JP2006244181A - Device and program for correcting online coordinate value - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an online coordinate value correcting device for temporally smoothly perform correction without deteriorating continuity in a corrected coordinate value when the coordinate value to be detected by automatic processing is corrected by manual processing. <P>SOLUTION: The online coordinate value correcting device 1 comprises: a detection information input means 10 for inputting a detection coordinate value and a reliability value relative to the detection coordinate value from a position detecting part 2; a specified information input means 20 for inputting a specified coordinate value, which indicates a position to be specified by an operator from a coordinate specification device 3 to be operated by the operator, and a pressure value, which is added to the coordinate specification device 3 by the operator; and a weighting and synthesizing means 30 for calculating a correction coordinate value, which is obtained by correcting the detection coordinate value, by weighting and synthesizing the detection coordinate value and the specified coordinate value, based on the reliability value and the pressure value. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an online coordinate value correcting device and an online coordinate value correcting program for correcting a detected coordinate value indicating a position of a detection target detected in time series.

2. Description of the Related Art Conventionally, a technique (position detection device) that detects an object whose position changes with time in image measurement or the like and outputs coordinates (detection coordinate values) indicating the position has been disclosed (for example, see Patent Document 1). .
However, when detecting an object that changes with time, the detected position is not necessarily accurate due to the presence of a plurality of objects or the occurrence of noise or the like in an image.
Therefore, in the video object detection / tracking device disclosed in Patent Document 1, the reliability of the position of the detected object (reliability value indicating the accuracy) is assigned to each object, and detection of the object to be detected is detected. Increases accuracy.

However, even if the detection accuracy is increased in this way, 100% detection accuracy cannot be obtained. Therefore, when the detection target is automatically detected by such a position detection device, when the time series of the automatically detected coordinate values is not a desired result, the operator switches from the automatic detection operation to the manual correction operation. It is necessary to correct the coordinate value by inputting with a pointing device such as a mouse or a pen tablet, or by numerical input from a keyboard.
In addition, when manually specifying the target coordinate position on the screen, the coordinate value input by the operator with an error is adaptively corrected based on information such as the object arrangement on the screen and re-inputted. Has been disclosed (for example, see Patent Document 2).
JP 2002-358526 A (paragraphs 0019 to 76, FIG. 1) JP-A-8-272526 (paragraphs 0017 to 0022, FIG. 1)

However, if the detection operation that was automatically performed is switched to manual and the coordinate value is corrected, the continuity is lost before and after the switching, so a “jump” of the position (coordinates) to be detected occurs and the time There is a problem that the smoothness is lost. Conversely, when switching from manual processing to automatic processing, the continuity is similarly impaired.
For example, when the position (coordinates) to be detected is broadcast together with the video, if the continuity of the coordinate values is impaired, the viewer may have experienced a failure on the broadcast station side or on the television receiver. There is a problem of letting go.
On the other hand, although the technique described in Patent Document 2 links manual processing and automatic processing, the input coordinate value is corrected to the coordinate position of the target (object), and the position of the detection target is changed. It is not intended to connect smoothly. That is, even in the technique described in Patent Document 2, if the coordinate values are manually corrected, there is a problem that the continuity of the coordinate values is lost and the temporal smoothness is lost.

  The present invention has been made to solve the above-described problems. Even when the coordinate value of the detection target detected by the automatic process is corrected by the manual process, the coordinate value of the detection target is changed. It is an object of the present invention to provide an on-line coordinate value correcting apparatus and an on-line coordinate value correcting program that can be corrected smoothly in time without impairing continuity.

  The present invention was devised to achieve the above object. First, the on-line coordinate value correcting device according to claim 1 provides a detected coordinate value indicating a position of a detection target detected by the position detecting device. An on-line coordinate value correction device for correction, comprising a detection information input means, an instruction information input means, and a weighted synthesis means.

In such a configuration, the on-line coordinate value correcting device inputs detection coordinate values from the position detection device in time series by the detection information input means, and an instruction output from the coordinate instruction device operated by the operator by the instruction information input means. The coordinate value and the pressurization value applied to the coordinate pointing device by the operator are input.
This position detection device may be anything as long as it outputs the position of the detection target as a coordinate value (detected coordinate value). For example, this position detection device detects a subject from an image and outputs a coordinate value indicating the position, a radio wave from a GPS satellite, tire rotation speed, gyro sensor information, etc. It may detect and output a coordinate value indicating the position.

However, the detected coordinate value may indicate an incorrect position due to noise or the like. Therefore, the online coordinate value correcting apparatus operates so as to correct the coordinate value to a position (indicated coordinate value) indicated by the operator. At this time, the online coordinate value correcting device weights and synthesizes the detected coordinate value and the designated coordinate value based on the reliability value indicating the accuracy of the predetermined detected coordinate value and the pressurization value by the weighted synthesizer. Then, a corrected coordinate value obtained by correcting the detected coordinate value is calculated.
Thus, the detected coordinate value is corrected to a position closer to the indicated coordinate value as the operator applies more pressure to the coordinate indicating device. For example, by making the coordinate pointing device a pen tablet device with a writing pressure sensing function, the detected coordinate value of the online coordinate value correcting device is corrected in accordance with the writing pressure value (pressing value). .

  The reliability value is set in advance to a high value when the accuracy of the position detection device to be used is high, and to a low value when the accuracy is low. Thus, even when the same pressurization value is input, the correction amount of the detected coordinate value is small when the reliability value is high, and the correction amount of the detection coordinate value is large when the reliability value is low.

  The on-line coordinate value correcting device according to claim 2 is an on-line coordinate value correcting device for correcting a detected coordinate value indicating a position of a detection target, which is output together with a reliability value indicating accuracy from the position detecting device. The detection information input unit, the instruction information input unit, and the weighted synthesis unit are provided.

In such a configuration, the on-line coordinate value correcting device inputs the detected coordinate value and the reliability value in time series from the position detecting device by the detection information input unit, and the coordinate indicating device operated by the operator by the instruction information input unit. And the pressurization value applied by the operator to the coordinate pointing device.
The on-line coordinate value correcting device corrects the detected coordinate value by weighted combining the detected coordinate value and the designated coordinate value based on the reliability value and the pressure value by the weighted combining unit. Is calculated.
Thereby, when the reliability value is high, the correction amount of the detected coordinate value is small, and when the pressurization value is high, the correction amount of the detection coordinate value is large.

  Furthermore, the online coordinate value correcting device according to claim 3 is the online coordinate value correcting device according to claim 1 or 2, wherein the weighting synthesis unit includes a first weighting factor calculating unit, a second weighting factor calculating unit, and a second weighting factor calculating unit. The calculation means and the coordinate value correction means are provided.

In such a configuration, the online coordinate value correcting apparatus calculates the weight of the reliability value for performing weighted synthesis as the first weighting factor by the first weighting factor calculating means using a predetermined calculation formula. For example, in this calculation formula, the value of the weighting factor is increased as the reliability value is higher, and the value of the weighting factor is decreased as the reliability value is lower.
In the online coordinate value correcting device, the second weighting factor calculating means calculates the weight of the pressurization value for performing weighted synthesis as the second weighting factor by a predetermined calculation formula. For example, in this calculation formula, the higher the pressurization value, the larger the weighting factor value, and the lower the pressurization value, the smaller the weighting factor value.

Then, the online coordinate value correcting device uses the coordinate value correcting means to calculate the added value of the multiplied value of the detected coordinate value and the first weighting factor and the multiplied value of the indicated coordinate value and the second weighting factor. A corrected coordinate value is calculated by dividing the added value of the first weight coefficient and the second weight coefficient.
Accordingly, the corrected coordinate value is calculated as a coordinate value proportionally proportional to the detected coordinate value and the designated coordinate value corresponding to the ratio between the first weighting factor and the second weighting factor.

