JP2003117045A - Swing form diagnosing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To secure a determining method of a specific motion point which is very important for diagnosing the swing form and to extract only an effective image. SOLUTION: This swing form diagnosing device comprises an input means for inputting animated images of a subject picked up by an image picking-up means, an extracting means for extracting frames relating to a series of swing motions of the subject from the inputted animated images, a difference operating means for comparing each frame in the swing data with a standard image and outputting a difference image, a determining means for determining a specific motion point in the swing motion in accordance with a result of the analysis of a moving part image in the difference image, and a display means for extracting the frames corresponding to every determined specific motion point and displaying the image.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a frame (still image) at an important time point (specific operation point) during a swing motion from a moving image (continuous image) of, for example, a golf swing of a subject photographed by a photographing means. The present invention relates to a swing form diagnosis apparatus capable of easily and effectively diagnosing a swing form of a subject by extracting and displaying an image.

[0002]

2. Description of the Related Art With the recent widespread use of home video cameras and the like, it is common to take a video of a golf swing of a subject and check the swing form while watching the reproduced image. Has become
In such a method, since the image displayed on the monitor is a moving image, it is difficult to check the swing form in detail. Therefore, a more advanced swing form diagnosis device is required.

The invention disclosed in Japanese Unexamined Patent Publication No. 8-196677 extracts a frame at a predetermined time during a swing motion based on a weight movement state accompanying a golf swing, and specifically, it applies to the left and right legs of a subject. By using a reaction force meter that can detect the weight movement state by measuring the load individually, the start point of the rapid decrease of the weight movement waveform (corresponding to the top) and the time point when the weight movement waveform becomes maximum (impact Impact using an optical sensor that detects when the golf ball has been hit and has disappeared from the set position, as well as determining when the weight movement waveform becomes zero (corresponding to when turning back) before The time is determined and the frames at each of the above-mentioned time points are extracted to diagnose the swing form of the subject.

However, in the above invention, in view of the fact that the swing form varies from person to person and the weight movement waveforms vary from subject to subject (especially remarkable for beginners who have not completed the form). There is no guarantee that each time point (excluding the impact time determined by the optical sensor) coincides with the true specific operating point.
As a result, an inappropriate frame (frame at a time point other than the specific operation point) is extracted, and the swing form may not be properly diagnosed.

The invention disclosed in Japanese Patent Application Laid-Open No. 7-213672 is to extract a frame at a constant timing from a predetermined range before and after the impact time, specifically, a photoelectric sensor for detecting an address state of a subject. By using a sensor and an impact sensor that detects a state where the golf ball has disappeared from the tee, the range is specified from the detection time of the photoelectric sensor to the detection time of the impact sensor and from the detection time of the impact sensor to, for example, one second. The frames are extracted at regular intervals within these ranges to diagnose the swing form of the subject.

However, in the invention, in the process from the state where the subject enters the address state to the time when, for example, 1 second elapses after the golf ball is hit, a constant interval around the moment when the golf ball is hit. Since multiple frames of each frame are extracted as decomposed photos (continuous photos), if you set a small number of frames to be extracted, the frames at a specific operating point may not be extracted, but the number of frames to be extracted is large. If set, a problem arises in that a large number of useless frames that are not the frames at the specific operation point are extracted, and in any case, swing form diagnosis may not be properly performed.

The invention disclosed in Japanese Unexamined Patent Publication No. 8-257191 is to extract frames that meet a predetermined condition from a predetermined range before and after the impact time, specifically, to analyze a captured image. Are sequentially performed starting from the final image, and the time when the golf ball is analyzed to appear (that is, before hitting) is determined to be the impact time, and then the predetermined time is traced back from that time. While the address is determined to be the address time, the time point after a predetermined period from the impact time is determined to be the follow-through time, and further, as a result of the image analysis, the time point when the head speed of the club head changes from the front to the back is determined to be the top time. The frames at these times are extracted to diagnose the swing form of the subject.

In this invention, however, there is no problem at the time of impact and at the time of top because it is recognized that it coincides with the true specific operating point. It depends on the subject (especially the head speed of the club head), and in consideration of the fact that the "interval between the impact and the time period", which is the condition for determining these points, differs from subject to subject, it always matches the true specific operating point. There is no guarantee. That is, the frame at the specific operation point is reliably extracted only at the time of impact and at the time of top,
Frames at other specific operating points may not be properly extracted, and it may happen that the swing form diagnosis cannot be properly performed in the same manner as in the above invention.

In short, in any of the inventions described in the above publications, the method of determining a specific operating point, which is extremely important for diagnosing a swing form, is not accurate. It is difficult for the diagnostician (subject in the case of self-diagnosis) to extract the frame that is intended, or there is a high possibility that an unintended frame is extracted. As an ideal golf swing form diagnosis, it is possible to extract frames at multiple specific operating points such as address, takeback, top, downswing, impact, follow through, and finish. Although preferred, the present invention provides a swing form diagnostic device that meets this need.

