embodiment.
FIGS. 3A and 3B are graphs illustrating the waveforms of operation signals, with time on the horizontal axis and feature quantities on the vertical axis, and FIG. 3A shows the waveforms of skilled workers. , (B) shows the waveform of an unskilled worker,
FIG. 4 is a graph illustrating a method for extracting differences by taking time on the horizontal axis and taking the absolute value of the feature amount difference on the vertical axis.

  First, as shown in step S <b> 1 of FIGS. 1 and 2, the data acquisition unit 11 of the work movement analysis system 1 acquires a moving image and a movement signal of work by a skilled worker. Specifically, for example, a plurality of marks colored in different colors are attached to a plurality of parts of a skilled worker's body. Note that there may be one mark. Then, in this state, a skilled worker performs work to be analyzed, and this is captured by a video camera as the moving image acquisition means 12 to acquire a moving image. In addition, the motion signal acquisition unit 13 performs image processing on the moving image, thereby acquiring a feature amount, for example, coordinate data of a target point in time series as an operation signal. As shown in FIG. 3A, this operation signal represents a temporal change in the feature value, and forms a constant waveform with respect to time.

  Next, as shown in step S2, the moving image and the operation signal acquired by the data acquisition unit 11 are stored in the storage unit 14 in association with the quality of the work result. For example, the moving image and the operation signal are stored by associating the result of the quality determination of the product manufactured by the work measured in step S1. At this time, the moving image and the operation signal are also stored in association with each other. The association between the moving image and the operation signal is possible by matching the time axes of both data. Normally, a product manufactured by a skilled worker is a “good product”.

  On the other hand, as shown in step S3, the data acquisition unit 11 acquires a moving image and an operation signal of work by an unskilled worker. The acquisition conditions at this time are the same as in step S1. That is, the unskilled worker is caused to perform the same type of work in the same work environment as the work by the skilled worker in step S1, and the moving image and the operation signal are acquired. At this time, the waveform of the operation signal is, for example, as shown in FIG. And as shown to step S4, this moving image and an operation signal are linked | related with the quality of the result of work, and are memorize | stored in the memory | storage means 14. FIG. Normally, a product manufactured by an unskilled worker is “defective”.

  Note that the process of steps S1 and S2 and the process of steps S3 and S4 are not limited, and either may be performed first. Further, each worker may be caused to repeatedly execute the same work a plurality of times, or a plurality of skilled workers and a plurality of unskilled workers may perform the same type of work. Thereby, a plurality of sets of moving images and operation signals can be acquired.

  Next, as shown in step S5, the difference extraction means 15 performs an operation signal (good product operation signal) when the work result is good and an operation signal (defective) when the work result is bad. And the difference is extracted. That is, a work part estimated to have a high possibility of affecting the quality of the product is extracted by calculation processing. In general, the non-defective operation signal is the signal acquired in step S1, and the defective operation signal is the signal acquired in step S3.

  For example, as shown in FIG. 4, the difference between a non-defective product operation signal and a defective product operation signal is taken, and a time range in which the absolute value of the difference is larger than a predetermined reference value is set as a difference D. Further, for example, when a plurality of types of operation signals are acquired in steps S1 and S3, a portion in which the absolute value of the difference is larger than a predetermined reference value is extracted from the plurality of operation signals. D. As a result, the portion of the entire work that greatly affects the work result is limited temporally or temporally and spatially. When a plurality of sets of operation signals are acquired in steps S1 to S4, the difference between the average value of the plurality of non-defective operation signals and the average value of the plurality of defective operation signals may be obtained. In addition, instead of comparing the absolute value of the difference with the reference value, multiway PCA etc. is used to analyze using multiple variables with different dimensions, and extract parts that are highly correlated with quality. It is good.

  Next, as shown in step S <b> 6, the difference extraction unit 15 selects a moving image corresponding to the difference D extracted in step S <b> 5 among the moving images of the skilled worker and the unskilled worker stored in the storage unit 14. It is displayed on the display means 16. In other words, a time range corresponding to the difference D, or a moving image of the time range and the spatial range is displayed. Then, by verifying this moving image, the difference between the work of the skilled worker who manufactured the non-defective product and the work of the unskilled worker who manufactured the defective product is detected. Thereby, the problem of the work by the unskilled worker is extracted.

Next, the effect of this embodiment will be described.
As described above, in the present embodiment, work improvement points can be presented by extracting problems of work by unskilled workers. Then, while conscious of this improvement point, by correcting the operation of the unskilled worker and bringing it closer to the operation of the skilled worker, the skill of the skilled worker can be passed on to the unskilled worker. Thus, according to the present embodiment, it is possible to objectively and quantitatively grasp the points of work that requires skill, and to improve the efficiency of skill transfer.

  Further, according to the present embodiment, the difference D is extracted based on the operation signal, and only the moving image corresponding to the difference D is displayed to detect the difference in operation between the skilled worker and the unskilled worker. Therefore, it is only necessary to verify a moving image in a temporally or temporally and spatially limited region. As a result, it is possible to efficiently extract the problem of work by an unskilled worker in a short time compared to the case of verifying the entire moving image.