  Further, the online coordinate value correcting device according to claim 4 is the online coordinate value correcting device according to claim 1 or 2, wherein the weighted synthesis means includes a first observation covariance matrix calculating means, An observation covariance matrix calculation unit, a first filter estimation unit, a second filter estimation unit, a prediction estimation unit, and a selection unit are provided.

In such a configuration, the online coordinate value correction apparatus calculates the covariance matrix of the reliability value as the first observation covariance matrix observed in the detection information input means by the first observation covariance matrix calculation means. Note that the reliability value is expressed as a covariance matrix by decreasing the value of the covariance matrix component when the reliability value is high and increasing the value of the covariance matrix component when the reliability value is low. be able to.
Further, the online coordinate value correction apparatus uses the second observation covariance matrix calculation means to convert the covariance matrix of the pressurization value as the second observation covariance matrix observed in the instruction information input means, similarly to the reliability value. calculate.

Then, the online coordinate value correcting device predicts a state vector and a state covariance matrix including the coordinate value indicating the position of the detection target based on the detected coordinate value and the first observation covariance matrix by the first filtering estimation unit. The temporary state vector and the temporary state covariance matrix that estimate the current state (each component of the state vector) are calculated from the estimated state vector prediction value and the state covariance matrix prediction value.
The accuracy of estimation can be improved by including not only the position of the detection target but also the movement amount (speed) per unit time that the detection target moves in the state vector.
In this way, the first filtering estimation means uses the detected coordinate value and the first observation, which are information currently observed, from the state vector prediction estimation value and the state covariance matrix prediction estimation value predicted by the information one time before. Based on the covariance matrix, it functions as a filter estimation unit of a Kalman filter that estimates (filters) the current state.

The on-line coordinate value correction apparatus also corrects the temporary state vector and the temporary state covariance matrix based on the designated coordinate value and the second observed covariance matrix by the second filtered estimation means. Calculate the dispersion matrix filtering estimate.
That is, the second filtered estimation means is based on the indicated coordinate value and the second observed covariance matrix, which are information currently observed, from the temporary state vector and the temporary state covariance matrix estimated by the first filtered estimation means. Furthermore, it functions as a filter estimation unit of a Kalman filter that estimates (filters) the current state.

  Then, the online coordinate value correcting device delays the state vector filtered estimated value and the state covariance matrix filtered estimated value calculated by the second filtered estimating unit by the prediction estimating unit, and based on a predetermined state transition rule, Predict state vector prediction estimates and state covariance matrix prediction estimates. That is, the prediction estimation means predicts the current state vector and state covariance matrix from the state vector and state covariance matrix one point before. The state vector prediction estimation value and the state covariance matrix prediction estimation value, which are the prediction results, are used for the filter estimation of the first filter estimation means. Thus, the online coordinate value correcting apparatus forms a loop in the filtering estimation process and the prediction estimation process, and the state vector and the state covariance matrix are estimated sequentially.

Further, the state vector filtered estimated value (state vector) calculated by the second filtered estimating means always includes a value obtained by correcting the detected coordinate value.
Therefore, the online coordinate value correcting apparatus selects the position (detected coordinate value) of the detection target from the state vector filtered estimated value (state vector) calculated by the second filtered estimating unit by the selecting unit, and the corrected coordinate value and To do.

  Furthermore, the online coordinate value correction device according to claim 5 is the online coordinate value correction device according to any one of claims 1 to 4, wherein the position detection device detects the position of the detection target from an image. And a video display means including a corrected coordinate position graphic drawing means and a video composition means.

In such a configuration, the online coordinate value correction apparatus draws a graphic corresponding to the position indicated by the corrected coordinate value by the corrected coordinate position graphic drawing means in the video display means. In addition, as long as this figure can identify a position visually, the shape, such as a point, a straight line, and a rectangle, does not ask | require.
The on-line coordinate value correcting apparatus combines the graphic drawn by the corrected coordinate position graphic drawing means with the video whose position has been detected by the position detecting device by the video synthesizing means in the video display means. As a result, the online coordinate value correction apparatus visually presents to the operator which position in the video the corrected position corresponds to.

  Moreover, the online coordinate value correction device according to claim 6 is the online coordinate value correction device according to any one of claims 1 to 4, wherein the position detection device detects the position of the detection target from an image. And a video display means including a detected coordinate position graphic drawing means and a video composition means.

  In such a configuration, the online coordinate value correcting apparatus draws a figure corresponding to the position indicated by the detected coordinate value by the detected coordinate position figure drawing means in the video display means. Then, the online coordinate value correcting apparatus synthesizes the graphic drawn by the detected coordinate position graphic drawing means and the video whose position has been detected by the position detecting device by the video synthesizing means in the video display means. As a result, the online coordinate value correcting apparatus visually presents to the operator at which position in the video the position detection apparatus has detected the detection target.

  Moreover, the online coordinate value correction device according to claim 7 is the online coordinate value correction device according to any one of claims 1 to 4, wherein the position detection device detects the position of the detection target from an image. And a video display means including a detected coordinate position graphic drawing means, a corrected coordinate position graphic drawing means, and a video composition means.

In such a configuration, the online coordinate value correcting device draws a figure corresponding to the position indicated by the detected coordinate value by the detected coordinate position figure drawing means in the video display means, and by the corrected coordinate position figure drawing means in the video display means, A figure corresponding to the position indicated by the corrected coordinate value is drawn. The on-line coordinate value correcting device detects the figure drawn by the detected coordinate position figure drawing means, the figure drawn by the corrected coordinate position figure drawing means, and the position detection device detects the position by the image synthesizing means in the image display means. Is combined with the video.
As a result, the online coordinate value correction device visually presents to the operator which position in the video the position detection device has detected the detection target and to which position the position has been corrected.

  Furthermore, the on-line coordinate value correction program according to claim 8 is a coordinate indicating device in which a detection coordinate value indicating a position of a detection target, which is output together with a reliability value indicating accuracy from the position detection device, is operated by an operator. In order to correct based on the indicated coordinate value output from the pressure value applied to the coordinate indicating device, the computer is connected to a first weight coefficient calculating means, a second weight coefficient calculating means, a coordinate value The configuration is made to function as correction means.

In such a configuration, the online coordinate value correction program calculates the weight of the reliability value for performing weighted synthesis as the first weighting factor by the first weighting factor calculation means. For example, the weighting factor value is increased as the reliability value is higher, and the weighting factor value is decreased as the reliability value is lower.
In addition, the online coordinate value correcting apparatus calculates the weight of the pressurization value for performing weighted synthesis as the second weighting coefficient by the second weighting coefficient calculating means. For example, the weighting factor value is increased as the pressure value is higher, and the weighting factor value is decreased as the pressure value is lower.
Then, the online coordinate value correcting device uses the coordinate value correcting means to calculate the added value of the multiplied value of the detected coordinate value and the first weighting factor and the multiplied value of the indicated coordinate value and the second weighting factor. A corrected coordinate value is calculated by dividing the added value of the first weight coefficient and the second weight coefficient.
Accordingly, the corrected coordinate value is calculated as a coordinate value proportionally proportional to the detected coordinate value and the designated coordinate value corresponding to the ratio between the first weighting factor and the second weighting factor.