[0010]

That is, a swing form diagnosis apparatus according to the present invention has an input means for inputting a moving image of a subject photographed by the photographing means and an inputted moving image as described in claim 1. Extraction means for extracting, as swing data, a frame relating to a series of swing motions performed by the subject, difference calculation means for comparing each frame in the swing data with a reference image, and outputting a difference image, and a difference image And a display unit for extracting a frame corresponding to each of the determined specific motion points and displaying the image according to the analysis result of the moving partial image. It is characterized by

According to the swing form diagnosis apparatus having the above-mentioned structure, the frame relating to a series of swing motions (from address to finish) performed by the subject is selected from the moving image of the subject composed of a plurality of frames arranged in time series. It is extracted and used as swing data (moving image related to swing motion). In the swing data, a frame at a specific operation point useful for diagnosing a swing form and a frame not so are mixed, and even if it is useful, the frame includes many overlapping portions. It is an object of the present invention to remove the non-useful frames and the useful but redundant frames by using the difference image processing.

As the specific operation point, generally, a characteristic time point in a series of swing operations, for example, the top time,
It can be used at impact, finish, etc., but in addition to this, at address, during takeback from the start of swing motion to top, during downswing from top to impact, after finish after impact Examples include follow-throughs up to the time. Among these, especially at impact, it is important for diagnosis of swing form.

When a plurality of specific operating points are determined, a frame corresponding to the specific operating points is extracted from the swing data and displayed as an image. Even if the frame is not the frame of the specific operation point or the frame of the specific operation point, a frame (noise) unnecessary for diagnosis such as an overlapping portion is excluded. Therefore, only the minimum necessary image can be obtained as an output from the moving image of the subject, and the swing form of the subject can be easily and effectively diagnosed.

In extracting the swing data for the golf swing, as described in claim 2, the frame at the time when the golf ball is lost by hitting is subjected to the difference image processing and the template matching processing for the golf ball. Alternatively, it is possible to employ a method of extracting as swing data by using an impact sensor and extracting from a frame that is a predetermined number before the reference frame to a frame that is a predetermined number after the reference frame.

The "frame which is a predetermined number before the reference frame" is a frame in which the subject is presumed to be in the address state (the start point of the swing motion), and the "predetermined number" is the starting point of the swing data. Is a condition for setting. The "frame after a predetermined number of reference frames" is a frame in which the subject is estimated to be in the finish state (at the time of completion of the swing motion), and this "predetermined number" is
It is a condition for setting the end point of the swing data. These values are obtained from empirical rules and can be changed arbitrarily.

As another swing data extraction method, as described in claim 3, the frame in which the subject is in the address state is subjected to the difference image processing or the template matching processing for the subject or the golf club, or It is possible to adopt a method of extracting as swing data by using an address sensor, and extracting from the reference frame to a frame after the reference frame by a predetermined number.

In this method, the reference frame is set as a starting point. The “frame after a predetermined number of reference frames” is a frame in which the subject is estimated to be in the finish state (at the time of completion of the swing motion), and the “predetermined number” is used to set the end point of the swing data. It becomes a condition. This value is derived from experience,
It can be changed arbitrarily.

However, where to set the reference frame can be arbitrarily changed. In short, as long as the starting point and the ending point of the swing data can be set, the reference frame may be a frame of a specific operation point other than at the time of impact or address, or a frame other than the specific operation point.

In determining the specific operating point for the golf swing, the analysis result that the moving partial image is hardly or not extracted from the difference image in the golf swing is obtained, and the address of the specific operating point is determined. Time, top time or finish time (claim 4), a specific action is obtained by obtaining an analysis result that there is little or no lateral displacement of the moving partial image extracted from the difference image in the golf swing. Among the points, there is a method of determining a take-back time, a down-swing time, or a follow-through time (claim 5), and in the golf swing, there is little or no vertical displacement of the moving part image extracted from the difference image. A method of determining the impact time of specific operating points by obtaining the analysis result (claim 6) It can be adopted.

Among the specific operating points, at the time of address, at the time of top, and at the time of finish, due to the fact that the subject and the golf club remain stationary for a certain time, little or no moving partial image is extracted from the difference image. At the time of impact, a moving partial image is extracted from the difference image, but this moving partial image has little or no vertical displacement amount. By performing image analysis on the moving partial image in the difference image using such an event as a criterion, it is possible to determine each specific operating point. For takeback, downswing, and follow-through, the moving part image is extracted from the difference image, but at the time when this moving part image has little or no lateral displacement, the takeback, downswing is performed. Time or follow-through time.

Then, as described in claim 7, the difference image processing or the template matching processing may be performed only on a partial area of a frame. This can reduce the processing load and the processing time. Examples of the partial area include an area including the set position of the golf ball at the time of addressing and impact, and a golf club (club head) at the time of takeback, downswing and follow through.
And an area including the subject's arm. These are based on the fact that the position of the subject's arm or golf club at each specific motion point is specified due to the nature of the swing motion.

