JP2008227636A - Rotating object image processor, rotating object image processing method and rotating object image processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To execute automatic adjustment so as to turn an object to the optimum size within a visual field of a camera when photographing the rotating object with the same camera. <P>SOLUTION: In the rotating object image processor, a rotating object detection part 13 inputs the image of the rotating object from a video camera 15 and detects an area occupied by the rotating object on the image by image processing. A magnification calculation part 14 determines the zoom magnification of the video camera 15 on the basis of the obtained occupancy area and executes adjustment. Thus, the rotating object is photographed in the optimum size from the photographing of the second rotation. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an apparatus for photographing and recording a rotating object (member), and relates to a rotating object image processing apparatus that automatically adjusts the photographed object so as to have an optimum size within the field of view of the camera. The present invention relates to an apparatus for recording an image (or a plurality of images) of a member being manufactured or manufactured.

  Conventionally, when a plurality of members having different sizes are recorded and stored sequentially by video manually, the operator has to adjust the zoom magnification each time so that the members can be accommodated in the camera field of view. In particular, in order to check the finished state of the member from each direction, when the member is rotated, it is necessary to finely adjust the zoom magnification so that the member does not deviate from the camera field of view.

  On the other hand, as a method for automatically adjusting the zoom magnification regardless of human hands, a method has been proposed in which an object becomes the optimum size within the camera field of view while tracking with a pan-tilt camera or the like when shooting a moving object. ing.

  (1) The image processing is used to detect the contour of the object to be photographed, and the zoom lens is controlled based on the size of the object, and the object is photographed so as to have an optimum size (Patent Document 1).

  (2) A person region is extracted by skin color detection, and tracking and zooming operations are realized by cutting out the corresponding region (Patent Documents 2 and 3).

(3) Further, the size of a person is acquired by detecting a face and a hand by skin color detection, and tracking and automatic zooming are performed with a pan / tilt camera (Patent Document 4).
Japanese Patent Laid-Open No. 9-238279 JP-A-6-268894 JP 2004-282535 A JP-A-5-219421

  When the methods described in Patent Documents 1 to 4 are used, when the member rotates and the apparent size changes, enlargement and reduction of the zoom magnification are repeated. However, if the zoom magnification changes in conjunction with the rotation of the member, the scale of the member when viewed from the front or side changes (corresponding to the front view or side view of the trigonometric drawing), so the overall appearance is grasped. I cannot answer the need to do. That is, there is a problem that the zoom magnification during rotation of the member is constant and the entire image cannot be taken.

  SUMMARY OF THE INVENTION An object of the present invention is to solve the above-described problems and provide an apparatus for photographing a rotating object at an optimum zoom magnification. In particular, when a rotating object is photographed with the same camera, the object is to automatically adjust the object so that it has an optimum size within the camera field of view.

  In order to solve the above-described problems, the present invention particularly includes a rotating body detecting unit and a magnification calculating unit, and the rotating body detecting unit inputs an image of a rotating object from a camera and the rotating object. Detects the area occupied by the image by image processing, and the magnification calculation means determines and adjusts the camera zoom magnification based on the obtained area, thereby capturing the rotating object at the optimum size. It is characterized by doing. Hereinafter, an example in which an object to be photographed is a product member will be described.

  In the present invention, first, a rotating member is photographed with the zoom magnification of the camera set to a wide angle, and the member region is detected by image processing. At this time, the member region in the frame image is stored until the turntable on which the member is placed makes one rotation. After one rotation, all the stored member regions are overlapped to determine the region occupied by the member during rotation. Based on the obtained member occupation area, the camera zoom magnification is determined and adjusted, and video recording is started.

  In addition, the present invention sets the zoom magnification of the camera to a wide angle, detects the member region in the image when the member makes one rotation, detects each member in each frame, and superimposes all stored member regions after one rotation. By combining, the area occupied by the member during rotation is obtained, and the zoom magnification of the camera is determined and adjusted based on the obtained area occupied by the member, and an image is recorded.

  Alternatively, in the present invention, the zoom magnifications of the first and second rotation cameras are kept constant, and the second rotation is performed by extracting and enlarging the rotating member occupation area from the acquired frame image, A pseudo zoom image is generated.

  Alternatively, the present invention associates the zoom magnification calculated from the rotating member occupation area with the frame image, and stores both the zoom magnification and the frame image. After one rotation, a minimum zoom magnification is obtained, and a zoomed image is obtained by enlarging a stored frame image based on the obtained minimum zoom magnification.

