JP2003070021A - Portable three-dimensional data input apparatus and stereoscopic model generating apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a high-portability three-dimensional data input apparatus, and to eliminate or reduce a calibration process for obtaining a positional relationship between a camera and a turntable in the all-around data input mode. SOLUTION: The data input apparatus has an input unit 10 composed of a digital camera 10b, the turntable 10e and a background board 10d into a single body. For inputting an object (in use), the camera 10b is erected with the board 10d erected at the opposite position and then tightly supported with a post 10c pivoted turnably to a cabinet, as shown by arrow in Fig. (A), as well as the background board 10d. While being not in use, the post 10c and the board 10d are bent down into the surface of the cabinet 10a to make it portable, as shown in Fig. (B).

Description

Detailed Description of the Invention

[0001]

2. Description of the Related Art Conventionally, as a three-dimensional data input device (three-dimensional scanner), one that uses laser light (light cutting method), two that uses pattern light such as slit light (coded light method) , 3: There are devices that use a camera-captured image (stereo method, silhouette method) and the like.

However, these are generally large-scale devices, and it has been customary to bring an object to the installation location of the input device and perform the input processing. Now, consider a case where a single three-dimensional data input device is used for a plurality of customers to provide a three-dimensional data input service. For the customer who makes the input request, bringing the object into the three-dimensional data input service company every time the input request is made causes a great loss in time and money.

Further, in the case of a valuable object, even moving the object to the location of the three-dimensional data input service provider becomes a problem.

Therefore, the 3D data input service provider himself brings in the 3D data input device to the customer side,
It is desired to input three-dimensional data. For this purpose, the three-dimensional data input device should not be a large-scale one, and it is necessary to improve portability.

For the above-mentioned purposes, a portable laser scanner (Minolta Vivid700) and a silhouette method input device (Sanyo Electric's RealModelist, 3Dscanboo 3Dscanboo) are used as a three-dimensional data input device with enhanced portability.
k) etc.

[0006]

In any of the methods, when it is attempted to input all-surroundings data of an object, the object is placed on a computer-controlled rotary table, and this is input while rotating it at every certain angular step. Will do the work.

In the former case, it is necessary to integrate the partial three-dimensional data obtained at each rotation step, which is actually a manual work using an editing tool in a host computer (personal computer or workstation). Become. Therefore, there is a problem that a great deal of time and skill required for editing work is required.

On the other hand, in the latter case, a calibration process for obtaining the positional relationship between the input section (for example, a digital camera) and the turntable is required. Specifically, a panel on which a known pattern is printed is placed on a turntable, a plurality of images are taken while rotating the turntable, and the calibration process is performed by automatic calculation by a host computer. However, since the calibration process is required in addition to the object input, there is a problem that the input time and the processing time increase. In this case, there is also a method in which a known pattern is printed on the turntable itself and the target object and the turntable are photographed together so that the calibration process is apparently eliminated. Although this method does not increase the input time, it causes an increase in processing time, and the number of pixels of the image allocated to the object itself is relatively decreased (because the rotating table must be also photographed). There is a problem that the quality of such three-dimensional data tends to deteriorate.

The present invention has been made in view of the above problems of the prior art, and realizes a three-dimensional data input device having high portability, and at the time of inputting all-around data, a camera and a turntable. The purpose of this is to eliminate or reduce the calibration process for obtaining the positional relationship of.

[0010]

In order to achieve the above object, the present invention is a portable device for inputting three-dimensional data of an object, which includes rotating means for rotating and driving the object. It is characterized in that the photographing means for photographing the object is provided on the same base in a predetermined positional relationship.

In the present input device, on the base, a background plate is further provided on the straight line connecting the photographing means and the rotating means and behind the object as seen from the photographing means. Is preferred.

Further, in the present input device, the photographing means is fixed to a support rotatably provided on the base,
It is preferable that the pillar can be rotated substantially vertically with respect to the surface of the base when in use, and the pillar can be rotated substantially parallel to the surface of the base when not in use.

When a background plate is used, the background plate is rotatably provided on the base, the background plate can be rotated substantially vertically with respect to the base when in use, and the background plate when not in use. It is preferable that the can be rotated substantially parallel to the surface of the base.