Further, in the determination of the specific operating point, as described in claim 8, the reference image is a frame which is one frame before in the swing data, that is, an arbitrary frame in the swing data and one frame before that. It is possible to adopt a configuration in which the difference image is output by comparing with the frame of. In such a case, it is not necessary to prepare a background image in which the subject is not photographed in advance, and image processing based on only the swing data becomes possible.

Alternatively, the reference image is a background image in which the subject is not photographed, that is, each frame in the swing data is compared with one background image and a difference image is output. A configuration can be adopted. In such a case, the subject, the golf club, and the golf ball can be image-processed separated from the background. The separated and extracted images of the subject, etc. are combined with other specific operating point frames, the frames at the time of personal best collected and stored previously, and the frames of professional golfers prepared in advance as master data. Overlaid display can be performed. As a result, the golf swing to be diagnosed can be compared with the golf swing of the personal best or the golf swing of the professional golfer, and more appropriate swing form diagnosis can be carried out.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of a swing form diagnosis apparatus according to the present invention will be described below with reference to the drawings.

FIG. 1 shows the system configuration of the swing form diagnosis apparatus 1. This swing form diagnosis apparatus inputs a video camera (photographing means) 2 for photographing a golf swing of a subject S and a personal computer (processing means) 3 for inputting a moving image photographed by the video camera 2 and performing image processing such as image analysis. A display (first output means) 4 for displaying a captured moving image and an image edited based on the captured moving image as a swing form diagnostic material, and a printer for printing a swing form diagnostic material created by image processing (first Two output means) 5.

The video camera 2 is a digital video camera equipped with an image pickup device composed of a CCD. For example, a region including the golf ball B set on the tee T and the subject S hitting the golf ball B is in front of the subject S. It is provided to shoot from the side, for example 3
It is configured such that each frame photographed by 0 frames is sequentially output in a moving image format of MPEG2 format. still,
A frame is an arbitrary still image that constitutes a moving image.

The moving image from the video camera 2 is recorded on the personal computer 3.
It is given to the primary processing unit 32 through the input processing unit 31 of. The primary processing unit 32 includes a buffer memory 321 such as a ring buffer memory capable of storing a frame given every 1/30 seconds for a predetermined time (for example, 7 seconds = 210 frames), and the buffer memory 321.
On the other hand, each time a new frame (image data) is input, the oldest frame is erased and new frames are sequentially stored.

Further, the primary processing section 32 carries out template matching processing of the golf ball B for each frame. The presence or absence of the golf ball B can be determined by comparing the difference between the background color in each frame and the previously created template of the golf ball.

In the template matching process, the fact that the golf ball B has disappeared indicates that the time point is the impact of the swing motion, and among the plurality of frames stored in the buffer memory 321. It shows that there is image data related to a series of swing motions performed by the subject S. The primary processing unit 32 triggers the change of the golf ball B from the state “present” to the state “absent”,
An operation of extracting frames before and after the frame as image data (swing data) related to a series of swing motions is performed.

Specifically, a predetermined time (for example, 4 seconds before = 12) is set with the time point when the golf ball B is exhausted as a reference point.
Since it can be inferred that the subject is in the addressing state (0 frame before), the subject makes a swing motion after a predetermined time (for example, after 1.5 seconds = 45 frames) from the reference point while the point is the starting point. Since it can be assumed that it is in a completed state, this point is the end point. Then, the frame between the start point and the end point (for example, 5.5 seconds = 165 frames) is
The swing data includes images of a series of swing motions from the address to the finish.

The extracted swing data and the frame number at the reference point are given to the secondary processing unit 33. The secondary processing unit 33 includes a format conversion unit 331 that converts swing data into an AVI moving image format, and a buffer memory 332 that can store all of the converted swing data. Therefore, in the buffer memory 332 of the secondary processing unit 33, among the frames stored in the buffer memory 321 of the primary processing unit 32, a frame before the swing data (for example, an image in which the golf ball B is set, Only useful image data in a state in which surplus data such as an image in which a swing is performed) and a frame after the swing data (image in a state in which the swing motion is completely completed) are removed is stored.

The secondary processing unit 33 also performs a process of determining a plurality of specific operating points based on the swing data. As the specific operation points, eight points are adopted in the order of swing operation: address, takeback, top, downswing, first impact, second impact, follow through, and finish. . The impact time is divided into the first impact time (the moment when the club head H hits the golf ball B or immediately before that) and the second impact time (immediately after the golf ball B is hit) every second that can be shot by the video camera. Due to the number of frames, club head H
This is because it is difficult to reliably capture the moment when the ball hits the golf ball B.