  The on-line coordinate value correcting program according to claim 9 is a coordinate indicating device in which a detected coordinate value indicating a position of a detection target, which is output together with a reliability value indicating accuracy from the position detecting device, is operated by an operator. In order to make correction based on the indicated coordinate value output from the pressure value applied to the coordinate indicating device, the computer is provided with first observation covariance matrix calculation means and second observation covariance matrix calculation. Means, first filtering estimation means, second filtering estimation means, prediction estimation means, and selection means.

In such a configuration, the online coordinate value correction program calculates the covariance matrix of the reliability value as the first observation covariance matrix by the first observation covariance matrix calculation means. The online coordinate value correction program calculates the covariance matrix of the pressurized value as the second observation covariance matrix by the second observation covariance matrix calculation means.
Then, the online coordinate value correction program predicts a state vector and a state covariance matrix including coordinate values indicating the position of the detection target based on the detected coordinate values and the first observation covariance matrix by the first filtering estimation unit. The temporary state vector and the temporary state covariance matrix that estimate the current state (each component of the state vector) are calculated from the estimated state vector prediction value and the state covariance matrix prediction value.
In addition, the online coordinate value correction program corrects the temporary state vector and the temporary state covariance matrix based on the designated coordinate value and the second observed covariance matrix by the second filtered estimation means. Calculate the dispersion matrix filtering estimate.

  Then, the online coordinate value correction program delays the state vector filtered estimated value and the state covariance matrix filtered estimated value calculated by the second filtered estimating means by the prediction estimating means, and based on a predetermined state transition rule, Predict state vector prediction estimates and state covariance matrix prediction estimates. Further, the online coordinate value correction program selects the position (detected coordinate value) of the detection target from the state vector filtered estimated value (state vector) calculated by the second filtered estimating unit by the selecting unit, and the corrected coordinate value and To do.

  According to the first aspect of the present invention, the detected coordinate value automatically detected by the position detecting device is corrected by the indicated coordinate value indicated by the operator using the coordinate indicating device, so that the position is corrected even during automatic detection. Is possible. Further, the present invention corrects the detected coordinate value based on the pressurization value applied to the coordinate pointing device, so that the operator can change the correction coordinate rapidly by controlling the pressurization value. The coordinate values can be corrected smoothly in time.

  According to the second aspect of the present invention, the detected coordinate value is based on the reliability value of the detected coordinate value automatically detected by the position detecting device and the pressurization value applied to the coordinate indicating device by the operator. Therefore, the corrected coordinate value is not corrected to the specified coordinate value at a time. For this reason, correction can be performed smoothly in time.

  According to the invention described in claim 3 or claim 8, the detected coordinate value automatically detected by the position detecting device and the indicated coordinate value indicated by the operator using the coordinate indicating device are represented by the reliability value and the pressurizing value. Therefore, the configuration can be simplified and the processing speed can be increased.

  According to invention of Claim 4 or Claim 9, the correction of a coordinate is implement | achieved by the regression type process by each process of a filtering and prediction. For this reason, even if one or both of the detected coordinate value automatically detected by the position detection device and the instruction coordinate value designated by the operator using the coordinate instruction device are temporarily lost, the time-series corrected coordinate value is not Therefore, the corrected coordinate value does not change greatly, and a corrected coordinate value that is smooth in time is output.

  According to the invention described in claim 5, the operator can visually confirm the corrected coordinate value corrected by a graphic in the video, so that the correctness of the position can be recognized quickly, Operability can be improved.

  According to the sixth aspect of the invention, since the operator can visually confirm the detected coordinate value automatically detected by the position detection device in the image with a figure, the operator views the image. The operation can be performed while improving operability.

  According to the seventh aspect of the present invention, the operator can confirm the detected coordinate value automatically detected by the position detecting device and the corrected corrected coordinate value on the video with a graphic in the video. Therefore, the operability of the coordinate value correction work can be improved.

Embodiments of the present invention will be described below with reference to the drawings.
<< First Embodiment >>
[Configuration of online coordinate value correction device]
First, the configuration of the online coordinate value correcting apparatus will be described with reference to FIG. FIG. 1 is a block diagram showing a configuration of an on-line coordinate value correcting apparatus according to the first embodiment of the present invention.
The online coordinate value correction apparatus 1 shown in FIG. 1 corrects the detected coordinate value indicating the position of the detection target output from the position detection apparatus 2. In addition to the position detecting device 2, a coordinate indicating device 3 and a display device 4 are connected to the online coordinate value correcting device 1.

  The position detection device 2 outputs detection coordinate values indicating the position of the detection target in time series. Here, it is assumed that the position detection device 2 detects an object (detection target) included in the video from the video, and outputs a detection coordinate value of the object and a reliability value indicating the accuracy. And For example, the position detection device 2 uses a video object detection / tracking device disclosed in Japanese Patent Application Laid-Open No. 2002-358526, uses an estimated position detected by the video object detection / tracking device as a detected coordinate value, and the estimated position. The reliability of the is a reliability value. The detected coordinate value and the reliability value may be a one-dimensional value or a multi-dimensional vector value.

  The coordinate pointing device 3 detects the coordinate pointed by the operator, detects a pressure value indicating the strength of the pressure applied by the operator to the coordinate pointing device 3, and indicates the pointed coordinate value that is the coordinate. And the pressure value are output. For example, the coordinate indicating device 3 uses a pen tablet device with a writing pressure sensing function or a touch panel device, and uses the detected two-dimensional coordinate value as the indicating coordinate value and the writing pressure value as the pressurizing value. The indicated coordinate value may be only a one-dimensional coordinate value that is one of the two-dimensional coordinate values, or may be converted to a one-dimensional to multi-dimensional coordinate by geometrically converting the two-dimensional coordinate. Good. The pressure value may be a pressure value detected by a stylus pen in a pen tablet device, or may be detected by a pressure sensor attached to the stylus pen, the touch panel device body, or the operator's body in the touch panel device. It is good also as using the made pressure value.

  The display device 4 is a general display means such as a cathode ray tube, a liquid crystal display, a plasma display, etc., and includes an image for the position detection device 2 to detect, a detected coordinate value detected by the position detection device 2, an online The corrected coordinate value corrected in the coordinate value correcting apparatus 1 is displayed. The display device 4 may be, for example, a liquid crystal display device in which a tablet (coordinate indicating device 3) is integrated. The display device 4 may be configured to include a transparent or translucent touch panel (coordinate indicating device) on a display device such as a liquid crystal display or a plasma display.

  Hereinafter, the configuration of the online coordinate value correcting apparatus 1 will be described in detail. Here, the online coordinate value correcting apparatus 1 includes detection information input means 10, instruction information input means 20, weighted synthesis means 30, and video display means 40.

  The detection information input means 10 inputs the detected coordinate value of the object detected from the video in the position detection device 2 and its reliability value. The input detected coordinate value and reliability value are output to the weighted synthesis means 30. The detected coordinate value is also output to the video display means 40.

  The instruction information input unit 20 inputs an instruction coordinate value and a pressure value (hereinafter referred to as a writing pressure value) detected by the coordinate indicating device 3 operated by the operator. The input designated coordinate value and writing pressure value are output to the weighted synthesis means 30.

  Based on the reliability value input from the detection information input unit 10 and the writing pressure value input from the instruction information input unit 20, the weighted synthesis unit 30 detects the detected coordinate value input from the detection information input unit 10 The corrected coordinate value obtained by correcting the detected coordinate value is calculated by performing weighted synthesis with the indicated coordinate value input from the indicated information input means 20. Here, the weighting synthesis unit 30 includes a first weighting factor calculation unit 31, a second weighting factor calculation unit 32, and a coordinate value correction unit 33.