In the present input device, it is preferable that the input device further includes a reference marker portion that is within the photographing range of the photographing means and is fixed with respect to the rotating means.

The present invention also provides a three-dimensional model generation device including the above three-dimensional data input device. The apparatus includes a processing unit that processes an image acquired by the image capturing unit of the three-dimensional data input device to generate three-dimensional data, and the processing unit includes the image capturing unit and the rotation unit that are stored in advance. It is characterized in that the three-dimensional data is generated based on the positional relationship data of the means.

When the input device includes a reference marker portion, the processing means detects the reference marker portion image from the image, the position of the reference marker portion image, and a theoretical value corresponding to the positional relationship data. It is preferable to have means for performing parallel movement processing of the image based on the difference between the reference position of the marker and the reference position.

The processing means detects the reference marker portion image from the image, a position of the reference marker portion image, and a difference between a theoretical marker reference position according to the positional relationship data. It is preferable to have means for enlarging or reducing the image based on the above.

Further, the processing means detects a difference between the reference marker portion image from the image, the position of the reference marker portion image, and a theoretical reference position of the marker according to the positional relationship data. And a means for correcting the positional relationship data based on the above.

On the other hand, when the input device does not include the reference marker section, the processing means extracts the silhouette information of the object from the image and a plurality of images extracted from a plurality of images having different photographing angles. It is preferable to have means for performing parallel movement processing of the image based on the position of the silhouette information.

Further, the processing means extracts the silhouette information of the object from the image, and the positional relationship data based on the positions of the plurality of silhouette information extracted from a plurality of images with different photographing angles. It is preferable to have a means for correcting

Further, the processing means extracts the first silhouette information of the object from the image, and the object based on the plurality of first silhouette information extracted from a plurality of images having different photographing angles. Means for generating three-dimensional shape data of an object, means for extracting second silhouette information based on the three-dimensional shape data, and parallel movement of the image based on a difference between the first silhouette information and the second silhouette information And means for processing.

Further, the processing means extracts the first silhouette information of the object from the image, and the object based on the plurality of first silhouette information extracted from a plurality of images having different photographing angles. Means for generating three-dimensional shape data of the object, means for extracting second silhouette information based on the three-dimensional shape data, and the positional relationship data based on the difference between the first silhouette information and the second silhouette information. It is preferable to have a correction means.

[0023]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

<First Embodiment> FIG. 1 shows an overall configuration diagram of an apparatus according to the present embodiment. This apparatus is composed of a rotating table, a background plate, a reference marker, an input unit 10 including an image capturing unit such as a digital camera, a control unit 12, and a data processing device 14. The input unit 10 is integrated by being set in the housing. This will be described later.

The control unit 12 controls the rotation of the turntable and the input control of an image pickup unit such as a digital camera.

The data processing unit 14 processes the input data (for example, an image) and calculates three-dimensional data (shape and color information) of the object. In addition, here, as a calculation method of three-dimensional data, the well-known silhouette method and its extension method (for example, the document “A Portable 3D Di
gitizer ”International Conference on Recent Advan
ces in 3-D Digital Imaging and Modeling, pp.197-20
4 (1997)).

FIG. 1B shows a block diagram of the configuration of the data processing device 14. The data processing device 14
It has an interface I / F, a CPU, a ROM, an image memory, and a positional relationship data memory.

The image data of the object obtained by the input unit 10, that is, a plurality of image data having different photographing angles according to the rotation angle of the rotary table, which will be described later, is the interface I / F.
Is stored in the image memory via.

The CPU reads out the image stored in the image memory and executes a predetermined three-dimensional calculation. The three-dimensional calculation is performed according to the positional relationship between the camera and the object, which is stored in advance in the positional relationship data memory, more specifically, the positional relationship between the camera and the turntable. By previously storing the positional relationship between the camera and the turntable, the calibration process becomes unnecessary. Further, when there is an error in the positional relationship between the camera and the turntable, the CPU corrects the image stored in the image memory so as to eliminate the error according to a program stored in advance in the ROM, or the positional relationship data. The positional relationship data stored in the memory is corrected.