The determination process of the specific operating point is performed by using the differential image processing method. The difference image processing method is a moving part after performing comparison processing of a reference image and an image to be compared to generate a difference image, and then performing preprocessing such as binarization processing and smoothing processing on the difference image. The amount of displacement of the moving part image is analyzed by extracting the image and determining the area of this moving part image. Specifically, the Nth frame is extracted as a reference frame, the (N + 1) th frame is extracted as a contrasting frame to generate a difference image between the two, and then a moving partial image in the difference image is extracted. By obtaining the area of the moving part image, the displacement amount in the X direction (horizontal direction) and the Y direction (longitudinal direction) of the moving part image is analyzed.

During takeback, downswing, impact, follow-through, and finish, the swing motion changes, so the difference image between frames has a relatively white portion indicating the changed portion. There are many. On the other hand, at the time of addressing, at the time of top, and at the time of finishing, almost no change in swing motion is observed, so the difference image between frames becomes black, indicating that there is no change. Based on this, the criteria for determining each specific operating point are as follows. The white and black image colors may be changeable by setting. That is, the difference image processing cannot be performed only based on the combination of white and black image data.

At address: At an arbitrary point in time when the subject S remains stationary for a certain time or more, going back from the reference point, that is, as a result of the differential image processing in the preceding stage of the swing data, a predetermined frame ( It is an arbitrary point in time when the displacement amounts in the X direction and the Y direction of the moving partial image are both minimal over two frames.

At the time of takeback: When the shaft of the golf club becomes horizontal after addressing, that is, when the displacement amount of the moving portion image in the X direction becomes minimum as a result of the difference image processing in the preceding stage portion of the swing data. Based on the empirical rule that the subject (right-handed) always swings his / her arm to the left when the swing motion is viewed from the front, the takeback time is the time when the arm (or golf club) extends to the left the most. It is something to do with time).

At the time of the top ... After the takeback, the subject S remains stationary for a certain time or more, that is, as a result of the differential image processing in the preceding stage of the swing data, a predetermined frame (for example, 2 frames). Is an arbitrary point in time when both the displacement amounts of the moving partial image in the X direction and the Y direction are minimal.

At the time of downswing: a time point before a predetermined time from the reference point, that is, a time point before a predetermined frame (for example, one frame) from the reference point in the middle section of the swing data (the difference image processing method is not used is the takeback). This is because the moving speed of the shaft is too fast compared to the time).

At the time of the first impact: The time immediately before the reference point, that is, the time one frame before the reference point in the middle part of the swing data.

At the time of the second impact ... It is a reference point.

At the time of follow-through ... A time point after a fixed time from the reference point, that is, a time point after a predetermined frame (for example, 3 frames) from the reference point in the middle stage of the swing data (the difference image processing method is not used is the take-back). This is because the moving speed of the shaft is too fast compared to the time).

At the finish time ... After the follow-through time, the subject S remains stationary for a certain time or more, that is, as a result of the differential image processing in the latter part of the swing data, a predetermined frame (for example, 10 It is an arbitrary point in time when both the displacement amounts in the X direction and the Y direction of the moving partial image are minimal over the frame).

When each specific operation point is determined, the secondary processing unit 32 carries out an operation of adding additional information (tag) indicating that it is any specific operation point to the frame at each specific operation point. . This is called keyframe extraction. The secondary processing unit 32 also performs a time calculation from the one specific operation point to the previous specific operation point, and the additional information of each key frame includes the time information up to the previous specific operation point. Is also included.

The swing data in a state where the additional information is added to the corresponding key frame is given to the output processing unit 33. The output processing unit 34 includes a monitor memory 341 capable of storing all swing data. In addition to the display 4 and the printer 5, a storage device 6 such as a hard disk drive, a CD-R drive, and an FD drive is connected to the output processing unit 34, and the output processing unit 34 is connected to these peripheral devices. It is possible to output the swing data on the monitor memory 341 or the data processed as a swing form diagnostic material.

For example, the display 4 can reproduce the image of the swing data, and can also reproduce the extracted key frames arranged in a matrix. The printer 5 can copy the contents of the display screen of the display 4, and can also copy arbitrary or all key frames as a disassembled photograph of the swing motion. The storage device 6 is for storing swing data and data processed for diagnosis.

FIG. 2 shows the screen of the display 4 for displaying the image of eight key frames processed for diagnosis. When such a display instruction is given by a diagnostician via an operation input means such as a keyboard (in the present embodiment, since the touch panel display is adopted, this corresponds to the operation input means), the output processing unit 34 of the personal computer 3 is displayed.
Displays frames with a tag, that is, key frames, extracted from the swing data on the monitor memory 341 and arranged in a matrix to display an image.

From the left to the right of the upper stage, and from the left to the right of the lower stage, at the time of address, takeback, top, downswing, first impact, so that a series of swing motions can be understood in time series, Images at 2nd impact, follow-through, and finish are displayed in 2x4. As an output mode on the screen, all key frames may be output as images at the same time, but the output processing unit 34 determines the time difference between the key frames based on the interval information between the specific operation points written on the tag. If the images are attached and the images are sequentially output, it becomes a display mode full of realism, and more accurate swing form diagnosis can be performed.