The first weighting factor calculation unit 31 calculates a first weighting factor based on the reliability value input from the detection information input unit 10. For example, the first weighting coefficient calculating unit 31 converts the value from the minimum value to the maximum value that the reliability value can take into a value in the range of “0” to “1”.
That is, the first weighting factor calculation means 31 has a reliability value r, and a function (first weighting factor calculation function) that outputs a large scalar quantity when the reliability value r is high is f (r). The first weighting coefficient w _a is calculated by the following equation (1). It is assumed that the value of the first weight coefficient w _a is “0” or more.

  The first weighting factor calculated by the first weighting factor calculating unit 31 is output to the coordinate value correcting unit 33.

The second weighting factor calculating unit 32 calculates a second weighting factor based on the writing pressure value input from the instruction information input unit 20. For example, the second weighting coefficient calculating unit 32 converts the value from the minimum value to the maximum value that the writing pressure value can take into a value in the range of “0” to “1”.
In other words, the second weighting factor calculating means 32 assumes that the writing pressure value is p and a function (second weighting factor calculating function) that outputs a large scalar quantity when the writing pressure value p is high is g (p). The second weight coefficient w _m is calculated by the following equation (2). It is assumed that the value of the second weighting coefficient w _m is “0” or more. However, w _a + w _m > 0.

Here, although the expression (2) has been described as outputting a large scalar amount when the writing pressure value p is high, for example, the writing pressure is always applied to a pen tablet (coordinate indicating device 3). When it is desired to perform an operation to reduce the writing pressure when it is desired to give an instruction, Equation (2) is a function that outputs a large scalar quantity when the writing pressure value p is low.
The second weighting coefficient calculated by the second weighting coefficient calculating means 32 is output to the coordinate value correcting means 33.

The coordinate value correcting means 33 is based on the first weight coefficient and the second weight coefficient input from the first weight coefficient calculating means 31 and the second weight coefficient calculating means 32, respectively, and the detection information input means 10 and the instruction information input means. The corrected coordinate value obtained by correcting the detected coordinate value is calculated by proportionally dividing the detected coordinate value and the designated coordinate value input from 20 respectively.
That is, the coordinate value correcting means 33, the detection coordinate value y _a, when the indicated coordinate value was y _m, using a first weighting coefficient w _a and the second weighting factor w _m, the following correction coordinate value y _out (3) is calculated.

The corrected coordinate value calculated by the coordinate value correcting means 33 is output to the outside as the coordinate corrected by the online coordinate value correcting device 1. The corrected coordinate value is also output to the video display means 40.
Accordingly, when the operator increases the writing pressure in the operation of the coordinate indicating device 3, the detected coordinate value is corrected (largely corrected) in a direction closer to the position instructed by the operator. Further, by reducing the writing pressure, the correction amount for the detected coordinate value is reduced.

  The video display means 40 visually presents the coordinates to be corrected in the online coordinate value correction apparatus 1. Here, the video display means 40 includes a detected coordinate position graphic drawing means 41, a corrected coordinate position graphic drawing means 42, and a video composition means 43.

  The detected coordinate position graphic drawing means 41 draws a graphic indicating the position corresponding to the detected coordinate value input from the detection information input means 10 in a memory (video memory or the like) (not shown). As a result, the position of the detection target detected by the position detection device 2 can be visually indicated.

  The corrected coordinate position graphic drawing means 42 draws a graphic indicating the position corresponding to the corrected coordinate value input from the coordinate value correcting means 33 of the weighted synthesis means 30 in a memory (video memory or the like) (not shown). is there. This makes it possible to visually indicate the position where the detected coordinate value is corrected.

The image synthesizing means 43 uses the detected coordinate position figure drawing means 41 and the corrected coordinate position figure drawing means 42 to display the figure drawn in the memory (not shown) on the same image as the image input to the position detection device 2. Are combined and output to the external display device 4.
As a result, the position of the detection target detected by the position detection device 2 and the position where the detected coordinate value is corrected are displayed on the screen of the display device 4, and the operator can visually confirm that the correction is correct. You can check whether.

Here, an image displayed on the display device 4 by the online coordinate value correcting apparatus 1 will be described with reference to FIG. FIG. 5 is a diagram illustrating an example of a display screen displayed on the display device by the online coordinate value correcting device.
As shown in FIG. 5, here, as an example, an example in which an image of a soccer stadium is displayed is shown. Further, here, the position detection device 2 detects the position of the offside line by tracking a player (not shown) on the image, and sequentially uses the coordinates in the sideline direction of the offside line as detection coordinate values. Output together with the reliability value.

That is, the online coordinate value correcting device 1 is detected coordinate geometry L _a (here, dashed line) were animation based on the detected coordinates are located offside line position detecting apparatus 2 detects the display screen Display on D. Thus, the operator can perform an operation while confirming the position of the offside line detected by the position detection device 2.
The online coordinate value correcting apparatus 1 displays a corrected coordinate figure L _out (here, a solid line segment) drawn based on the corrected coordinate value on the display screen D. Thereby, the operator can visually confirm the correctness of the position of the offside line after the position correction.

  By configuring the online coordinate value correcting apparatus 1 in this way, the online coordinate value correcting apparatus 1 can detect the position of the detection target (detected coordinate value) detected by the position detecting apparatus 2, and the operator can use the coordinate indicating apparatus 3. Since the designated position (designated coordinate value) is proportionally divided based on the reliability value and the pen pressure value (pressurized value), the corrected coordinate value changes from the detected coordinate value to the designated coordinate value at a time. There is nothing. For this reason, the detected coordinate value is smoothly corrected as the operator's writing pressure changes.

  The online coordinate value correcting apparatus 1 can be realized by causing a general computer to function as each of the means described above. This program (online coordinate value correction program) can be distributed via a communication line, or can be distributed by writing on a recording medium such as a CD-ROM.

The configuration of the online coordinate value correcting apparatus according to the present invention has been described above, but the present invention is not limited to this configuration. For example, when the position detection device 2 does not output the reliability value but outputs only the detected coordinate value in time series, the detection information input means 10 inputs only the detected coordinate value and calculates the first weighting coefficient. The means 31 outputs a fixed weighting factor as the first weighting factor based on a reliability value of a predetermined detected coordinate value. For example, in the formula (3), the first weighting coefficient w _{a is set} to “0.5”.

  Further, in the video display means 40, the video composition means 43 synthesizes both the graphic drawn by the detected coordinate position graphic drawing means 41 and the graphic drawn by the corrected coordinate position graphic drawing means 42 to the video. However, only one of them may be combined with the video. In this case, either the detected coordinate position figure drawing means 41 or the corrected coordinate position figure drawing means 42 may be omitted from the configuration.

  Here, the weighting synthesis unit 30 calculates the weighting coefficient from the reliability value and the writing pressure value, and corrects the detected coordinate value by the proportional proportion based on the weighting coefficient. The detected coordinate value may be corrected based on the covariance of the pen pressure value. Note that an online coordinate value correcting apparatus that corrects the detected coordinate value based on the covariance of the reliability value and the writing pressure value will be described in detail later as a second embodiment.

[Operation of online coordinate value correction device]
Next, the operation of the online coordinate value correcting apparatus will be described with reference to FIG. FIG. 2 is a flowchart showing the operation of the online coordinate value correction apparatus according to the first embodiment of the present invention. Here, the operation of the weighted combining means 30 for calculating the corrected coordinate value in the online coordinate value correcting apparatus 1 will be mainly described.