Although the silhouette method is used for the three-dimensional calculation in this embodiment, it is not necessarily limited to the silhouette method, and various three-dimensional shape reconstruction algorithms such as a stereo method may be applied.

The control unit 12 and the data processing device 14 can be specifically configured by a computer.

The input section 10 will be described below.

<Compactness by folding> Input unit 1
The digital camera 0, the digital camera, the rotary base, the background plate, and the reference marker are integrally configured as described above, but the state is as shown in FIG. The housing (base) 10a has a substantially rectangular shape, and the digital camera 10b is erected on one of two facing short sides,
On the other hand, a background plate 10d is erected. Digital camera 10
b is fixed to the column 10c. The column 10c is rotatably supported on the housing 10a as shown by an arrow in the figure, and the background plate 10d is also rotatably provided on the housing 10a.

When not in use, both the column 10c and the background plate 10d can be folded in the plane of the housing 10a as shown in FIG. 7B, and when in use, both the column 10c and the background plate 10d can be folded from the housing 10a. By "raising", the digital camera 10b and the turntable 10e and the background plate 10d have a predetermined positional relationship, and the target object 100 and the background plate 10d placed on the turntable 10e exist within the shooting range of the digital camera 10b. Become.

The background plate 10d can be folded by itself, and when the apparatus is used, it can be expanded to a required area by unfolding the folding portion as shown in FIG. The manner of deployment is arbitrary, and some of them are illustrated in FIG. In FIG. 3 (A), the background plate 10d is composed of parts 10d1, 10d2 and 10d3. In the folded state, the parts 10d2 and 10d3 are folded over the part 10d1.
This is a mode in which d2 and 10d3 are expanded from 10d1. Also in FIG. 3B, the portions 10d1, 10d2, and 10d3 are folded, but the portion 10d2 is folded on the portion 10d1 and the portion 10d3 is further folded on the portion 10d2. When unfolding, first part 10d3
The background plate 10d can be set to a desired area by unfolding and the portion 10d2. Furthermore, FIG.
Is an example in which the background plate 10d is composed of a portion 10d1 and a portion 10d2, is folded in half, and is expanded to set a desired area.

By making the background plate 10d foldable in this manner, it is possible to cover a sufficiently wide background area corresponding to the angle of view of the digital camera 10b.

When in use, the digital camera 10
Since the brace 10c of b is raised from the housing 10a and fixed, the position of the digital camera 10b is fixed with respect to the rotary table 10e fixed to the housing 10a. That is, since the positions of the digital camera 10b and the rotary base 10e are always constant during use, data (calibration data) indicating the positional relationship between the digital camera 10b and the rotary base 10e is registered in the data processing unit 14 in advance. ,
The actual three-dimensional shape calculation process may be performed based on this data. This is an external parameter of the digital camera 10b and is represented by a rotation matrix and a translation vector.

That is,

[Equation 1] Is. However, Pc (x, y, z) is the digital camera 1
0b is the coordinate system, and Pt (x, y, z) is the rotary table coordinate system. Further, T (x, y, z) is a translation vector.

In other words, the calibration process (the process of calculating the positional relationship between the digital camera 10b and the rotary base 10e) may be performed only once after the device is manufactured or according to the original design specifications, and the calibration for each input process may be performed. No processing is required.

The three-dimensional data input process and the three-dimensional data calculation process using the above apparatus are as follows.

(1) Data input (2) Silhouette extraction (3) Voxel generation (4) Polygon generation (5) Texture acquisition These will be sequentially described below.

<Data input> For data input, first, the input unit 10 in a folded state is brought to the input site, each folding unit is unfolded, and each movable unit is fixed, whereby the data is ready for input. To Specifically, the housing 1
0a is arranged at a predetermined position, and the support column 10c is erected from the housing 10a to fix the digital camera 10b. And
The background plate 10d is raised and unfolded. After that, the target object 1
00 is placed on the turntable 10e, and the turntable 10e is rotated at a designated step (for example, every 10 degrees) while
An all-circumference image of 0 is taken by the digital camera 10b.

Here, it is preferable to appropriately select the type of the background plate 10d according to the color of the object. For example, in the case of the target object 100 that does not include blue, the blue background plate 10d can be used as the background plate 10d.