Since the swing form diagnosis device determines the specific operating point based on the actual image,
Inappropriate image (image at a point other than the specific operation point)
Is not extracted, and the output images are both highly reliable and match the image at the true specific operating point. Further, the number of specific operating points to be determined is larger than that of the conventional one. In this respect, it can be said that the swing form diagnosis device provides a very useful diagnostic material for diagnosing the swing form.

Further, in the above-described embodiment, since all the specific operating points are determined by using image processing, when diagnosing a swing form, the swing form diagnosing device, photographing means (for example, a digital video camera), golf, etc. A club, a golf ball, and a means (for example, a tee) for setting the golf ball are sufficient, and all are general-purpose products except the swing form diagnostic device.
It is intended to provide a versatile swing form diagnosis method regardless of location. Moreover, since the troublesome preparation work such as attaching the reflective tape to the body of the subject or the golf club is not necessary, the diagnosis can be performed quickly and easily.

[0050]

EXAMPLE A specific operation point determination process will be described below as a specific example of the above embodiment. Figure 3
FIG. 4 is a flowchart of a swing data extraction process, and FIG. 4 is a flowchart of a plurality of specific operation point determination processes using the swing data.

Referring to FIG. 3, with the subject standing at a predetermined position (golf swing execution position), the video camera 2
When a shooting command is issued to the computer (S1), the image data of the first frame is input to the personal computer 3 (S1).
2). The control unit (primary processing unit 32) of the personal computer 3 performs template matching processing on this frame (S3) and checks whether or not there is a golf ball in the frame (S4). When it is determined that the golf ball is still in the frame (NO in S4), the process returns to S2,
Image data of the next frame is sampled after 1/30 seconds.

Frames are sequentially accumulated by continuing between S2 and S4 (serial numbers are given to each frame), but when the subject makes a golf swing, the golf ball B is hit. Eventually S4
Is YES. If S4 is YES, the time point becomes the reference point, and the frame number (reference frame) becomes the reference frame number _Fbs (S5).

When the reference frame number _Fbs is set, the starting frame number _Fst is calculated (S6). The starting frame is a frame that is a predetermined time before the reference frame, and is a frame in which the subject is estimated to be in the address state. The starting frame number F _st is the reference frame number F
It is obtained by subtracting a predetermined number of frames B from _bs . The number of frames B is obtained by an empirical rule and is a variable that can be changed arbitrarily.

As a result of calculating the starting frame number F _st ,
If there is no corresponding frame (S7),
Assuming that the image data of the address state has not been collected,
An error is displayed (S8). If S7 is YES, then the end frame number F _ed is calculated (S9).
The end point frame is a frame after a predetermined time from the reference frame, and is a frame in which it is estimated that the subject has completed the swing motion. End frame number F _ed
Is calculated by adding a predetermined number of frames A from the reference frame number _Fbs . The number of frames A is obtained by an empirical rule and is a variable that can be changed arbitrarily.

As a result of calculating the end frame number F _ed ,
When there is no corresponding frame (S10)
Assumes that the image data in the finish state has not been collected, and after 1/30 seconds, the image data related to the next frame is collected (S11). Then, by continuing from S10 to S11, the frames after the reference point are sequentially accumulated, but if the photographing is continued and the predetermined number of frames A is accumulated after the reference point, S10 becomes YES and S12 becomes S12. Move to.

In S12, assuming that each frame from the starting point to the ending point has been accumulated, these frames are extracted as swing data including image data of a series of swing motions from the address to the finish. The frame number of each frame in the swing data is an integer serial number in order of oldest time (S13).

After the swing data is extracted, a process of determining a plurality of specific operating points is then performed based on the swing data. Referring to FIG. 4, first, from the swing data,
The N (= 1) th frame (that is, the starting frame) is extracted (S20), the N (= 2) th frame is extracted (S21), and a difference image is generated from both (S2).
2). Then, after the difference image is pre-processed (binarization process, smoothing process, etc.) (S23), the moving partial image is extracted, and the area of this moving partial image is obtained to determine the moving partial image in the X direction. The displacement amount and the displacement amount in the Y direction are calculated (S24).

In S25, it is checked whether or not the amount of displacement in the X direction is near 0, that is, whether the amount of displacement in the X direction is below a predetermined threshold value. In S26, it is checked whether the displacement amount in the Y direction is near 0, that is, whether the displacement amount in the Y direction is lower than a predetermined threshold value. If S25 is YES and S26 is YES, 1 is added to the variable C (default is 1) and the variable D becomes 0 (S27). The variable C indicates the number of frames in which the motion of the subject S is minimal in both the X and Y directions. The variable D indicates a flag that sets the state in which the movement of the subject S is minimal only in the X direction to 1. Also, S2
If 5 is YES and S26 is NO, the variable C is cleared to 1 and the variable D flag is set (S28). If S25 is NO, the variable C is cleared to 1,
The flag of the variable D is laid down (S29).