First, the online coordinate value correction apparatus 1 determines whether or not the detection information input means 10 has input from the position detection apparatus 2 a detection coordinate value that is a position to be detected and a reliability value that indicates its accuracy. (Step S1) Wait until input (No in Step S1).
Then, the online coordinate value correcting apparatus 1 determines the reliability value from the reliability value by the first weighting factor calculation unit 31 of the weighting synthesis unit 30 when the detected coordinate value and the reliability value are input (Yes in step S1). One weighting coefficient is calculated by the equation (1) (step S2).

In the online coordinate value correcting apparatus 1, the instruction information input unit 20 receives an instruction coordinate value that is a position instructed by the operator and a writing pressure value indicating the writing pressure intensity from the coordinate indicating apparatus 3. It is determined whether or not (step S3).
Then, when the designated coordinate value and the writing pressure value are input (Yes in step S3), the online coordinate value correcting device 1 uses the second weighting coefficient calculating unit 32 of the weighting synthesis unit 30 to calculate the second value from the writing pressure value. The weighting coefficient is calculated by the equation (2) (step S4), and the process proceeds to step S6.
On the other hand, when the designated coordinate value and the writing pressure value are not input (No in step S3), the online coordinate value correcting apparatus 1 sets “0” in the second weighting coefficient in the coordinate value correcting unit 33 ( Step S5).

Then, the online coordinate value correcting apparatus 1 uses the coordinate value correcting unit 33 to calculate the first weighting factor calculated in step S2 and the second weighting factor calculated in step S4 (or set in step S5). The detected coordinate value is corrected by proportionally dividing the detected coordinate value and the designated coordinate value input from the detected information input means 10 and the designated information input means 20 according to the ratio as shown in the equation (3). A corrected coordinate value is calculated and output (step S6). And the online coordinate value correction apparatus 1 returns to step S1, and continues operation | movement.
Thereby, the online coordinate value correcting apparatus 1 can smoothly correct the detected coordinate values input in time series in accordance with the change of the writing pressure in the coordinate indicating apparatus 3 operated by the operator.

  Although not shown in the flowchart, the online coordinate value correcting apparatus 1 is configured to detect a figure indicating a position corresponding to the detected coordinate value by the detected coordinate position figure drawing means 41 when the detected coordinate value is input. Drawing in a memory not shown. Further, in step S6, the online coordinate value correcting apparatus 1 draws a graphic indicating the position corresponding to the corrected coordinate value in the memory by the corrected coordinate position graphic drawing means 42 when the corrected coordinate value is calculated.

Then, the online coordinate value correcting apparatus 1 combines the graphic drawn on the memory with the same video as the video input to the position detecting device 2 by the video synthesizing unit 43, and displays it on the external display device 4. Output.
As a result, the operator can visually recognize the positions of the detected coordinate value and the corrected coordinate value.

<< Second Embodiment >>
[Configuration of online coordinate value correction device]
Next, the configuration of the online coordinate value correcting apparatus according to the second embodiment will be described with reference to FIG. FIG. 3 is a block diagram showing the configuration of the online coordinate value correcting apparatus according to the second embodiment of the present invention. Since the online coordinate value correcting apparatus 1B according to the second embodiment is different from the online coordinate value correcting apparatus 1 described in FIG. 1 only in the weighting synthesis means 30 (30B), here, Only the weighted combining means having the different configuration will be described. Further, since the configuration other than the weighted synthesis means is the same as that of the online coordinate value correcting apparatus 1, FIG. 1 will be referred to as appropriate, and illustration and description thereof will be omitted.

  The weighting synthesis unit 30B is input from the detection information input unit 10 based on the respective covariances of the reliability value input from the detection information input unit 10 and the writing pressure value input from the instruction information input unit 20. The corrected coordinate value obtained by correcting the detected coordinate value is calculated by performing weighted synthesis of the detected coordinate value and the specified coordinate value input from the specified information input means 20. Here, the weighted synthesis unit 30B includes a first observation covariance matrix calculation unit 31B, a second observation covariance matrix calculation unit 32B, and a coordinate value correction unit 33B.

The first observation covariance matrix calculation means (first weighting coefficient calculation means) 31B determines the covariance (first observation covariance matrix) that becomes the first weighting coefficient from the reliability value input from the detection information input means 10. Is to be calculated. For example, the first observation covariance matrix calculation means 31B is the following formula (4), converts the reliability value r to the first observation covariance matrix R _a.

This function J (r) is a function (first observation covariance matrix calculation function) that outputs a semi-positive definite matrix (all eigenvalues are “0” or more). For example, when the detected coordinate value is two-dimensional (for example, x coordinate value, y coordinate value), the function J (r) is a 2 × 2 half positive definite matrix.
When the detected coordinate value is N-dimensional, the function J (r) can use a diagonal matrix such as N rows and N columns shown in the following equation (5).

Note that f _k (r) (k = 1, 2,..., N) indicates the variance of the reliability value r of the k-th component of the detected coordinate value, and it can be said that the smaller the variance value, the higher the reliability. Therefore, in all k, if f _k (r)> 0 and the reliability value r indicates that the reliability of the k-th component of the detected coordinate value is high, a small value is set for f _k (r). Shall be set. For example, f _k (r) sets a value from the maximum value to the minimum value that can be taken by the reliability value r of the k-th component of the detected coordinate value in the range of “0” to “1”.
The first observation covariance matrix calculated by the first observation covariance matrix calculation means 31B is output to the coordinate value correction means 33B.

The second observation covariance matrix calculation means (second weighting coefficient calculation means) 32B calculates the covariance (second observation covariance matrix) that becomes the second weighting coefficient from the writing pressure value input from the instruction information input means 20. Is to be calculated. For example, the second observation covariance matrix calculating unit 32B converts the writing pressure value p into the second observation covariance matrix R _m by the following equation (6).

The function G (p) is a function (second observation covariance matrix calculation function) that outputs a semi-positive definite matrix (all eigenvalues are “0” or more). For example, when the designated coordinate value is two-dimensional (eg, x coordinate value, y coordinate value), the function G (p) is a 2 × 2 half positive definite matrix.
When the designated coordinate value is N-dimensional, the function G (p) can use a diagonal matrix such as N rows and N columns shown in the following equation (7).

Note that g _k (p) (k = 1, 2,..., N) indicates the variance of the writing pressure value p of the k-th component of the indicated coordinate value, and g _k (p)> 0 for all k. To do. When the writing pressure value p is high, a small value is set for g _k (p). For example, g _k (p) sets a value from the maximum value to the minimum value that can be taken by the writing pressure value p of the k-th component of the indicated coordinate value in the range of “0” to “1”.

In addition, here, g _k (p) in Expression (7) has been described as setting a small value when the writing pressure value p is high, but for example, it is always applied to a pen tablet (coordinate indicating device 3). If you want to perform an operation that reduces the pen pressure when you want to give a pen pressure and give an instruction, set g _k (p) to a large value when the pen pressure value p is low. To do.
The second observation covariance matrix calculated by the second observation covariance matrix calculation means 32B is output to the coordinate value correction means 33B.

The coordinate value correction means 33B is based on the first observation covariance matrix and the second observation covariance matrix input (observed) from the first observation covariance matrix calculation means 31B and the second observation covariance matrix calculation means 32B, respectively. The corrected coordinate value obtained by correcting the detected coordinate value is calculated by synthesizing the detected coordinate value and the specified coordinate value respectively input from the detected information input means 10 and the specified information input means 20.
The coordinate value correcting means 33B estimates the internal state in time series by performing “filtering” for estimating the current internal state and “prediction” for estimating the future internal state, and detecting based on the internal state. Coordinate values and designated coordinate values are combined. Here, the internal state is a state vector including a corrected coordinate value that is an output value. For example, when the coordinate value (corrected coordinate value) is y and the unit time difference (hereinafter referred to as speed) is v, the state vector x is defined by the following equation (8).