  In the techniques of Patent Documents 1 to 4, since there is no means for accumulating and superimposing the object area, automatic zoom magnification adjustment for a rotating object cannot be realized. The present invention has means for adjusting a zoom magnification or generating a pseudo zoom image so that a rotating member that is an object to be photographed has an optimum size within a camera field of view. And very different.

  According to the present invention, when the images of a plurality of objects having different sizes are rotated and sequentially recorded and stored, the zoom magnification is automatically set. Therefore, the operator needs to adjust the zoom magnification each time. The work burden associated with shooting can be greatly reduced. Further, according to the present invention, since an image to be recorded is obtained at a constant zoom magnification while one object is rotating, the apparent size of the object does not change during one rotation. .

  FIG. 1 shows a block diagram of an apparatus according to an embodiment of the present invention. In the following description, the rotation direction of the object is described as being about the vertical axis for easy understanding, but the same applies even when the rotation axis is orthogonal to the vertical axis or when the rotation axis changes with time. This method can be used.

  In FIG. 1, the rotating object image processing apparatus 10 includes signal input / output interfaces (I / F) 11 and 12 for inputting / outputting signals from / to the outside, a rotating body detection unit 13, and a magnification calculation unit 14. Further, in this example, a video camera 15 and a turntable 16 are connected to the signal input / output interface 11 of the rotating object image processing device 10, and the display device 17 and a magnetic field for storing video are stored in the signal input / output interface 12. A disk device or other recording medium 18 is connected. The rotating body detection unit 13 and the magnification calculation unit 14 can be realized by a CPU, a memory, a software program, and the like.

  Hereinafter, the processing contents of each unit in the block diagram shown in FIG. 1 will be described in detail. In the present embodiment, an example in which a moving image is captured by the video camera 15 will be described. However, when a shutter is continuously released at regular intervals by a normal camera instead of a video camera, the image is captured like a continuous photograph. In addition, the present invention can be applied as in the case of a video camera. In other words, the present invention can be applied to either continuously captured images (moving images or videos) or intermittently captured images.

[Video camera 15]
The video camera 15 forms an image at a constant sampling time interval (for example, 33 ms) from the light intensity detected by the CCD or the like, and further generates a moving image.

  One pixel of an image is given by a gradation value of a red component (R), a green component (G), and a blue component (B) in the case of a color image, and given by a gradation value of luminance in the case of a monochrome image. . For example, the gradation values of the red component (R), green component (G), blue component (B), and luminance (I) of the pixel at coordinates (x, y) indicated by integers x and y are digital values, respectively. R (x, y), G (x, y), B (x, y), I (x, y). Note that the luminance (monochrome gray value) is used as the pixel value used in the following calculation.

[Rotating body detection unit 13]
The rotating body detection unit 13 inputs an image (or an image) from the video camera 15 via the signal input / output interface 11, and an object placed on the rotating table 16 by a rotation sensor provided on the rotating table 16. And the frame image of the video input from the video camera 15 and the rotation angle of the object are stored in association with each other. The rotation angle may not be input from the turntable 16, but may be input from an actuator that rotates the turntable 16, and the rotation angle may be detected by image analysis.

  In the present embodiment, a method for detecting a rotating body from a frame image is, for example, a method combining a background difference method and a local correlation method.

  The background difference method is a method in which an image that does not include a rotator is prepared in advance as a background image, and then a difference from an image that includes a rotator is obtained, and a portion having a large difference is used as an object region. The background subtraction method can be processed in a shorter processing time, but has a drawback that it is vulnerable to noise such as brightness fluctuations.

  In the local correlation method, first, an image is divided into grid-like blocks (for example, 8 × 8 pixels). Next, a correlation value between neighboring blocks is obtained between a frame at a certain sampling time and a frame acquired at the next sampling time, and it is determined whether the block has moved according to the magnitude. This process has a large amount of calculation, so there are disadvantages such as that the image size cannot be increased, the distance between blocks that perform correlation calculation is limited, and it is not possible to follow fast-moving objects. High tolerance. Therefore, a rough rotator region is obtained by the background subtraction method, and then a procedure for determining the rotator region more finely by the local correlation method is used.