Then, the three-dimensional data calculation process is performed.

<Silhouette Extraction> For the silhouette extraction processing, for example, the well-known chroma key method can be used. This is a technique often used in broadcasting stations and the like, and is a method for synthesizing a plurality of videos to create a new video.

The image obtained by the digital camera 10b is
An image of the background plate 10d and the target object 100 is formed. In the case of the target object 100 that does not include blue, the background plate 10d can be, for example, a blue background plate 10d. In this case, if the blue portion is removed from the image, the target object 100d can be used.
Only the part of can be cut out.

FIG. 4 schematically shows the silhouette extraction processing. In the image A1 obtained by photographing the target object 100 on the turntable 10e, the target object 100 and the blue background are present as shown in (A). By removing only this blue background, it is possible to extract only the image of the target object 100 as shown in FIG.
Blue is black, other than blue is represented by white).

Of course, instead of extracting the silhouette using chroma key processing, for example, Japanese Patent Application No. 9-23261
As disclosed in No. 7, the silhouette may be extracted by the difference processing between the background image and the object image.

<Voxel Generation> Since the field of view of the digital camera 10b is limited in advance, it is possible for the measurer to preset the possible area. This possible area is expressed in a voxel space. Here, the voxel space means that a three-dimensional space is divided into fine three-dimensional elements (voxels) so that a two-dimensional image is represented by a set of surface elements (pixels).
It is expressed as a collection of. Most simply, the three-dimensional shape can be reconstructed as follows.
That is, each voxel forming the voxel space is projected onto one silhouette image. At this time, if the voxel is projected outside the object silhouette, it can be determined that the voxel is clearly not the target object.

By performing such processing on all silhouettes, only voxels projected inside the object silhouette are extracted from all silhouette images, and this is determined as the target object 100.

Intuitively, this process can be considered as a method of assuming the plaster block 200 corresponding to the object existing region and gradually cutting the portion other than the silhouette region. Then, the finally remaining plaster part is regarded as the three-dimensional shape of the object.

If there is an error in the silhouette image, a large shape error will occur in the final result in the above process. In such a case, the voting method (voting method) described below is effective.

That is, instead of immediately deleting the portion outside the silhouette area, the points are first added to the space corresponding to the silhouette area. Then, the part where the score finally exceeds a certain threshold value is regarded as the object part.

The flow of the voting process is as follows. Note that here, as described above, it is premised that the object existing region is expressed as a voxel space.

(1) Initialization: The score of each voxel is set to zero.

(2) For each voxel, the projection points on the entire silhouette image are obtained, and the number of projection points included in the silhouette area is used as the point.

(3) Voxels whose scores are equal to or higher than a certain value are regarded as an object region.

FIG. 6 schematically shows the voting method. Here, it is shown that the darker the part is, the more voxels the score is. In addition, FIG.
(B) and (C) are obtained by adding 1 to the projection points included in the silhouette area in the three directions. (D) is the addition of these projection points, and the total of the projection points of each voxel is shown by a number. The voting result is, for example, 1
It is indicated by a number from 0 to 16. Since the number of voxels in which the target object 100 truly exists is large, by extracting voxels having a certain value or more, for example, 16 or more,
Voxels can be extracted as shown in (E).

An important point in this processing is information about where the viewpoint of the digital camera 10b exists with respect to the voxel space. If this information is incorrect, the calculation result of the three-dimensional shape will be significantly different from the actual result. If the voxel space is defined according to the coordinates of the rotary base 10e, the position information of the viewpoint of the digital camera 10b is nothing but the positional relationship between the rotary base 10e and the digital camera 10b. By the way, the positional relationship between the rotary base 10e and the digital camera 10b is fixed as described above. Therefore, this data is previously stored in the data processing unit 14.
The calibration data registered in can be used.

<Polygon Generation> The three-dimensional shape data of the target object 100 is once obtained as a set of voxel data, but in a general shape measurement tool, it is assumed that the target object is represented by polygon data. There is also. In this case, as shown in FIG. 7, it is necessary to convert the voxel data 300 into polygon data 400.
There are also cases where it is necessary to show the measurement shape three-dimensionally with display software (viewer). Also in this case, it is effective to convert the voxel data into polygon data.