Next, a second embodiment of the present invention will be described.
FIG. 5 is a block diagram illustrating a work operation analysis system according to this embodiment.
As shown in FIG. 5, in the work motion analysis system 2 according to the present embodiment, in addition to the configuration of the work motion analysis system 1 according to the first embodiment described above, a standard that normalizes the time axis of the motion signal. A conversion means 17 is provided. Then, the difference extraction unit 15 compares the standardized operation signals and extracts the difference. The standardization means 17 is constituted by a CPU, for example. In addition, the normalization means 17 and the difference extraction means 15 may be comprised by the same CPU. The other configuration of the work motion analysis system 2 according to the present embodiment is the same as that of the work motion analysis system 1 according to the first embodiment described above.

Next, an operation of the work motion analysis system according to the present embodiment, that is, a work motion analysis method will be described.
FIG. 6 is a flowchart illustrating the work operation analysis method according to this embodiment.
As shown in FIG. 6, in this embodiment, as shown in steps S1 to S4, the data acquisition unit 11 acquires a moving image and an operation signal for the work of the skilled worker and the unskilled worker, After being stored in the storage means 14 in association with pass / fail, the operation signal is normalized as shown in step S7. Hereinafter, the normalization process will be described in detail.

FIG. 7 is a diagram illustrating an operation signal when three workers perform three tasks in succession. In the graph, the horizontal axis represents time, and the vertical axis represents a feature amount. ing.
FIGS. 8A to 8C are graphs illustrating the operation signal of the work 1 of each worker shown in FIG. 7, in which the horizontal axis represents time and the vertical axis represents the feature amount. Yes.
FIGS. 9A to 9C are graphs illustrating the results of normalizing the time axis of the operation signals shown in FIGS. 8A to 8C, and the horizontal axis of each figure represents time. The vertical axis represents the feature amount.
FIGS. 10A to 10C are diagrams exemplifying operation signals when three workers perform three tasks in succession with time on the horizontal axis and feature amounts on the vertical axis. (A) shows an operation signal before normalization, (b) shows an operation signal after normalization, and (c) shows a case where the operation signals shown in (b) are connected.

  In the above-described first embodiment, the case where the time required for work is equal among workers has been described. However, in the case of work in which the timing of work is not regulated from the outside, the time required for work may be different for each worker or even for the same worker. In such a case, the operation signals cannot be simply compared. Therefore, in the present embodiment, the time axis of the operation signal is standardized so that the time length of the operation signal is made constant. Then, the normalized operation signals are compared.

  This will be specifically described below. As shown in FIG. 7, when three workers are allowed to perform a series of tasks whose work timing is not regulated from the outside, the time required for the entire work differs for each worker. Therefore, the motion signal is divided into a plurality of unit operations such as a tact unit, for example, by comparing the moving image with the operation signal and grasping the contents of the operation by the moving image. For example, in the example illustrated in FIG. 7, a series of work is divided into three unit works of “work 1”, “work 2”, and “work 3”. In addition, when the worker performs the work 1, the work 2, and the work 3, a means for transmitting a signal may be provided so that the operation signal is automatically divided into unit work.

Next, as shown in FIGS. 8A to 8C, the time required for the work 1 of each worker is measured. As a result, for example, the time required for the work 1 of three workers is assumed to be t ₁ , t ₂ , and t ₃ , respectively. On the other hand, a reference time T1 for operation ₁ is set. The length of the reference time T ₁ is not particularly limited. For example, the reference time T ₁ is set to a time longer than the longest time among the times t ₁ , t ₂ , and t ₃ .

Next, as shown in FIGS. 9A to 9C, the time axes of the operation signals when the three workers perform the work 1 are respectively (T ₁ / t ₁ ) times (T ₁ / t ₂ ) times and (T ₁ / t ₂ ) times. Thus, the temporal length of the work 1 in the operation signal of each worker is unified to the reference time T _1.

Then, as shown in FIGS. 10A and 10B, the operation signals for the work 2 and the work 3 also normalize the time axis in the same manner as the work 1. Expressing the above processing in general, when the time required for the j-th unit work by the i-th worker is t _ij and the reference time of the j-th unit work is T _j , the i-th work The time axis of the operation signal when the person performs the j-th unit work is multiplied by (T _j / t _ij ). Thereby, the time length of the operation signal of the j-th unit work by each worker is unified and standardized to the reference time T _j .

  Next, as shown in FIG. 10C, the standardized operation signals of the work 1 to the work 3 are connected to each worker. As a result, an operation signal is generated which integrally represents a series of operations and whose time axis is standardized for each unit operation. Thereby, the normalization of the operation signal is completed.

Then, as shown in step S5 of FIG. 6, the difference extraction means 15 shown in FIG. 5 compares the standardized non-defective product operation signal with the standardized defective operation signal, and extracts the difference. At this time, the work time t _ij measured in step S7 is also referred to as a variable representing the work time. Subsequent processing is the same as in the first embodiment. That is, as shown in step S6, a moving image corresponding to the difference is displayed on the display means 16, and a problem in the operation is extracted.

