JP2003256842A - Device and method for detecting position of moving object - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To configure an inexpensive system by reducing the arithmetic quantity of the position detection of a moving object projected on a highly precise image obtained by photographing the scene of a whole lecture by a remote lecture system. <P>SOLUTION: A difference between an image and the temporally past image obtained by photographing the same range is calculated, and the changed position is defined as the movement destination of a moving object. In this case, only the pixel data of linearly arranged comparison points extracted from the image are compared to calculate the different so that the arithmetic quantity required for detecting the position of the moving object can be reduced. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for detecting the position of a moving body existing in front of a background where no moving body exists in images continuously photographed in the same range. INDUSTRIAL APPLICABILITY The present invention is used in a technique of detecting position information of a lecturer in a remote lecture system that delivers image information to a remote place via a communication network.

[0002]

2. Description of the Related Art In a remote lecture system in which an instructor and a student are connected via a communication network and images and voices are distributed, the scenery of the entire lecture is photographed at a high image quality capable of reading the character information on the blackboard. By providing high-definition image information or realizing sound image localization at the student terminals, the presence of lectures is improved.

In such a remote lecture system for image distribution, since it is not always possible to secure a sufficient bandwidth available for communication, various techniques for reducing the amount of information to be distributed have been adopted. For example, a technique is adopted in which moving image information is extracted by image processing and only the moving image information is distributed.

In a remote lecture system, a moving part, that is, a moving body is often a lecturer. Therefore, in an image distribution system such as a remote lecture system,
Research has been conducted on detecting the position of a moving lecturer and delivering the position information.

As a technique for detecting the position of such a moving body, an image format such as JPEG which is highly efficient coding is used for recording image information, and the moving body is detected because the amount of information in the image changes as the subject changes. There is a detection method. This method is proposed by the inventors in Japanese Patent Laid-Open No. 2001-2181.
It is used in the image distribution system disclosed in Japanese Patent Publication No. 80, and in a remote lecture or the like, it is used for sound image localization at a terminal to which an image is distributed, thereby improving the realism of the lecture.

[0006]

However, unless the clothes of the instructor are significantly different from the background, the instructor's position detection accuracy is not so good because the amount of information in the image does not change significantly. In addition, since it is impossible to distinguish the change in the amount of information due to the movement of the teacher and the change in the amount of information due to the characters written on the blackboard in the background, the change in the amount of information due to the change in the background is erroneously detected as the change in the amount of information due to the movement of the teacher. There was an occasion.

In order to detect only the teacher without erroneously detecting background characters, it is necessary to perform image processing such as pattern matching on the image information. However, the time that can be used to detect the position of the instructor is until the next image is taken by the camera. In order to perform image processing on image information, an expensive information processing device with high processing capability is required, and there is a problem that the image distribution system becomes expensive. Furthermore, when performing sound image localization from the position information of the instructor, there is a problem that it can be applied to a stereo system, but cannot be applied to a surround system that generates more realistic sound when there is a depth in the background.

The present invention has been made against such a background, and the present invention reduces the amount of calculation for detecting the position of a target object that moves in the background and reduces the time required for position detection. An object of the present invention is to provide a moving body position detection device and method that can be shortened. Another object of the present invention is to provide a moving object position detection apparatus and method capable of easily detecting the position of a target object even when the background has a depth.
INDUSTRIAL APPLICABILITY The present invention can be used in a remote lecture system that delivers high-definition image information of a scene of an entire lecture captured to a student terminal with an image quality capable of reading the character information on a blackboard, and the position information of the lecturer can be obtained with an inexpensive configuration. An object and an object of the present invention are to provide a moving body position detecting device and method capable of delivering and giving a realistic lecture.

[0009]

According to the present invention, in a system for capturing an image including a moving body that continues in time, a previously captured image and a background image in which a moving body which is an object to be detected does not exist in a subject are compared. The position of the target object is determined by estimating the presence or absence of the target object by performing image processing including comparison processing on the image data at that position with the point at which the target object is desired to be detected as the comparison point.