  Here, the coordinate value correcting unit 33B includes a first filtered estimating unit 33a, a second filtered estimating unit 33b, a predicted estimating unit 33c, and a selecting unit 33d.

Based on the detected coordinate value input from the detection information input means 10 and the first observation covariance matrix calculated by the first observation covariance matrix calculation means 31B, the first filtered estimation means 33a is a prediction estimation described later. From the state vector prediction estimated value predicted by the means 33c and the state covariance matrix prediction estimated value, a temporary state vector and a temporary state covariance matrix obtained by estimating (filtering) the current state are calculated.
That is, the first filtered estimation means 33a is configured to detect the state vector (state vector predicted estimated value) and the state predicted by the prediction estimating means 33c based on the observed information (detected coordinate values and first observed covariance matrix). A current state vector (provisional state vector) and a state covariance matrix (provisional state covariance matrix) are estimated (filtered) from a dispersion matrix (state covariance matrix prediction estimation value). Note that the current state is further estimated by the second filtering estimation unit 33b described later, and thus is in a temporary state at this stage.

For example, the first filtered estimation means 33a performs estimation by a sequential estimation process by a filtered estimation unit of a general Kalman filter. Here, the detected coordinate value, which is the observed information, is y _a , the first observation covariance matrix is R _a , and the state at time t predicted from the state up to time [t−1] in the prediction estimation means 33c. When the vector prediction estimation value is x _{t | t−1} and the state covariance matrix prediction estimation value is P _{t | t−1} , the first filtering estimation means 33a uses the temporary state vector x _tmp and the temporary state covariance matrix P. _tmp is calculated by the following equation (9).

Here, K _a indicates a gain (Kalman gain) in the Kalman filter. H _a represents an observation matrix indicating the relationship between the state quantity (state vector) and the observation quantity (coordinate values). T represents transposition.
For example, as shown in the equation (8), the observation matrix H _a uses the following equation (10) when an N-dimensional coordinate value (corrected coordinate value) y and its velocity v are used as the state vector x. ). Note that I _N × _N represents a unit matrix of N rows and N columns, and O _N × _N represents a zero matrix of N rows and N columns.

Further generalized, and to be estimated coordinate state quantity including (as the output value correction coordinate value) (the state vector x), the relation between the observed quantity (the detection coordinate value y _a) function (observation function) h _a When represented by (x), the first filtering estimation means 33a calculates the temporary state vector x _tmp and the temporary state covariance matrix P _{tmp according} to the following equation (11). Incidentally, the observation matrix H _a is calculated by partially differentiating the observation function h _a (x).

The temporary state vector x _tmp and the temporary state covariance matrix P _tmp calculated in this way are output to the second filtered estimation means 33b.

The second filtered estimation means 33b is configured to perform first filtered estimation based on the indicated coordinate value input from the indicated information input means 20 and the second observed covariance matrix calculated by the second observed covariance matrix calculating means 32B. From the temporary state vector estimated by the means 33a and the temporary state covariance matrix, a state vector filtered estimated value and a state covariance matrix filtered estimated value obtained by estimating (filtering) the current state are calculated.
In other words, the second filtered estimation means 33b, based on the observed information (indicated coordinate values and second observed covariance matrix), the state vector (provisional state vector) and the state shared by the first filtered estimation means 33a. A current state vector (state vector filtered estimated value) and a state covariance matrix (state covariance matrix filtered estimated value) are estimated (filtered) from the dispersion matrix (provisional state covariance matrix).

For example, the second filtered estimation means 33b performs estimation by a successive estimation process by a filtered estimation unit of a general Kalman filter. Here, the indicated coordinate value which is the observed information is y _m , the second observation covariance matrix is R _m , the temporary state vector estimated (filtered) by the first filter estimating means 33a is x _tmp , and the temporary state covariance. When the matrix is P _tmp , the second filtering estimation means 33b calculates the state vector filtering estimated value x _{t | t} and the state covariance matrix filtering estimated value P _{t | t} by the following equation (12).

Here, K _m represents a gain (Kalman gain) in the Kalman filter. H _m represents an observation matrix indicating the relationship between state quantities (state vectors) and observation quantities (coordinate values). T represents transposition.
For example, when the observation matrix H _m uses an N-dimensional coordinate value (corrected coordinate value) y and its velocity v as the state vector x as shown in the equation (8), the following equation (13) ). Note that I _N × _N represents a unit matrix of N rows and N columns, and O _N × _N represents a zero matrix of N rows and N columns.

Further generalized, the relationship between the state quantity (state vector x) including the coordinates to be estimated (corrected coordinate values serving as output values) and the observation quantity (indicated coordinate value y _m ) is a function (observation function) h _m. When represented by (x), the second filtered estimation means 33b calculates the state vector filtered estimated value x _{t | t} and the state covariance matrix filtered estimated value P _{t | t} by the following equation (14). Note that the observation matrix H _m is calculated by partial differentiation of the observation function h _m (x).

The state vector filtered estimated value x _{t | t} calculated in this way is output to the prediction estimating means 33c and the selecting means 33d. Further, the state covariance matrix filtered estimated value P _{t | t} is output to the prediction estimating means 33c.

The prediction estimating unit 33c delays the state vector filtered estimated value and the state covariance matrix filtered estimated value calculated by the second filtered estimating unit 33b, and estimates (predicts) the current state based on a predetermined state transition rule. ) State vector (state vector prediction estimation value) and a covariance matrix (state covariance matrix prediction estimation value) of the state vector.
Here, the prediction estimation unit 33 c includes a delay unit 331 and a prediction estimated value calculation unit 332.

  The delay means 331 delays the state vector filtered estimated value and the state covariance matrix filtered estimated value calculated by the second filtered estimating means 33b. This delay means 331 holds a state vector filtered estimated value and a state covariance matrix filtered estimated value in a general memory (not shown), and is one time before (when the current time is t [t− 1) The state vector filtered estimated value and the state covariance matrix filtered estimated value at time 1) are read from the memory and output to the predicted estimated value calculating means 332.

The predicted estimated value calculating means 332 is a state vector predicted estimated value obtained by estimating (predicting) the current state from the state vector filtered estimated value and the state covariance matrix filtered estimated value delayed and output by the delay means 331, and the state A covariance matrix prediction estimated value is calculated.
For example, the predicted estimated value calculating means 332 sets the state vector filtered estimated value one time before (time [t−1]) to x _{t−1 | t−1} and the state covariance matrix filtered estimated value to P _{t−1 | t-1} , where the state transition matrix indicating the state transition rule is F and the process noise covariance matrix is Q, the current state (time t) state vector prediction estimate x _{t | t-1} and state covariance matrix prediction The estimated value P _{t | t−1} is calculated by the following equation (15).

Note that the state transition matrix F in Expression (15) is stored in advance in a storage unit (not shown) as a rule for performing state transition (state transition rule). As shown in the equation (8), the state transition matrix F uses the following equation (16) when an N-dimensional coordinate value (corrected coordinate value) y and its velocity v are used as the state vector x. It can be. Note that I _N × _N represents a unit matrix of N rows and N columns, and O _N × _N represents a zero matrix of N rows and N columns.

  By using the state transition matrix F shown in the equation (16), the prediction of the state transition in the equation (15) is based on a constant velocity model. Note that the state transition matrix F may be set with rules for predicting not only the uniform velocity model but also other models (for example, the uniform acceleration model).