  Prior to the description of the processing procedure, a sub-partition, a sub-partition template that is a collection of sub-partitions of the same object, and a tracking block are shown in FIG. 2 and will be described. In FIG. 2, 20 is an acquired image obtained from the video camera 15, 21 is a sub-partition, 22 is a sub-partition template, and 23 is a tracking block.

  The sub-partition 21 is composed of a square of vertical n pixels and horizontal n pixels. The sub-partition template 22 is a collection of sub-partitions 21 and indicates a region of the rotating body. The tracking block 23 is an area which is set so as to surround the sub-partition template 22 and is provided between the upper, lower, left and right m pixels outside the sub-partition template 22. The search by the local correlation method is executed in the area surrounded by the tracking block 23. In the present embodiment, the image area of the rotating body is represented by the sub-partition template 22.

  An area where a rotator may exist is obtained using the background subtraction method, and then an effective sub-partition for detecting the rotator is calculated using a local correlation method. FIG. 3 is a diagram illustrating a rotating body detection procedure by the rotating body detection unit 13. Hereinafter, a procedure in which the rotating body detection unit 13 confirms whether or not a rotating body is included in the captured image and detects an effective sub-section will be described with reference to FIG.

(A) Processing in Step S10 Bitmap format images obtained by the video camera 15 are fetched at regular intervals (every sampling time). The acquired images at the current timing are divided into two series and passed through low-pass filters having different cutoff frequencies in the next steps S11 and S12.

(B) Processing in Step S11 A background image of the current frame with a slow motion in which the cutoff frequency of the low-pass filter is set low is obtained. This cut-off frequency can be obtained corresponding to the moving speed of the rotating body, and can be set variably according to the target rotating body. The background image is divided into sub-partitions (n × n pixel squares).

(C) Processing in Step S12 The cut-off frequency of the low-pass filter is set higher than that of the low-pass filter in Step S11, and an image from which influences such as flickering of a fluorescent lamp are removed is obtained while leaving a moving object. This cut-off frequency can also be set variably according to the target rotating body, similarly to the cut-off frequency of the low-pass filter used in step S11. This image is the current image at the current sampling time. As in step S11, the image is divided into sub-partitions (n × n pixel squares).

(D) Processing in Step S13 In the current image (hereinafter referred to as the current frame image) and the background image, one of the sub-sections (n × n pixel square) at the same position is extracted.

(E) Processing in Step S14 The total sum T of the differences between the pixels in the sub-partitions at the same position in both images is calculated by the following equation.

T = Σ | I ^C (x, y) −I ^B (x, y) |
(Σ represents the sum of x, y∈R _SUB )
Here, I ^C (x, y) is the black and white gray value of the pixel position (x, y) in the current image, I ^B (x, y) is the black and white gray value of the pixel position (x, y) in the background image, R _SUB represents a sub-partition. When the calculated value T is equal to or greater than a predetermined threshold value T _th, it indicates that the calculated value T is different from the background image. In other cases, the process returns to step S13 to take out the next sub-partition.

(F) Processing in Step S15 Using the local correlation method, it is searched from which sub-section of the previous frame image the moving sub-part determined to have motion in step S14. Search using correlation calculation to determine which sub-section Sq (x ′, y ′) of the previous frame image has moved from the sub-section Sp (x, y) of the current frame image that is focused as a moving sub-section. To do. The correlation value C with the sub-partitions Sp (x, y) and Sq (x ′, y ′) is calculated by the following equation, and it is determined whether or not the movement is from a predetermined threshold value.

C = Σ | I ^p (x, y), I ^q (x ′, y ′) |
(Σ represents the sum of x, y∈Rp and x ′, y′∈Rq)
In the above equation, I ^p (x, y) and I ^q (x ′, y ′) are the black and white grayscale values of the pixels in each sub-partition Sp (x, y), Sq (x ′, y ′), Rp , Rq indicates that it is within the sub-partition area.

  The presence or absence of movement is determined using the so-called zero comparison method. The difference Q between the correlation value C0 of the sub-part of the previous frame at the same position as the sub-part of interest and the minimum value C1 of the correlation value obtained from the correlation calculation with all the sub-parts within the scanning range is expressed as follows: Calculate with the following formula.

Q = C0-C1
(G) Processing in Step S16 The calculated difference Q is compared with the movement determination threshold value _Qth .

Q> Q _th
If Q is larger than the threshold value Q _th , the sub-part of interest Sp (x, y) is determined as a sub-part moved from Sq (x ′, y ′), and the process of step S17 is performed as an effective sub-part. Do. Otherwise, the process of step S17 is skipped.