Next, the effect of this embodiment will be described.
In the present embodiment, by standardizing the time axis of the operation signal, it is possible to compare work with different work times. The effects of the present embodiment other than those described above are the same as those of the first embodiment described above.

  In this embodiment, in step S7, not only the time axis but also the space axis may be normalized. That is, you may arrange the movable range of a certain mark between workers. Thereby, the difference of the operation signal due to the physique difference of the worker can be eliminated, and the difference can be extracted with higher accuracy.

  In step S7 described above, not only the time axis of the operation signal but also the time axis of the moving image may be standardized. Thereby, in step S6, the moving image which imaged the skilled worker and the moving image which imaged the unskilled worker can be displayed so that each worker's operation | movement may become the same speed. As a result, it becomes easy to compare the operations of both, and it becomes easy to detect the difference.

  Furthermore, in step S7 described above, instead of normalizing the time axis of the operation signal, an appropriate number of variables may be generated by equally dividing each operation signal of the same tact. Thereby, the same number of variables can be obtained from operation signals having different lengths. In step S5, analysis of multiway-PCA or the like is performed using these variables. In this case, the time required for each tact, that is, the time required for each unit work may be added as one variable.

  The present invention has been described above with reference to the embodiment. However, the present invention is not limited to this embodiment. For example, those in which the person skilled in the art appropriately added, deleted, or changed the design of the above-described embodiments, or those in which processes were added, omitted, or modified in conditions also included the gist of the present invention. As long as it is provided, it is included in the scope of the present invention. Further, in each of the above-described embodiments, the example of analyzing the work in the manufacturing process of the product has been shown. However, the present invention is not limited to this, and the present invention can be applied to transmission of any skill. For example, the present invention can also be used for passing down sports competition skills.

1 is a block diagram illustrating a work operation analysis system according to a first embodiment of the present invention. It is a flowchart figure which illustrates the operation | work operation | movement analysis method which concerns on 1st Embodiment. (A) And (b) is a graph which takes time on a horizontal axis and takes a feature-value on a vertical axis | shaft, and illustrates the waveform of an operation signal, (a) shows the waveform of a skilled worker, b) shows the waveform of an unskilled worker. It is a graph which illustrates the extraction method of a difference, taking time on a horizontal axis and taking the absolute value of the difference of a feature-value on a vertical axis | shaft. It is a block diagram which illustrates the work movement analysis system concerning a 2nd embodiment of the present invention. It is a flowchart figure which illustrates the work operation | movement analysis method which concerns on 2nd Embodiment. It is a figure which illustrates the operation signal at the time of three workers performing three work continuously, the horizontal axis of the graph figure in a figure expresses time, and the vertical axis expresses the feature-value. (A) thru | or (c) is a graph which illustrates the operation signal of the operation | work 1 of each worker shown in FIG. 7, the horizontal axis of each figure represents time, and the vertical axis | shaft represents the feature-value. (A) thru | or (c) is a graph which illustrates the result which normalized the time axis | shaft of the operation signal shown to Fig.8 (a) thru | or (c), the horizontal axis of each figure represents time and the vertical axis | shaft Represents a feature amount. (A) thru | or (c) is a figure which illustrates the operation signal at the time of a horizontal axis | shaft, taking a feature-value on a vertical axis | shaft, when three workers performed three work | work continuously. (A) shows an operation signal before normalization, (b) shows an operation signal after normalization, and (c) shows a case where the operation signals shown in (b) are connected.

Explanation of symbols

1, 2 Work motion analysis system, 11 Data acquisition means, 12 Movie acquisition means, 13 Motion signal acquisition means, 14 Storage means, 15 Difference extraction means, 16 Display means, 17 Normalization means, D Difference

Claims (4)

Data acquisition means for acquiring a motion signal representing a change in a feature amount of a motion video of the worker who is working and the worker;
Storage means for storing the moving image and the operation signal for a plurality of the workers in association with the quality of the result of the work;
A difference extraction means for extracting a difference by comparing the operation signal when the result of the work is good and the operation signal when the result of the work is bad;
Display means for displaying the video corresponding to the difference;
A work motion analysis system characterized by comprising: Further comprising normalization means for normalizing the time axis of the operation signal;
The work motion analysis system according to claim 1, wherein the difference extraction unit compares the standardized motion signals. For each of a plurality of workers doing the same type of work, obtaining a motion signal representing a change in the motion video of the worker and a feature amount of the worker's motion;
Storing the moving image and the operation signal in association with the quality of the result of the work;
Comparing the operation signal when the result of the work is good and the operation signal when the result of the work is bad, and extracting a difference;
Displaying the video corresponding to the difference and extracting the cause of the defect;
A work motion analysis method characterized by comprising: Further comprising the step of normalizing the time axis of the operation signal,
4. The work motion analysis method according to claim 3, wherein in the step of extracting the difference, the standardized motion signals are compared with each other.

Images