Further, the covariance matrix Q of the process noise in the equation (15) should be considered in the transition from the state vector filtered estimated value x _{t−1 | t−1} to the state vector predicted estimated value x _{t | t−1} . Uncertainty is expressed as a covariance matrix, and is set in consideration of disturbance due to modeling error of a constant velocity model, noise applied from the outside, external vibration, and the like.
For example, the process noise covariance matrix Q can be given as a semi-positive definite matrix, and as shown in the equation (8), as a state vector x, an N-dimensional coordinate value (corrected coordinate value) y, When the velocity v is used, the following equation (17) can be obtained using the non-negative scalar value q. Note that I _N × _N represents a unit matrix of N rows and N columns, and O _N × _N represents a zero matrix of N rows and N columns.

  The state vector prediction estimation value calculated (predicted) by the prediction estimation value calculation unit 332 and the state covariance matrix prediction estimation value are output to the first filter estimation unit 33a.

The selection means 33d selects the position of the detection target from the state vector filtered estimated value calculated by the second filtered estimating means 33b, and outputs it as a corrected coordinate value.
That is, the selecting unit 33d extracts a coordinate value (corrected coordinate value y _out ) by defining a function ζ as shown in the following equation (18) from the state vector predicted estimated value x _{t | t at} time t.

As shown in the equation (8), this function ζ is expressed by the following equation (19) when an N-dimensional coordinate value (corrected coordinate value) y and its velocity v are used as the state vector x. be able to. Note that I _N × _N represents a unit matrix of N rows and N columns, and O _N × _N represents a zero matrix of N rows and N columns.

The corrected coordinate value y _{out that} is output need not necessarily have the same coordinate system and the same dimensions as the detected coordinate value input from the detection information input unit 10 and the instruction coordinate value input from the instruction information input unit 20. There is no. For example, two-dimensional coordinates on the image (on the screen of the display device 4) are used for the detected coordinate values and the designated coordinate values, and the corrected coordinate values y _out correspond to the ground on the soccer field shown in FIG. Two-dimensional coordinates may be used. In this case, camera parameters such as zoom, pan, and tilt of the camera that captured the soccer stadium are input to the selection means 33d corresponding to the video, and the function ζ of the equation (18) is the camera. Coordinate conversion is performed based on the parameters.

  As described above, the coordinate value correcting unit 33B estimates (predicts) the current state from the past state and estimates (filters) the current state based on the observation information, so that the observation information is not input. Even in this case, the coordinate value can be estimated and output as appropriate, and the correction can be performed smoothly.

  Note that the online coordinate value correcting apparatus 1B can be realized by causing a general computer to function as each of the above-described means. This program (online coordinate value correction program) can be distributed via a communication line, or can be distributed by writing on a recording medium such as a CD-ROM.

[Operation of online coordinate value correction device]
Next, referring to FIG. 4 (refer to FIG. 1 and FIG. 3 as appropriate for the configuration), the operation of the online coordinate value correcting apparatus will be described. FIG. 4 is a flowchart showing the main operation of the online coordinate value correcting apparatus according to the second embodiment of the present invention. Here, the operation of the weighted combining means 30B for calculating the corrected coordinate value in the online coordinate value correcting apparatus 1B will be mainly described.

First, the online coordinate value correction apparatus 1B determines whether or not the detection information input means 10 has input from the position detection apparatus 2 a detection coordinate value that is a position to be detected and a reliability value that indicates its accuracy. (Step S10).
Here, when the detected coordinate value and the reliability value are input (Yes in step S10), the online coordinate value correcting apparatus 1B uses the first observation covariance matrix calculating unit 31B to input the reliability value. A covariance (first observation covariance matrix) serving as a first weighting coefficient is calculated by the equation (4) (step S11).

Then, the online coordinate value correcting apparatus 1B is based on the detected coordinate value input in step S10 by the first filter estimating unit 33a of the coordinate value correcting unit 33B and the first observation covariance matrix calculated in step S11. From the state vector predicted by the prediction estimation unit 33c (state vector prediction estimation value) and the state covariance matrix (state covariance matrix prediction estimation value), the temporary state vector and the temporary state covariance estimated from the current state The matrix is calculated by the equation (9) (first filtering estimation: step S12).
On the other hand, when the detected coordinate value and the reliability value are not input (No in step S10), the first filter estimation unit 33a uses the state vector prediction estimation value and the state covariance matrix prediction predicted by the prediction estimation unit 33c. The estimated value is directly used as a temporary state vector and a temporary state covariance matrix (step S13).

Next, in the on-line coordinate value correcting apparatus 1B, the instruction information input means 20 inputs the instruction coordinate value that is the position instructed by the operator and the writing pressure value indicating the writing pressure intensity from the coordinate indicating apparatus 3. It is determined whether it has been performed (step S14).
Here, when the indicated coordinate value and the writing pressure value are input (Yes in step S14), the online coordinate value correcting device 1B uses the second observation covariance matrix calculating unit 32B to calculate the input writing pressure value from the input writing pressure value. A covariance (second observation covariance matrix) serving as a second weighting coefficient is calculated by the above equation (6) (step S15).

Then, the online coordinate value correction apparatus 1B is based on the indicated coordinate value input in step S14 and the second observation covariance matrix calculated in step S15 by the second filtering estimation unit 33b of the coordinate value correction unit 33B. From the state vector (provisional state vector) and the state covariance matrix (provisional state covariance matrix) calculated (filtered) by the first filtering estimation means 33a, the state vector filtered estimated value and the state estimated from the current state The covariance matrix filtered estimated value is calculated by the equation (12) (second filtered estimation: step S16).
On the other hand, when the designated coordinate value and the writing pressure value are not input (No in step S14), the second filtered estimating unit 33b calculates the temporary state vector and the temporary state calculated (filtered) by the first filtered estimating unit 33a. The covariance matrix is directly used as the state vector filtered estimated value and the state covariance matrix filtered estimated value (step S17).

  Then, the online coordinate value correcting apparatus 1B selects the position of the detection target from the state vector filtered estimated value calculated (filtered) by the second filtered estimating unit 33b by the selecting unit 33d, and outputs it as a corrected coordinate value ( Step S18).

Further, the online coordinate value correction apparatus 1B uses the prediction estimation unit 33c to calculate the current state vector (state vector dispersion estimated value) from the state vector (state vector filtered estimated value) and the state covariance matrix (state covariance matrix filtered estimated value) one point before. A state vector prediction estimation value) and a state covariance matrix (state covariance matrix prediction estimation value) are calculated by the equation (15) (prediction estimation: step S19).
And the online coordinate value correction apparatus 1 returns to step S10, and continues operation | movement. Further, the state vector prediction estimated value and the state covariance matrix prediction estimated value calculated (predicted) in step S19 are used in step S12 and step S13.

Note that the operation of the video display means 40 is the same as that of the online coordinate value correcting apparatus 1 described in the first embodiment, and thus the description thereof is omitted.
With the above operation, the online coordinate value correcting apparatus 1B can smoothly correct the detected coordinate values input in time series in accordance with the change in the writing pressure of the operator. Further, the online coordinate value correction apparatus 1B can output the corrected coordinate values in time series by appropriately performing filtering estimation and prediction estimation even when there is no input of the detected coordinate value or the designated coordinate value. .