(H) Processing in Step S17 The speed and direction of movement from the previous frame image of the effective sub-section are calculated. The upper left point of the sub-part of the previous frame image is (x _LeftTop1 , y _LeftTop1 ), the upper left point of the sub-part of the current frame image of interest is (x _LeftTop2 , y _LeftTop2 ), and the sub-part movement vector (x _LeftTop2− x _LeftTop1 , y _LeftTop2− y _LeftTop1 ), the speed of movement and the direction of movement are as follows.

Speed of movement =
{(X _LeftTop2 −x _LeftTop1 ) ² + (y _LeftTop2 −y _LeftTop1 ) ² } ^1/2
Direction of movement =
tan ⁻¹ {(y _LeftTop2 −y _LeftTop1 ) / (x _LeftTop2 −x _LeftTop1 )}
(I) Processing in Step S18 The process returns to Step S13 and the processing is repeated in the same manner until the background has been confirmed in all the sub-sections. Through a series of processing, all effective sub-parts, the speed of movement from the previous frame of each effective sub-partition, and the direction of the movement vector are obtained.

  Next, the image area of the rotating body is extracted based on the effective sub-section obtained above. FIG. 4 is a diagram illustrating a procedure for extracting candidates for a rotating body image region by the rotating body detection unit 13. From the set of effective sub-partitions obtained in the processing of FIG. 3, an effective sub-partition template for confirming the rotating body is generated, and tracking block candidates are set from the effective sub-partition template. Hereinafter, the procedure by which the rotator detection unit 13 extracts candidates for the rotator image area will be described with reference to FIG.

(A) Processing in Step S20 One is taken out from the set of effective sub-partitions and is set as an effective sub-partition template.

(B) Processing in Steps S21 and S22 The next one of the effective sub-sections is taken out, and it is compared whether the speed and direction of movement of both coincide. If they match, the process proceeds to step S23, and if they do not match, the process proceeds to step S24.

(C) Processing in Step S23 Both effective sub-partitions are combined to generate one sub-partition template. This sub-partition template is set as an effective sub-partition template.

(D) Processing in Step S24 The effective sub-partition is set as another effective sub-partition template.

(E) Processing in step S25 The processing from step S21 to step S24 is repeated until confirmation with all the valid sub-partitions is completed.

(F) Processing in Step S26 The area obtained by adding m pixels from the upper, lower, left and right ends of all the effective sub-partition templates defined in Steps S23 and S24 is set as a tracking block candidate.

The tracking block candidate area is calculated as follows. The top, bottom, left and right edges of the effective sub-partition template in step S26 are (x _sub-min ), (x _sub-max ), (y _sub-min ), and (y _sub-max ). Assuming that an area given a predetermined pixel value m outside the area surrounded by the upper, lower, left, and right ends is (> 0), the tracking block candidate area is as follows.

x _sub-min −m ≦ x ≦ x _sub-max + m
y _sub-min −m ≦ y ≦ y _sub-max + m
The center position (x _trk-cen , y _trk-cen ) of the tracking block candidate is as follows.

x _trk-cen = (x _sub-min + x _sub-max ) / 2
y _trk-cen = (y _sub-min + y _sub-max ) / 2
A tracking block for determining a rotator is extracted from the tracking block candidate region obtained by the processing of FIG. 4, and the rotator is determined. The rotation body is determined by calculating a movement vector of the rotation body.

  FIG. 5 is a diagram showing a rotating body determination procedure by the rotating body detection unit 13. Hereinafter, the procedure in which the rotating body detection unit 13 determines the rotating body will be described with reference to FIG.

(A) Processing in Step S30 The movement vector from the previous frame image of the rotator existing in the tracking block candidate is calculated. If the tracking block center position of the previous frame image (x _trk-cen1 , y _trk-cen1 ) and the tracking block candidate center position of the current frame image are (x _trk-cen2 , y _trk-cen2 ), the current frame image is rotated. Speed v _trk ^(k) and direction vector (d _trk-x , d _trk-y ) of the body movement vector (x _{trk-cen 2} -x _{trk-cen 1} , y _{trk-cen 2} -y _{trk-cen 1} ) ^(k) is as follows.