That is, a first aspect of the present invention relates to a moving body position detecting apparatus, which is a photographing means for photographing a position-changeable target object in front of a background in time series, and the photographed image. The image processing means includes an accumulating means for accumulating in series, and an image processing means for detecting a position of the target object by comparing a captured image and the accumulated image, and the image processing means.
With respect to the image of the accumulating unit, a unit for extracting pixel data of a comparison point arranged so that at least one overlaps with the target object in the image, and pixel data of the comparison point of the extracted image and a captured image. And comparing the pixel data of the comparison point with the pixel data of the comparison point, and determining the position of the comparison point where the change is detected as the moving position of the target object.

The principle of the present invention will be described with reference to FIGS. The background image 20 is captured and prepared in advance. Since the size of the instructor 1 is approximately known for the shooting range 2 shown in FIG. 1, when the person is included in the image, the image is set as the comparison point 5 so that at least one pixel overlaps the person image. Set multiple measurement pixels inside. The current position of the instructor is estimated by comparing the measurement pixel at the comparison point 5 with the previous image. As a result, the range of image processing is limited to the pixel data of a specific comparison point from the entire surface of the image, so that the amount of calculation is reduced and the time required to detect the position of the instructor can be shortened.

Further, in the present invention, it is preferable that the comparison points have an interval equal to or smaller than the size of the target object, and the comparison points are arranged in one or more lines in the image.

This is to make the image cut out at the comparison points arranged linearly smaller than the size of the instructor 1. That is, as shown in FIG. 1, the distance between the comparison points 5 in the horizontal direction is set smaller than the width of the instructor 1,
The vertical direction is set smaller than the waist height in consideration of the case where the instructor 1 bends. As a result, any one of the comparison points 5 will always overlap the instructor 1 within the detection range of the instructor 1, so that the instructor 1 can be detected reliably.

In the present invention, the comparison point is
It is preferable that they are arranged linearly along the movement of the target object, and at least one line is arranged so as to intersect with the target object.

This is to set the moving body position detection line 6 along the moving direction of the instructor 1 as shown in FIG. 2 and set the comparison point 5 on the line.
When the shooting range is always fixed like the entire lecture landscape as in the remote lecture system, the height of the instructor 1 in the shooting range 2 can be estimated. By using this, the position of the instructor 1 is detected for one or more heights in a region lower than the height in which the instructor 1 is imaged, rather than for the entire position of the imaging range 2 to be the position detection target of the instructor 1. is there. Further, since the moving direction of the instructor 1 is mainly in the horizontal direction on the screen, the moving body position detection line 6 is set on the horizontal line and the image comparison point 5 is set on the line. By doing so, the moving object detection range can be efficiently specified, and the amount of calculation required for moving object position detection can be reduced.

Further, since the teacher 1 is photographed as a vertically long object, the change in the image information due to the movement of the teacher 1 appears in the vertical direction. At this time, by detecting the change in the image information at the same position in the horizontal direction for the two different heights, it is possible to detect the instructor 1 as a vertically moving object. Thus, by making the interval between the two different heights larger than the letters written on the background blackboard, it is possible to distinguish the letters from the lecturer and detect the moving lecturer.

Further, in the present invention, the distance between the lines of the comparison points is set to be small so that at least one line of the comparison points intersects the target object as the distance between the target object and the photographing means increases. Is preferred.

This is because when the size of the instructor 1 to be photographed changes due to the depth of the background as shown in FIG. 3, it is necessary to narrow the observation height interval in the portion where the size of the instructor 1 appears small. Is there. The position of the instructor 1 in the depth direction can be estimated from the set position of the moving body position detection line 6 where the moving body is detected. With this depth position, a surround system can be used, and a sound with a realistic sensation can be realized at the delivery destination.

Further, in the present invention, it is preferable that the means for determining the position of the moving body includes a means for detecting a difference in color or brightness of the pixel data at the immediately preceding comparison point.

This is to detect the image change by comparing the pixel data at the comparison point 5 with respect to each component of red, blue and green, or the luminance, brightness and saturation, and these information are used as the pixel data at the comparison point. Can be used.

Further, according to the present invention, the means for determining the position of the moving body is the difference between the color or brightness of the pixel data of the comparison point immediately preceding in time and the difference of the pixel data two or more previous in time. It is preferable to include means for detecting

This is the latest image 23 as shown in FIG.
The difference information between the previous captured image 22 and the previous captured image 22 is compared with the difference information between the previous captured image 22 and the pre-previous captured image 21 to detect the position of the moving body. The difference image generated at that time when taken with the passage of time 28 is as shown in FIG. 5, and in the difference image 26 between the latest image 23 and the previous taken image 22, the character 4 disappears when the lecturer moves. An image change is detected at two points of the source and destination of the lecturer 1. Since this alone cannot determine which of the two points is the destination, the change between this result and the difference image between the previous captured image 22 and the previous captured image 21 is detected. As a result, only the image change of the moving destination can be detected by the difference image 27 generated by filtering the image change of the moving source and used for detecting the moving body position.