It is a block diagram which shows the structure of the online coordinate value correction apparatus which concerns on 1st Embodiment in this invention. It is a flowchart which shows operation | movement of the online coordinate value correction apparatus which concerns on 1st Embodiment in this invention. It is a block diagram which shows the structure of the online coordinate value correction apparatus which concerns on 2nd Embodiment in this invention. It is a flowchart which shows operation | movement of the online coordinate value correction apparatus which concerns on 2nd Embodiment in this invention. It is a figure which shows the example of the display screen which the online coordinate value correction apparatus which concerns on this invention displays on a display apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1, 1B Online coordinate value correction apparatus 2 Position detection apparatus 3 Coordinate instruction | indication apparatus 4 Display apparatus 10 Detection information input means 20 Instruction information input means 30, 30B Weighting synthesis means 31 1st weight coefficient calculation means 31B 1st observation covariance matrix calculation Means (first weighting factor calculation means)
32 Second weighting coefficient calculating means 32B Second observation covariance matrix calculating means (second weighting coefficient calculating means)
33, 33B Coordinate value correction means 33a First filter estimation means 33b Second filter estimation means 33c Prediction estimation means 33d Selection means 40 Video display means 41 Detected coordinate position figure drawing means 42 Correction coordinate position figure drawing means 43 Video composition means 43

Claims (9)

An online coordinate value correcting device for correcting a detected coordinate value indicating a position of a detection target detected by a position detecting device,
Detection information input means for inputting the detection coordinate value from the position detection device;
An instruction information input means for inputting an instruction coordinate value indicating a position for instructing correction and a pressure value applied by the operator to the coordinate instruction apparatus from a coordinate instruction device operated by an operator;
Based on a predetermined reliability value indicating the accuracy of the detected coordinate value and the pressure value, the detected coordinate value is corrected by weighting and combining the detected coordinate value and the indicated coordinate value. A weighted synthesis means for calculating a value;
An on-line coordinate value correction device comprising: An online coordinate value correcting device that corrects a detected coordinate value indicating a position of a detection target, which is output together with a reliability value indicating accuracy from the position detecting device,
Detection information input means for inputting the detection coordinate value and the reliability value from the position detection device,
Instruction information input means for inputting an instruction coordinate value indicating a position indicated by the operator and a pressure value applied by the operator to the coordinate instruction apparatus from a coordinate instruction device operated by the operator ,
Based on the reliability value and the pressurization value, weighted synthesis means for calculating a corrected coordinate value by correcting the detected coordinate value by weighted synthesis of the detected coordinate value and the indicated coordinate value;
An on-line coordinate value correction device comprising: The weighted synthesis means includes
First weighting factor calculating means for calculating a first weighting factor from the reliability value based on a predetermined calculation formula;
A second weighting factor calculating means for calculating a second weighting factor from the pressure value based on a predetermined calculation formula;
For the added value of the product of the detected coordinate value and the first weighting factor and the product of the indicated coordinate value and the second weighting factor, the first weighting factor and the second weighting factor Coordinate value correcting means for calculating the corrected coordinate value by dividing the added value of
The on-line coordinate value correction apparatus according to claim 1, wherein the on-line coordinate value correction apparatus is provided. The weighted synthesis means includes
First observation covariance matrix calculation means for calculating the covariance matrix of the reliability value as a first observation covariance matrix observed in the detection information input means;
Second observation covariance matrix calculation means for calculating the covariance matrix of the pressurization value as a second observation covariance matrix observed in the instruction information input means;
Based on the detected coordinate value and the first observation covariance matrix, a state vector including a coordinate value indicating the position of the detection target and a state vector prediction estimation value and a state covariance matrix prediction estimation in which the state covariance matrix is predicted First filtering estimation means for calculating a temporary state vector and a temporary state covariance matrix that estimate the current state from the values;
Based on the indicated coordinate values and the second observation covariance matrix, the temporary state vector calculated by the first filter estimating means and the state vector filtered estimated value and the state covariance matrix filtered estimation obtained by correcting the temporary state covariance matrix are corrected. A second filter estimating means for calculating a value;
The state vector filtered estimated value and the state covariance matrix filtered estimated value calculated by the second filtered estimating means are delayed, and the state vector predicted estimated value and the state covariance matrix predicted are determined based on a predetermined state transition rule. A predictive estimating means for predicting an estimated value;
A selecting unit that selects a coordinate value indicating the position of the detection target from the state vector filtered estimated value calculated by the second filtered estimating unit, and sets the corrected coordinate value;
The on-line coordinate value correction apparatus according to claim 1, wherein the on-line coordinate value correction apparatus is provided. The position detection device detects a position of the detection target from an image,
Correction coordinate position figure drawing means for drawing a figure corresponding to the position indicated by the correction coordinate value;
Video composition means for synthesizing the graphic created by the corrected coordinate position graphic drawing means and the video;
The on-line coordinate value correction apparatus according to claim 1, further comprising: an image display unit including The position detection device detects a position of the detection target from an image,
Detected coordinate position figure drawing means for drawing a figure corresponding to the position indicated by the detected coordinate value;
A video synthesizing unit that synthesizes the graphic drawn by the detected coordinate position graphic drawing unit and the video;
The on-line coordinate value correction apparatus according to claim 1, further comprising: an image display unit including The position detection device detects a position of the detection target from an image,
Detected coordinate position figure drawing means for drawing a figure corresponding to the position indicated by the detected coordinate value;
Correction coordinate position figure drawing means for drawing a figure corresponding to the position indicated by the correction coordinate value;
Video composition means for combining the graphic created by the corrected coordinate position graphic drawing means, the graphic drawn by the detected coordinate position graphic drawing means, and the video;
The on-line coordinate value correction apparatus according to claim 1, further comprising: an image display unit including The detected coordinate value indicating the position of the detection target output together with the reliability value indicating the accuracy from the position detecting device, the indicated coordinate value output from the coordinate indicating device operated by the operator, and the coordinate indicating device To correct the computer based on the applied pressure value,
First weighting factor calculating means for calculating a first weighting factor based on the reliability value;
Second weighting factor calculating means for calculating a second weighting factor based on the pressure value;
For the added value of the product of the detected coordinate value and the first weighting factor and the product of the indicated coordinate value and the second weighting factor, the first weighting factor and the second weighting factor Coordinate value correcting means for calculating a corrected coordinate value by correcting the detected coordinate value by dividing the added value of
An on-line coordinate value correction program characterized by functioning as The detected coordinate value indicating the position of the detection target output together with the reliability value indicating the accuracy from the position detecting device, the indicated coordinate value output from the coordinate indicating device operated by the operator, and the coordinate indicating device To correct the computer based on the applied pressure value,
First observation covariance matrix calculating means for calculating a covariance matrix of the reliability values as a first observation covariance matrix;
Second observation covariance matrix calculating means for calculating the covariance matrix of the pressure value as a second observation covariance matrix;
Based on the detected coordinate value and the first observation covariance matrix, a state vector including a coordinate value indicating the position of the detection target and a state vector prediction estimation value and a state covariance matrix prediction estimation in which the state covariance matrix is predicted A first filtering estimation means for calculating a temporary state vector and a temporary state covariance matrix that estimate the current state from the values;
Based on the indicated coordinate values and the second observation covariance matrix, the temporary state vector calculated by the first filter estimating means and the state vector filtered estimated value and the state covariance matrix filtered estimation obtained by correcting the temporary state covariance matrix are corrected. Second filter estimating means for calculating a value;
The state vector filtered estimated value and the state covariance matrix filtered estimated value calculated by the second filtered estimating means are delayed, and the state vector predicted estimated value and the state covariance matrix predicted are determined based on a predetermined state transition rule. Predictive estimation means for predicting the estimated value;
A selection unit that selects a coordinate value indicating the position of the detection target from the state vector filtered estimation value calculated by the second filtering estimation unit, and sets the detected coordinate value as a corrected coordinate value;
An on-line coordinate value correction program characterized by functioning as

Images