Speed of movement v _trk ^(k) =
{(X _trk-cen ² −x _{trk-cen 1} ) ² + (y _trk-cen ² −y _{trk-cen 1} ) ² } ^1/2
Direction of movement =
tan ⁻¹ {(y _trk-cen2 −y _trk-cen1 ) / (x _trk-cen2 −x _trk-cen1 )}
(B) Processing in Step S31 One of the tracking block candidates of the current frame image is extracted.

(C) Processing in Step S32 One of the tracking blocks of the previous frame image is taken out.

(D) Processing in Steps S33 and S34 The movement vector of the current frame image is calculated, and the tracking condition is calculated from the movement vector in the previous frame image.

The tracking conditions are as follows. v _th and d _th are predetermined threshold _values .

| V _trk ^(k) -v _trk ^(k-1) | <v _{th
│d trk-x} ^(k)・ d _trk-x ^(k-1) + d _trk-y ^(k)・ d _trk-y ^(k-1) │ ＞ d _th
Here, v _trk ^(k-1) and (d _trk-x , d _trk-y ) ^(k-1) indicate the speed of movement of the movement vector of the previous frame image and the direction of the movement vector.

  When the speed difference between the tracking block candidate and the tracking block movement vector is less than a certain value and the length is more than a certain value, the rotating bodies of both tracking blocks are regarded as the same tracking object. The length is determined by the inner product of the moving direction vectors. If the tracking condition is satisfied, the process proceeds to step S35. If the tracking condition is not satisfied, the process proceeds to step S38.

(E) Processing in Step S35 Setting and confirmation of the continuous number when the frame continuous condition is set as the tracking condition is performed. Let FR be the number of consecutive frames, and add 1 to FR.

(F) depending on whether the process FR in step S36 is larger than the threshold value FR _th, to determine whether the frame sequential condition is satisfied, when the frame sequential condition is satisfied, the process proceeds to step S37, if not satisfied, step Proceed to S39.

(G) Processing in Step S37 It is recognized that a rotator exists in the tracking block candidate area, and the tracking block candidate is replaced with the tracking block of the current frame image and held. Thereafter, the process proceeds to step S39.

(H) Processing in Step S38 When the tracking condition is not satisfied, it is regarded as another rotator newly appearing in the current frame image, and the tracking block candidate is held as a new tracking block. The tracking condition FR = 1 is set, and the process proceeds to step S39.

(I) Processing in Step S39 The processing in steps S32 to S38 is repeated until confirmation of all the tracking blocks in the previous frame is completed. All tracking blocks in which a rotator recognized in the current frame image obtained by this series of processing exists are obtained. The rotating body of each tracking block is a different rotating body.

[Magnification calculator 14]
FIG. 6 is a diagram for explaining a method of calculating the zoom magnification.

  The magnification calculation unit 14 calculates a magnification for the object to have an optimum size within the camera field of view. It is assumed that the zoom magnification is set to a wide angle so that the rotating body does not protrude from the camera field of view at the start of shooting.

By the processing described above, the upper and lower sides and the left and right sides of the sub-partition template of the tracking block obtained by the rotator detection unit 13 are
y = y _sub-min
y = y _sub-max
x = x _sub-min
x = x _sub-max
It is represented by At this time, as shown in FIG. 6, the intersection of each side of the sub-partition template and the diagonal line of the image is obtained, and the visual field range is determined by the distance from the center of the image. When the pixel size in the horizontal and vertical directions of the image is x _size and y _size , respectively,
y = (y _size / x _size ) · x
Therefore, the intersection with each side is as follows.

((X _size / y _size ) · y _sub-min , y _sub-min )
((X _size / y _size ) · y _sub-max , y _sub-max )
((Y _size / x _size ) _xsub-min , _xsub-min )
((Y _size / x _size ) _xsub-max , _xsub-max )
The distance between the calculated intersection and the center of the image is obtained, and the intersection with the maximum distance is determined. For example, the distance from the image center point of the intersection of the upper side and the diagonal line is
[{(X _size / y _size ) · y _sub-min − (x _size / 2)} ² +
{Y _sub-min − (y _size / 2)} ² ] ^1/2
It becomes. _Let L _cross ^(k) be the maximum distance when comparing the distance between the image center point and each intersection.

While the turntable 16 makes one rotation, the above calculation is sequentially performed to calculate the maximum distance L _cross ^(k) for all the frame images. Maximum; (however, k _st rotation starting frame value, k _en represents the frame value at rotation end _{k = k st, ..., k} en) from the after one rotation end, the distance L _cross ^(k) The value L _max is determined.