Next, in the state where 1 is added to the frame number N (S30), the determination process of each specific operating point is performed using the values of the reference frame number _Fbs , the variable C and the variable D (S).
31). And these processing (S20-S31),
It is looped until the final frame of the swing data is subjected to the differential image processing (S32).

Here, the value of the variable C is used for the address determination process. If the variable C is, for example, the number of two frames shown above (that is, if the state in which the movement of the subject S is minimal in both the X and Y directions continues for two frames), the second frame is addressed. And

The value of the variable D is used in the determination process at the time of takeback. Assuming that the address time is determined, the time when the flag of the variable D is first set is taken as the takeback time. However, "D = 0" in S27 is changed to "D
= 1 ”, S28“ D = 1 ”to“ D = D + 1 ”, S
29, "D = 0" is changed to "D = 1", and the variable D is, for example, 2 frames (that is, the state in which the movement of the subject S in the X direction is minimal continues for 2 frames). Then, the second frame may be determined to be the time of takeback.

The value of the variable C is used in the determination process at the top. If this variable C is, for example, the number of two frames shown above on condition that the address time has already been determined, the second frame is set to the top time. However, if the area of the swing data that will include the top time is recognized in advance, and if a frame that meets the determination conditions is found in the differential image processing in that area regardless of whether the determination at the address is successful, this May be at the top.

The value of the variable C is used for the determination process at the finish. If this variable C is, for example, the number of 10 frames shown above, on condition that the top time has already been determined, the 10th frame is set as the finish time. However, if the area of the swing data that will include the finish time is recognized in advance and regardless of the success or failure of the determination at the time of top, if a frame that matches the determination condition is found in the differential image processing in that area, May be the finish time.

In the determination processing at the time of downswing, at the time of the first impact, at the time of the second impact, and at the time of the follow-through, each frame number is calculated based on the reference frame number _bs , and the corresponding frame is set as the specific operating point. .

Although the above embodiments of the present invention have been described above, the present invention is not limited to the above embodiments and examples, and various modifications can be made without departing from the gist of the present invention. .

<Type of photographing means> For example, the photographing means is
There is no particular limitation as long as the captured image of the golf swing of the subject is output as image data over time, and known capturing means can be used.

<Motion Picture Acquisition Mode> In the above embodiment, the swing data and the key frame are extracted while the subject is photographed. However, the present invention is not limited to this. Alternatively, the swing data may be extracted and the key frame may be extracted based on the input moving image.

<Extraction of Swing Data> In the above embodiment, the buffer memory 32 of the primary processing unit 32 is used.
Since the ring buffer memory is used in 1, the swing data in which the starting point and the ending point are defined is extracted. However, even if the swing data is not the ring buffer memory or the ring buffer memory, only the starting point is defined. Swing data (the end point is inevitably determined at the time point corresponding to the last address of the buffer) or swing data that defines only the end point (the start point is necessarily determined at the time point corresponding to the buffer start address) It is possible. That is,
Regardless of the type of buffer memory, from the frame before the reference frame, or from the reference frame, from the frame after the reference frame, or from the frame before the reference frame, to the reference frame , Or up to a frame after a predetermined number of reference frames can be extracted as swing data.

<Use of Impact Sensor> Further, an impact sensor may be used to detect the moment when the golf ball is hit (at the time of impact). As the impact sensor, for example, a microphone that detects a hitting sound, a means that optically, magnetically, or mechanically detects a state where the golf ball disappears from the tee, a state where the golf club passes on the tee or immediately before the tee, Optically, magnetically, or mechanically detecting means can be used. If an impact sensor is used, the swing data extraction process may be started by inputting the detection signal as a trigger signal to the primary input unit of the personal computer, with the generation time of the detection signal as a reference point. Good. The detection signal of the impact sensor can also be used to extract the frame at the time of addressing.

<Use of Address Sensor> Further, the address sensor may be used when detecting the state (at the time of addressing) held by the subject. As the address sensor, for example, a means for optically, magnetically or mechanically detecting a state in which the golf club is placed immediately before the tee can be used. If an address sensor is used, the swing data extraction process may be started by inputting this detection signal as a trigger signal to the primary input unit of a personal computer, with the detection signal generation time as a reference point. Good. Specifically, it can be inferred that the subject has completed a swing motion after a predetermined time (after a predetermined frame) from the time of addressing as a starting point. Is used as swing data. Then, the frame corresponding to the generation of the detection signal can be set as the address frame.