Further, in the present invention, the means for determining the position of the moving body preferably includes means for detecting a difference in color or brightness with respect to the background image data in which the target object is not captured. Is preferably the latest image data in which it is determined that the target object is not captured.

As for the image change, a difference image is generated between the latest image 23 of FIG. 5 and the background image 20 in which the moving body is not shown, and the image change is detected. As the background image 20, it is desirable to use the latest image even if it is determined that there is no moving body. As a result, even if a character is written on the background blackboard 3 in a lecture or the like, it is possible to detect only the moving body as the difference information. Also, the previous captured image 2
When detecting a moving body from the latest image 23 using two images of 2 and the background image 20 as comparison images, a difference image 26 between the latest image 23 and the previous captured image 22 and a difference between the latest image 23 and the background image 20 are detected. Image 24 is generated. At this time, the image change between the moving source of the instructor 1 and the background character 4 can be filtered to generate a differential image used for detecting the moving body position, and the position of the moving body can be detected. further,
When the background changes, for example, even when the surrounding brightness changes, the position detection accuracy of the moving body is improved.

Further, the present invention can include means for detecting the distance between the target object and the photographing means by the number of comparison points intersecting the target object.

This is to estimate the position in the depth direction from the amount of comparison points where the change is detected on the moving body position detection line 6. In the shooting landscape as shown in FIG. 4, the nearby person 7 obviously has a larger amount of overlap with the comparison point than the distant person 8. Accordingly, since the depth position of the moving body can be detected, it is possible to provide a realistic sound by using the surround system at the delivery destination.

Further, according to the present invention, there are two or more sets of comparison points in which the overlap with the target object in the image is detected,
When the size of the set of comparison points is substantially equal to the size of the target object, it is preferable to determine one of the sets of comparison points as the position of one target object and detect the positions of two or more target objects. .

Even if there are a plurality of target objects, if the size of the target object is known, it overlaps with the target object, and if the size of the set is substantially equal to the size of the target object, One can be regarded as overlapping with one target object, and the position of the set of the comparison points can be determined as the position of the target object. For example, if the set of comparison points is twice as large as the target object, it is recognized as two objects, and if the size is less than double, it is recognized as one. Even if an OHP screen is included in the image, the size of the OHP screen is different from that of the instructor.
There is no mistakenly recognizing the screen of as a teacher.

Another aspect of the present invention relates to a moving body position detecting method, in which a variable position target object in front of the background is continuously photographed in time series, and the photographed image data is recorded. Pixel data at a position where at least one of the comparison points linearly arranged in the image and the target object are overlapped is extracted from the past image data and the captured image data that are accumulated in time series. When the extracted pixel data are compared and a change is detected, the position of the changed comparison point is determined as the moving position of the target object.

As a result, by limiting the range in which the image processing is performed to the pixel data of a specific comparison point, it is possible to reduce the amount of calculation for detecting a moving object and reduce the calculation time.

According to the present invention, the information processing apparatus is installed in an information processing apparatus to which a photographing means for time-sequentially photographing a variable-position target object in front of the background is connected, and the image taken by the information processing apparatus is recorded. An image processing unit that detects the position of the target object by comparing a captured image with an accumulated image, and the image processing unit, for the image of the accumulating unit, Means for extracting pixel data of a comparison point arranged so that at least one overlaps with the target object in the image, and pixel data of the comparison point of the extracted image and pixel data of the comparison point of the captured image, And means for determining the position of the comparison point at which the change is detected as the moving position of the target object to operate as a moving body position detecting device. This program can be provided via a recording medium or a communication line.

[0032]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. The following embodiment is an example in which the moving body position detecting device of this embodiment is used as an image distribution device for a remote lecture.