L _max = max (L _cross ^(kst) ,..., L _cross ^(ken) )
Here, max () is a function that gives the maximum value.

Zoom magnification Z _r seeking is the distance from the upper left end point of the image to the image center is given by the ratio of the maximum value Lmax.

Z _r = {(x _size / 2) ² + (y _size / 2) ² } ^1/2
[First processing flow]
Next, FIG. 7 shows a first processing flow until the zoom magnification is automatically calculated and a rotating object is photographed at an optimum size. It is assumed that the zoom magnification can be automatically set by giving an appropriate signal to the video camera 15, and the description of the control method on the video camera 15 side is omitted. Hereinafter, the flow of processing will be described according to (a) to (f) shown in FIG.

  (A) The zoom magnification of the video camera 15 is set to a wide angle, a member is placed on the turntable 16, and rotation is started.

  (B) The rotating body detection unit 13 detects the rotating body in the frame image, and calculates the position of each side of the sub-partition template from the image area occupied by the rotating body.

(C) While the turntable 16 makes one rotation, the intersection of each side of the sub-partition template and the diagonal line is obtained, and the distance to the image center is calculated. The maximum distance value L _cross ^(k) is stored in a memory or register in the apparatus.

(D) When the turntable 16 completes one rotation, the maximum value L _max is obtained from the distance L _cross ^(k) of all frame images. The zoom magnification _Zr is calculated based on the ratio between the distance from the upper left end point of the image to the center of the image and the maximum value _Lmax .

  (E) A signal based on the obtained zoom magnification is given to the video camera 15 to start photographing for video recording.

  (F) When the turntable 16 reaches an appropriate number of rotations, the photographing is terminated, and a process of displaying the recorded video on the display device 17 or storing it in the recording medium 18 is executed.

[Second processing flow]
In the method according to the first processing flow described above, the member is photographed at an optimum size by setting the wide angle at the first rotation and adjusting the zoom magnification at the second rotation. In the following, the zoom magnification of the video camera 15 of the first rotation and the second rotation is kept constant, and the second rotation is simulated by extracting and enlarging the rotating member occupation area from the acquired frame image. Next, a method for generating a zoom image will be described.

FIG. 8 is a diagram for explaining a method of setting an extraction area. As described in the magnification calculator 14, the maximum value L _max is obtained after the object has made one revolution. When a rectangle as shown in FIG. 8 is drawn using L _max , the object enters the rectangle regardless of the posture. Therefore, in the second rotation, a pseudo zoom image can be acquired by extracting a rectangular area determined by L _max and enlarging the cut image to the original image size. Hereinafter, the flow of processing according to this method will be described with reference to (a) to (h) shown in FIG.

  (A) The zoom magnification of the video camera 15 is set wider, a member is placed on the turntable 16, and rotation is started.

  (B) The rotating body detection unit 13 detects the rotating body in the frame image, and calculates the position of each side of the sub-partition template from the image area occupied by the rotating body.

(C) While the turntable 16 makes one rotation, the intersection of each side of the sub-partition template and the diagonal line is obtained, and the distance to the image center is calculated. The maximum distance value L _cross ^(k) is stored in a memory or register in the apparatus.

(D) When the turntable 16 completes one rotation, the maximum value L _max is obtained from the distance L _cross ^(k) of all frame images.

(E) A rectangle determined by L _max is set.

(F) Shooting for video recording is started, and a rectangular area determined by L _max is extracted from the acquired image.

  (G) The extracted rectangular area is enlarged to obtain the original image size.

  (H) When the turntable 16 reaches an appropriate number of rotations, photographing is terminated, and processing for displaying the recorded video on the display device 17 or storing it in the recording medium 18 is executed.

[Third processing flow]
This section describes a method for obtaining a zoom image by obtaining the minimum zoom magnification after one rotation and enlarging the stored frame image based on the obtained minimum zoom magnification.

First, in the first rotation, the acquired image is stored together with the maximum distance L _cross ^(k) between the image center point and each intersection. After the object has made one revolution, the magnification calculator 14 determines the maximum value L _max . As described above, when a rectangle as illustrated in FIG. 8 using L _max is drawn, the object enters the rectangle regardless of the posture. Therefore, a pseudo zoom image can be acquired by extracting a rectangular region determined by L _max from the stored first rotation image and enlarging the cut image to the original image size. Hereinafter, the flow of processing according to this method will be described with reference to (a) to (h) shown in FIG.