<Specific Operating Point Judgment Method (1)> Further, in the above embodiment, the position of the golf ball in the frame cannot be specified, so the template matching method is used for the judgment at the time of impact. If this golf ball is set at a fixed position (that is, the golf ball setting area in the frame is fixed), the image processing is performed only on this predetermined area, so that the golf ball "is present". The frame that changes to the “absence” state can be determined by the differential image processing method. Therefore, it is possible to determine the first impact time and the second impact time by using the difference image processing method. In addition, if an image capturing means capable of high-speed image capturing (image capturing of 240 frames or more per second) is used, the number of frames of swing data increases (image capturing interval shortens). As in the case of takeback, it is possible to make a determination using the differential image processing method. Then, it becomes possible to determine all the specific operation points of the series of swing operations from the address to the finish by the differential image processing method.
In addition, when high-speed shooting is performed, it becomes possible to obtain an image at the very moment when the club head catches the golf ball, so in this case, the first impact time and the second impact time are integrated into one. Good.

<Specific Operating Point Judging Method (Part 2)> In the above embodiment, the time when the golf ball is exhausted is used as the reference point for extracting the swing data and the key frame, and at the time of impact. But
The club head Y is adjusted so that the swing form can be appropriately diagnosed even when the golf ball is not set or when the club head does not hit the golf ball.
The time when the amount of movement in the direction becomes the minimum may be obtained by the difference image processing, and this may be used as the reference point or at the time of impact.

<Frame Division / Partial Designation> Further, in the above embodiment, when determining the take-back time, the difference image is generated using all the image data of the frame, but the frame is specified. Divide it into numbers,
A part of the area may be designated, and the differential image processing may be performed using the image data of the target divided area or the designated area. Thereby, the image processing time can be shortened. For example, if the frame is divided into four vertically, in the case of a right-handed subject, the subject's arm and shaft may be reflected in the left divided frame, so differential image processing is performed using image data of this divided frame. I do.

<Use of Background Image> Further, in the above embodiment, when the specific motion point is determined, one frame in the swing data is compared with the preceding frame and a difference image is output. However, the present invention is not limited to this, and a background image in which the subject is not captured may be prepared in advance, and each frame in the swing data may be compared with one background image to output the difference image.

<Recognition of subject etc. from frame> By preparing a background image, it is possible to compare this background image and a frame in which a swing motion is performed at the same place, and extract image data of only the subject etc. it can. This allows
Subject, golf club, and golf ball remain as a result of background removal. If you apply a labeling process to this,
"Subject", "golf club", and "golf ball" can be classified respectively.

<Creation of Synthetic Image> In order to absorb image differences due to individual differences, such as differences in body types of people and perspectives of subjects for shooting in different places, when creating a synthetic image,
An image of the subject or the like on one side of the two target frames is enlarged or reduced to make the subject or the like in both frames isometric (scaling process). Next, using the image data of the subjects separated in each frame, the following formula is used for the portion where the images of the two subjects overlap, while the original pixel value is directly displayed for the non-overlapping portion, the contrast decreases. Can be prevented and a clearer composite image can be obtained. Combined result = pixel value of frame 1 × transparency 1 + pixel value of frame 2 × transparency 2 (Equation)

<Utilization of Synthesized Image> Image synthesis is used for superimposing display between key frames. In addition, the key frame relating to the best swing motion of the user prepared in advance and the key frame to be diagnosed this time are displayed in an overlapping manner for each specific operation point so as to be compared with each other or prepared in advance. For example, a professional golfer (ie,
It is also used for comparing and displaying the key frame of a swing form modeler) and its own key frame to be diagnosed for each specific operating point in an overlapping manner. In general, since the time required for the swing motion differs depending on the person, proper comparison diagnosis cannot be performed only by reproducing the images of the swing motion of the subject and the professional golfer, for example. INDUSTRIAL APPLICABILITY The present invention extracts and compares key frames (images) related to swing motions of two persons, and therefore, it is possible to perform appropriate comparative diagnosis without being affected by temporal disagreement of swing motions. Becomes

<Simultaneous shooting> The shooting direction is not only from the front side direction but also the back side direction of the subject (the specific operation point determination processing can be performed from this direction), or the upper side of the head. Direction, front side direction, rear side direction, or the like. When shooting with two or more cameras, synchronize the frames output from each camera. By doing so, the frame corresponding to the key frame extracted from the image data captured by one camera can be extracted from the image data captured by the other camera, which allows, for example, the front side direction and the rear side direction. It is possible to simultaneously display mutually corresponding key frames photographed from at least two directions.

<Tracking of Club Head> In a golf swing, a swing operation in which the trajectory of the club head from the address to the top and the trajectory of the club head from the top to the impact are exactly the same is good.
It is known that the position of the club head reciprocates between a diagonally lower right position and a diagonally upper left position (in the case of a right-handed subject photographed from the rear side direction) of the frame due to the nature of the swing motion. Based on this empirical rule, the club head can be tracked by the difference image processing between the background image and the captured image and the labeling processing between the subject and the golf club. If the trajectory of the club head from the address to the top and the trajectory of the club head from the top to the impact are displayed in different colors or different line types, for example, in the frame at impact, diagnosis from another perspective can be performed. It will be possible.