FIG. 6 shows the configuration of an image distribution device of a remote lecture system including the moving body position detecting device 16 according to the embodiment of the present invention. That is, the camera 13 that continuously photographs the instructor 1 as a subject in time series, and the camera 13
A moving body position detecting device 16 including a configuration for detecting and determining the position of the instructor 1 by processing the image captured by the device, and the distribution of delivering the captured image together with the detected position information of the instructor 1 to the terminal of the student at the remote place. And a device 17.

Here, the feature of the embodiment of the present invention is that the moving body position detecting device 16 has a photographed image recording unit 9 for accumulating images photographed by the camera 13 in time series, and this photographed image recording unit 9. From the accumulated images, the pixel data of the comparison point 5 arranged so that at least one overlaps with the instructor 1 as the moving body to be detected in the image is extracted, and the pixel data of the comparison point 5 is used for the new image and the past image. The pattern comparison unit 10 and the moving body detection determination unit 11 are provided as means for determining the position of the comparison point where the change is detected by comparing with the position as the moving position of the lecturer 1.

The detected position information of the instructor 1 is output to the distribution device 17 via the image / moving object detection information combining unit 12 that combines the photographed image and the detected moving object position information.

FIG. 7 shows the configuration of a remote lecture system to which this embodiment is applied.

The teacher 1 is standing and moving in front of the blackboard 2,
The camera 13 continuously photographs it with time. This photographed image is a prior application, Japanese Patent Application No. 2000-02656.
A high-definition still image as shown in 8 or a moving image usable in a remote lecture system may be used. The blackboard 3 may be a whiteboard or a screen of a projector.

The image photographed by the camera 13 is image-processed by the moving body position detecting device 16, and the signal of the microphone receiver 15 which receives the radio signal of the voice signal from the radio microphone transmitter 14 which collects the pronunciation of the instructor is processed. Then, the distribution device 17 that is combined with the image processing result of the moving body position detection device 16 and is sent to the distribution destination, the communication network 18 such as the Internet or a LAN, and the distribution device 1 via this communication network 18.
This remote lecture system is provided with a student terminal 19 that outputs and displays the distribution data from 7. There may be a plurality of student terminals 19, and the distribution device 17 sends the captured image signal from the moving body position detection device 16 and the position information and voice information of the instructor to the student terminal 19, and the student terminal 19
Sound image localization is performed based on the instructor's position information and output to demonstrate a realistic distance lecture.

Next, the moving body position detecting operation for detecting the position of the instructor will be described with reference to FIGS. 1 to 7.

The image photographed by the camera 13 is taken into the photographed image recording section 9. In the captured image recording unit 9, at least three images, that is, the latest image 23 shown in FIG. 5, the previous captured image 22 captured immediately before, and the background image 20 in which the lecturer is not captured are recorded. For these three images, three image data are input from the photographed image recording unit 9 to the pattern comparison unit 10 every time photographing is performed to detect the position of the moving body. In addition, the latest image 23 for distribution is input to the image / moving image detection information combining unit 12.

The pattern comparison section 10 generates a difference image of the color or brightness of the pixel data of the comparison point 5 of the moving body position detection line 6 shown in FIGS. Comparison point 5 is as shown in Figure 1.
It may be set linearly or linearly in the image as shown in FIGS. As the difference information used for detecting the moving body position, two pieces of difference information, that is, the difference information between the latest image 23 and the previous captured image 22 and the difference information between the latest image 23 and the background image 20 in FIG. 5 are generated. The pixel data is compared for each of the corresponding comparison points with respect to the two pieces of difference information, and a value with a small change is set as the final difference information of the comparison point. The generated difference information is input to the moving body detection determination unit 11 that determines the position of the instructor 1 from the difference information.

When the difference information is generated, the previously photographed image 22
The pre-previously captured image 21 or the background image 20 may be used as a sample image for comparison. Thus, by using the difference image 25 between the previous captured image 22 and the pre-previous captured image 21 and the difference image 24 between the background image 20 and the latest image, the change in the source image is removed, and the change in the destination image is removed. Only detect. As a parameter for detecting an image change, a difference in color or brightness is used to determine the image change.

When the moving body detection / judgment unit 11 judges that there is the lecturer 1, the moving body position detection result is input to the image / moving body detection information combining unit 12 and is combined with the latest image 23, and then the distribution device 17 is provided. Is input to the student terminal 19 via the communication network 18. If it is determined that the instructor 1 does not exist, an update request for the background image 20 is made to the captured image recording unit 9 to update the background image 20, and the latest image 23 is recorded as the background image 20. Further, the latest image 23 is updated and is also used as the previous captured image 22 in the next capturing process. If there is a latest image in which it is determined that the instructor 1 is not shown, the image is updated in the image recording unit 9 as the background image data 20.