  (A) The zoom magnification of the video camera 15 is set wider, a member is placed on the turntable 16, and rotation is started.

  (B) The rotating body detection unit 13 detects the rotating body in the frame image, and calculates the position of each side of the sub-partition template from the image area occupied by the rotating body.

(C) While the turntable 16 makes one rotation, the intersection of each side of the sub-partition template and the diagonal line is obtained, and the distance to the image center is calculated. The maximum distance L _cross ^(k) and the acquired image are stored in a memory or register in the apparatus.

(D) When the turntable 16 completes one rotation, the maximum value L _max is obtained from the distance L _cross ^(k) of all frame images.

(E) A rectangle determined by L _max is set.

(F) The stored first rotation image is reproduced, and a rectangular area determined by L _max is extracted for each image.

  (G) The extracted rectangular area is enlarged to obtain the original image size.

  (H) When the turntable 16 reaches an appropriate number of rotations, photographing is terminated, and processing for displaying the recorded video on the display device 17 or storing it in the recording medium 18 is executed.

  The above-described member video recording process can be realized by a computer and a software program. The program can be provided by being recorded on a computer-readable recording medium or can be provided through a network.

  The features of the present embodiment are listed as follows.

(Appendix 1)
A rotating object image processing apparatus that displays or records an image of a rotating object,
An image obtained by intermittently or continuously capturing a rotating object by a camera is input, an object area occupied by the rotating object is detected on each image by image processing, and the object rotates once. Rotating body detection means for calculating an area occupied by an object on the image by rotating the object area in the captured image and storing the area information in a storage device;
Based on the area information stored in the storage device, the zoom magnification of the camera that captures the rotating object is determined so that the area occupied by the object during rotation is maximized within a predetermined range on the image. A scaling factor calculating means;
A rotating object image processing apparatus comprising: means for re-photographing the rotating object with the camera set with the determined zoom magnification, and displaying or recording the photographed image.

(Appendix 2)
In the rotating object image processing device according to attachment 1,
The means for displaying or recording the image extracts an area occupied by the object during rotation from an image obtained by re-imaging the rotating object with a camera having the determined zoom magnification, and enlarges the image to a predetermined image size. A rotating object image processing device characterized by displaying or recording an enlarged image.

(Appendix 3)
A rotating object image processing apparatus that displays or records an image of a rotating object,
An image obtained by intermittently or continuously capturing a rotating object by a camera is input, an object area occupied by the rotating object is detected on each image by image processing, and the object rotates once. Rotating body detection means for calculating an area occupied by an object on the image by rotating the object area in the captured image and storing the area information in a storage device;
Based on the area information stored in the storage device, a magnification for enlarging an image taken until the object makes one rotation is set so that an area occupied by the object is within a predetermined range on the image. A magnification calculation means for determining the maximum value at
A rotating object image processing apparatus comprising: means for enlarging an image taken until the object makes one rotation at the determined magnification, and displaying or recording the enlarged image.

(Appendix 4)
A rotating object image processing method for displaying or recording an image of a rotating object, comprising:
An image obtained by intermittently or continuously capturing a rotating object by a camera is input, an object area occupied by the rotating object is detected on each image by image processing, and the object rotates once. Rotating object detection process in which the area occupied by the object on the image is calculated by superimposing the object area in the captured image, and the area information is stored in the storage device
Based on the area information stored in the storage device, the zoom magnification of the camera that captures the rotating object is determined so that the area occupied by the object during rotation is maximized within a predetermined range on the image. Magnification calculation process,
A rotating object image processing method comprising: taking a picture of the rotating object again with a camera set with the determined zoom magnification, and displaying or recording the taken image.

(Appendix 5)
In the rotating object image processing method according to attachment 4,
In the process of displaying or recording the image, an area occupied by the object during rotation is extracted from an image obtained by photographing the rotating object again with the camera having the determined zoom magnification and enlarged to a predetermined image size. And displaying or recording the enlarged image. A rotating object image processing method.