<Swing Type> In addition to golf, the swing form diagnosis device (method) advises proper form correction based on analysis of swing forms such as batting of tennis and baseball (or self-diagnosis). It can be used as a training system.
When using a moving ball in baseball or the like, a high-speed shooting camera or the like may be used. When a moving ball is not used, a baseball ball or the like may be placed on a pole standing on the ground, and the static ball may be hit with a bat or the like.

[0081]

As described above, the swing form diagnosis apparatus according to the present invention extracts, as swing data, frames related to a series of swing motions performed by the subject from the input moving image of the subject, and the swing data of this swing data is extracted. A frame corresponding to a specific motion point during a swing motion is extracted from the inside by using differential image processing, and each specific motion point is determined based on a frame (image) of a live-action image. The specific operating point, which is extremely important for diagnosis, is accurate, and as a result, it is possible to accurately extract the frame intended by the diagnostician (subject in the case of self-diagnosis) from multiple frame groups involved in swing motion. Therefore, the swing form of the subject can be easily and effectively diagnosed.

[Brief description of drawings]

FIG. 1 shows a system configuration of a swing form diagnosis device according to an embodiment of the present invention.

FIG. 2 shows a display screen of a display that displays images of eight key frames processed for diagnosis according to the same embodiment.

FIG. 3 shows a flowchart of a swing data extraction process of the same embodiment.

FIG. 4 shows a flowchart of a process of determining a plurality of specific operating points using swing data according to the same embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Swing form diagnostic device, 2 ... Video camera (photographing means), 3 ... Personal computer (processing means), 31 ... Input processing part, 32 ... Primary processing part, 321 ... Buffer memory, 33
... Secondary processing unit, 331 ... Format conversion unit, 332 ...
Buffer memory, 34 ... Output processing unit, 341 ... Monitor memory, 4 ... Display (first output means, operation input means), 5 ... Printer (second output means), 6 ... Storage device,
S ... Subject, T ... Tee, B ... Golf ball, H ... Club head

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Toyofumi Tamura             2-9-14 Minamisenba, Chuo-ku, Osaka-shi, Osaka Prefecture               Takasago Electric Industrial Co., Ltd. (72) Inventor Shinya Iwasaki             2-9-14 Minamisenba, Chuo-ku, Osaka-shi, Osaka Prefecture               Takasago Electric Industrial Co., Ltd.

Claims (10)

[Claims] 1. An input unit for inputting a moving image of a subject photographed by using the photographing unit, and an extracting unit for extracting a frame relating to a series of swing motions performed by the subject as swing data from the inputted moving image. A difference calculating means for comparing each frame in the swing data with a reference image and outputting a difference image; and a determining means for determining a specific operation point during the swing operation according to an analysis result of the moving partial image in the difference image. And a display unit for extracting and displaying an image of a frame corresponding to each of the determined specific operating points, and a swing form diagnosis apparatus comprising: 2. In a golf swing, a frame at the time when the golf ball is lost by hitting is extracted by the difference image processing or template matching processing for the golf ball or by using an impact sensor, and is used as a reference. 2. The swing form diagnosis apparatus according to claim 1, wherein the swing form diagnosis is performed from a frame before a predetermined number of frames to a frame after a predetermined number of the reference frames as swing data. 3. In a golf swing, a frame in which a subject is in an address state is extracted by the difference image processing or template matching processing for the subject or the golf club, or by using an address sensor, which serves as a reference. 2. The swing form diagnosis apparatus according to claim 1, wherein the swing form diagnosis is performed from a frame to a frame after a predetermined number of the reference frames as the swing data. 4. In a golf swing, an analysis result that almost or no moving partial images are extracted from a difference image is obtained, and at the time of addressing at a specific operating point,
The swing form diagnosis apparatus according to claim 1, wherein the top form or the finish time is determined. 5. In a golf swing, an analysis result that there is little or no lateral displacement of a moving partial image extracted from a difference image is obtained, and a specific operating point is taken back or downswing. Alternatively, the swing form diagnosis device according to any one of claims 1 to 3, wherein the time of follow-through is determined. 6. In a golf swing, it is possible to determine an impact time of specific operating points by obtaining an analysis result that there is little or no vertical displacement of a moving portion image extracted from a difference image. The swing form diagnosis device according to any one of claims 1 to 3, which is characterized. 7. The swing form diagnosis apparatus according to claim 1, wherein the difference image processing or the template matching processing is performed only on a partial area of a frame. 8. The reference image is a frame immediately before in the swing data.
7. The swing form diagnosis device according to any one of 1 to 7. 9. The swing form diagnosis apparatus according to claim 1, wherein the reference image is a background image in which a subject is not captured. 10. The swing form diagnosis apparatus according to claim 1, wherein a position of the club head in each frame is specified and a locus is displayed by the difference image processing in a golf swing. .