(Modification of the Embodiment) As shown in FIG. 2, in the example of the remote lecture in which the blackboard 3 is in the shooting range 2 and the instructor 1 moves in the shooting range 2, characters are written on the blackboard 3 in the background. It is necessary not to detect the entry or deletion of item 4 as the movement of the instructor 1 by mistake. For this purpose, it is necessary to perform not only the presence / absence of the change in the image information but also the pattern matching to confirm whether or not the change in the image information occurs due to the movement of the teacher 1. This pattern matching can be performed by detecting the moving body position with respect to the two moving body position detection lines 6. Two moving object position detection lines 6
If the size of the interval is set to be wider than the size of the character 4, the character 4 is detected as a change in the image information only in one moving body position detection line 6, and the instructor 1 makes an image in the two moving body position detection lines 6. Detected as a change. Therefore, if it is determined that the teacher 1 moves when the two moving body position detection lines 6 change at each comparison point 5, and that the character 4 is entered or deleted when only one moving body position detection line 6 changes. Good. As a result, the instructor 1 and the character 4 can be detected separately by using the plurality of moving body position detection lines.

Further, in the captured image of the remote lecture, the moving direction of the instructor 1 is the horizontal direction as shown in FIGS. 2 to 4. Therefore, it is desirable that the moving body position detection line is the horizontal direction with respect to the captured screen.

(Modification of Embodiment) It is also possible to use it as a monitoring system for a place having a depth as shown in FIGS. In FIG. 3, the distance between the comparison point rows (position detection lines 6) is shortened as the distance between the target object as a subject and the camera increases, so that no matter where the person is, one moving object position detection line is always Place it so that it overlaps with the person. As a result, the position of the detection target person can be detected in the horizontal direction or the depth direction. Also, FIG.
As shown in, the number of comparison points in which a change is detected on the moving body position detection line 6 is larger in the person 7 near the person than in the person 8 far away,
If the number of comparison points where the change in the image information is detected is large, it is determined that the comparison points exist near the camera, and if the number is small, it is determined that the comparison points are located far. As a result, the position of the person can be detected in the horizontal direction and the depth direction of the screen.

(Modification of Embodiment) FIG. 8 is a diagram for explaining position detection when a plurality of target objects are present in an image.

When the assistant 7 in addition to the instructor 1 is on the screen, there can be two sets of comparison points 5 overlapping the target object. The number of comparison points 5 overlapping with the instructor 1 is three, and the number of comparison points 5 overlapping with the assistant 7 appearing in the image is also three. There are two sets of comparison points that overlap with. Further, since the number of comparison points constituting the set is three, the set of comparison point sizes is twice the size of the target object, and the size of one set is less than double the size of the target object. There are two detection targets, and it is possible to determine the collective position of the respective comparison points as the positions of the instructor 1 and the assistant 7. Even if the image includes an OHP screen, it is different from the size of the instructor. Therefore, by judging the size of the set of comparison points,
The change in OHP will not be mistaken for the movement of the teacher's position.

In this embodiment, the information processing apparatus is installed in a general-purpose information processing apparatus, and an image captured by the information processing apparatus is processed. It can also be realized as a program that realizes the functions of the pattern comparison unit 10, the moving body detection determination unit 11, and the image / moving body detection information combining unit 12 to function as the moving body position detection device 16. In this case, the program can be recorded on a recording medium or provided via a communication line.

[0050]

As described above, according to the present invention, a moving object position detection of a high-definition image, which conventionally requires an expensive image processing apparatus, can be realized by an inexpensive information processing apparatus, and the accuracy of moving object detection is improved. You can also Further, by using the detected position information of the instructor, it is possible to transmit a realistic audio signal compatible with the surround system to the distribution destination.
According to the present invention, it is possible to construct a distance learning system that is inexpensive and can perform distance learning with high-definition images.

Further, the moving body position detecting apparatus and method of the present invention can be used not only for remote lectures but also for a monitoring system or a video conference for detecting the position of a moving body and transmitting it to the transmission destination.