(Appendix 6)
A rotating object image processing method for displaying or recording an image of a rotating object, comprising:
An image obtained by intermittently or continuously capturing a rotating object by a camera is input, an object area occupied by the rotating object is detected on each image by image processing, and the object rotates once. Rotating object detection process in which the area occupied by the object on the image is calculated by superimposing the object area in the captured image, and the area information is stored in the storage device;
Based on the area information stored in the storage device, a magnification for enlarging an image taken until the object makes one rotation is set so that an area occupied by the object is within a predetermined range on the image. A magnification calculation process for determining the maximum at
A rotating object image processing method comprising: enlarging an image captured until the object makes one rotation at the determined magnification, and displaying or recording the enlarged image.

(Appendix 7)
A rotating object image processing program for causing a computer to execute the rotating object image processing method according to appendix 4 to appendix 6.

It is a block diagram which shows the structural example of a rotating object image processing apparatus. It is a figure which shows the relationship between a subdivision, a subdivision template, and a tracking block. It is a figure which shows the rotary body detection procedure by a rotary body detection part. It is a figure which shows the extraction procedure of the candidate of the rotary body image area | region by a rotary body detection part. It is a figure which shows the rotary body determination procedure by the rotary body detection part. It is a figure for demonstrating the calculation method of zoom magnification. It is a 1st process flowchart. It is a figure for demonstrating the setting method of an extraction area | region. It is a 2nd processing flowchart. It is a 3rd process flowchart.

Explanation of symbols

10 Rotating Object Image Processing Device 11, 12 Signal Input / Output Interface (I / F)
DESCRIPTION OF SYMBOLS 13 Rotating body detection part 14 Magnification calculation part 15 Video camera 16 Turntable 17 Display apparatus 18 Recording medium 20 Acquired image 21 Subdivision 22 Subdivision template 23 Tracking block

Claims (5)

A rotating object image processing apparatus that displays or records an image of a rotating object,
An image obtained by intermittently or continuously capturing a rotating object by a camera is input, an object area occupied by the rotating object is detected on each image by image processing, and the object rotates once. Rotating body detection means for calculating an area occupied by an object on the image by rotating the object area in the captured image and storing the area information in a storage device;
Based on the area information stored in the storage device, the zoom magnification of the camera that captures the rotating object is determined so that the area occupied by the object during rotation is maximized within a predetermined range on the image. A scaling factor calculating means;
A rotating object image processing apparatus comprising: means for re-photographing the rotating object with the camera set with the determined zoom magnification, and displaying or recording the photographed image. A rotating object image processing apparatus that displays or records an image of a rotating object,
An image obtained by intermittently or continuously capturing a rotating object by a camera is input, an object area occupied by the rotating object is detected on each image by image processing, and the object rotates once. Rotating body detection means for calculating an area occupied by an object on the image by rotating the object area in the captured image and storing the area information in a storage device;
Based on the area information stored in the storage device, a magnification for enlarging an image taken until the object makes one rotation is set so that an area occupied by the object is within a predetermined range on the image. A magnification calculation means for determining the maximum value at
A rotating object image processing apparatus comprising: means for enlarging an image taken until the object makes one rotation at the determined magnification, and displaying or recording the enlarged image. A rotating object image processing method for displaying or recording an image of a rotating object, comprising:
An image obtained by intermittently or continuously capturing a rotating object by a camera is input, an object area occupied by the rotating object is detected on each image by image processing, and the object rotates once. Rotating object detection process in which the area occupied by the object on the image is calculated by superimposing the object area in the captured image, and the area information is stored in the storage device;
Based on the area information stored in the storage device, the zoom magnification of the camera that captures the rotating object is determined so that the area occupied by the object during rotation is maximized within a predetermined range on the image. Magnification calculation process,
A rotating object image processing method comprising: taking a picture of the rotating object again with a camera set with the determined zoom magnification, and displaying or recording the taken image. A rotating object image processing method for displaying or recording an image of a rotating object, comprising:
An image obtained by intermittently or continuously capturing a rotating object by a camera is input, an object area occupied by the rotating object is detected on each image by image processing, and the object rotates once. Rotating object detection process in which the area occupied by the object on the image is calculated by superimposing the object area in the captured image, and the area information is stored in the storage device;
Based on the area information stored in the storage device, a magnification for enlarging an image taken until the object makes one rotation is set so that an area occupied by the object is within a predetermined range on the image. A magnification calculation process for determining the maximum at
A rotating object image processing method comprising: enlarging an image captured until the object makes one rotation at the determined magnification, and displaying or recording the enlarged image.   A rotating object image processing program for causing a computer to execute the rotating object image processing method according to claim 3.

Images