[Brief description of drawings]

FIG. 1 is a diagram showing an arrangement example of comparison points in an image.

FIG. 2 is a diagram showing an example in which comparison points in an image are linearly arranged.

FIG. 3 is a diagram showing an arrangement example of a moving body position detection line 6 when a moving body is movable in the depth direction.

FIG. 4 is a view showing an arrangement example of a moving body position detection line 6 when a moving body is movable in the depth direction.

FIG. 5 is a diagram showing an example of a captured image and a difference image arranged on a time axis.

FIG. 6 is a block diagram showing the configuration of a moving body position detection device according to an embodiment of the present invention.

FIG. 7 is a diagram showing a configuration of a remote lecture system according to an embodiment of the present invention.

FIG. 8 is a diagram showing a case where a plurality of target objects are present on the screen.

[Explanation of symbols]

1 teacher 2 shooting range 3 blackboard 4 letters on the blackboard 5 comparison points 6 Moving object position detection line 7 assistant 9 Captured image recording section 10 pattern comparison section 11 Moving object detection determination unit 12 Image / moving object detection information combining section 13 cameras 14 microphone transmitter 15 microphone receiver 16 Moving object position detector 17 Delivery device 18 communication network 19 Student terminal 20 background images 21 Image taken before last 22 Previous image 23 latest image 24 Difference image between latest image and background image 25 Difference image between the previous and previous images 26 Difference image between latest image and previous image 27 Difference image used for moving object position detection 28 hours have passed

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Tomohide Takano             3-1, Hikarinooka, Yokosuka City, Kanagawa Stock Association             Company Optowave Research Center F term (reference) 2C028 AA00 BB04 BD02 CA12 CA13                 5C054 DA09 FC01 FC03 FC07 FC12                       FC13 GA04 GB01 GB14 HA16                 5L096 AA02 CA04 FA59 FA69 GA08                       GA40 GA41 HA04

Claims (11)

[Claims] 1. A photographing means for photographing a variable position target object in front of a background in time series, a accumulation means for accumulating the photographed images in time series, the photographed image and the accumulated image. And an image processing unit that detects the position of the target object by comparing the image processing unit and the image processing unit, so that at least one of the stored image and the captured image overlaps the target object in the image. Means for extracting the pixel data of the arranged comparison points and means for comparing the pixel data of the comparison points of the extracted image and determining the position of the comparison point where the change is detected as the moving position of the target object. A moving body position detecting device including. 2. The moving body position detecting apparatus according to claim 1, wherein the comparison points are arranged at intervals equal to or smaller than the size of the target object, and the comparison points are arranged in one or more lines in the image. 3. The moving body position detecting apparatus according to claim 2, wherein the comparison points are linearly arranged along the movement of the target object, and at least one line is arranged so as to intersect the target object. . 4. The distance between the comparison points is set to be small so that at least one line of the comparison points intersects the target object as the distance between the target object and the photographing means increases. 2. The moving body position detection device according to 2 or 3. 5. The moving body position detection according to claim 1, wherein the determining means includes means for detecting a difference in color or brightness of pixel data of the comparison point immediately preceding in time. apparatus. 6. The determining means includes means for detecting a difference between a color or brightness difference of pixel data of the comparison point immediately preceding in time and pixel data two or more previous in time. The moving body position detecting device according to claim 1. 7. The moving body position detecting apparatus according to claim 1, wherein the determining means includes means for detecting a difference in color or brightness with respect to background image data in which the target object is not captured. . 8. The moving body position detecting apparatus according to claim 7, wherein the background image data is the latest image data in which it is determined that the target object is not captured. 9. The moving body position detecting device according to claim 1, further comprising means for detecting a distance between the target object and the photographing means by the number of comparison points intersecting with the target object. 10. When there are two or more sets of comparison points in which an overlap with the target object is detected in the image and the size of the set is substantially equal to the size of the target object,
10. The moving body position detecting apparatus according to claim 1, further comprising means for determining one of the sets of comparison points as a position of one target object. 11. A variable-position target object in front of a background is continuously photographed in time series, the photographed image data is accumulated in time series, and accumulated past image data and photographed images are collected. From the data, pixel data at a position where at least one of the comparison points linearly arranged in the image and the target object overlap is extracted, and when a change is detected by comparing the extracted pixel data, A moving body position detecting method for determining the changed position of the comparison point as the moving position of the target object.