Here, it is required to reduce the number of polygons required for expression and to maintain the accuracy of the expression shape. For example, a suitable polygon expression is possible by taking the steps shown below.

(1) Vertex setting Adjacent vertices of voxels corresponding to the object surface are connected to generate an initial polygon.

(2) Polygon reduction The shape change before and after merging is evaluated, and adjacent polygons with small changes are sequentially merged to reduce the number of polygons to a target value.

<Texture Generation> In this step, color (texture) information of all polygons is acquired based on the omnidirectional photographed image. Specifically, after determining an input image (hereinafter referred to as a reference image) to which texture information of each polygon is given, the polygon is projected on the reference image, and the color of the projected portion is used as the texture information. It should be noted here that, for most polygons, there are multiple input images whose regions are visible. Therefore, it is necessary to determine an appropriate reference image for each polygon. Points to be considered in determining an appropriate reference image are (1) selecting a reference image with a large amount of texture information (2) improving the continuity of the image on the polygon boundary line.

From the viewpoint of (1), it is desirable that the polygon projection area on the reference image is large. On the other hand, (2)
From the viewpoint of, it is desirable that the reference images of adjacent polygons are the same, and even if they are different, it is desirable that the shooting angle difference is small. Since these required points are often in a trade-off relationship, the evaluation function of each consideration point is determined, and the reference image is determined so that the sum function thereof becomes maximum. Incidentally, such a method is disclosed in Japanese Patent Application No. 9-234829.

Through the above processing, three-dimensional data of the target object can be acquired.

<Second Embodiment> In the first embodiment, the processing is performed on the assumption that the positions of the digital camera 10b and the rotary base 10e are known. However, in an actual device, play of the fixed portion of the digital camera 10b (joint portion between the digital camera 10b and the support column 10c),
Also, there are cases where the registered calibration data cannot be used as it is because of play in the joint between the support column 10c and the housing 10a. In other words, in the actual input work,
The positional relationship between the digital camera 10b and the rotary base 10e corresponding to the registered calibration data may be different. If the three-dimensional shape calculation is performed in this state, a large error will be included in the generated three-dimensional data.

However, based on the property that the positional relationship is different, but the difference is extremely small, as shown in FIG. 8 or 9, the reference mark 16 whose position is fixed to the rotary table 10e is introduced. At the same time, the reference mark 16 along with the target object 100 when data is input
The correction process can be performed by capturing an image including the.
In this case, corresponding to the registered calibration data,
The position (two-dimensional) data of the reference mark 16 is also obtained in advance and registered in the data processing unit 14. Further, a window surrounding the position (two-dimensional) of the reference mark 16 is set, and the window position data is also registered in the data processing unit 14.

The reference mark 16 may be one, but a plurality of reference marks 16 are preferable. In FIG. 8, the reference mark 16 is movably attached to the housing 10a similarly to the background plate 10d, is folded when not in use, and is raised and used like the background plate 10d when in use. Further, in FIG. 9, the reference mark 16 is provided on the background plate 10d itself.

It is preferable that the reference mark 16 has a color which can be easily distinguished from the background and is arranged near the edge of the input image. Since the object 100 normally exists near the center of the input image and the reference mark 16 is easily hidden, the arrangement around the center should be avoided. On the other hand, when it exists in the periphery, silhouette parts other than the object will appear in the silhouette extraction, but in general, in the voxel generation described above, even if silhouette noise exists, the distance between the noise and the object is The farther away it is, the less harmful it is. In addition, noise that does not correspond to an object generally corresponds to a low score point in the voting process in the voxel generation process in the three-dimensional space (voxel space). Therefore, there is almost no adverse effect on voxel generation.

The outline of the correction process is shown below.

(1) Object image input (2) As shown in FIG. 10, the reference mark 600 existing in the registered window 500 from each input image.
To extract.

(3) The reference mark position registered in advance (this is used as a standard reference mark) is compared with the actual reference mark position obtained in step (2) (this is used as the actual reference mark). , Find the difference in position.

(4) Correction processing is performed based on the position difference obtained in (3) above.

The following two types of methods can be considered for the correction process in the final step (4).

(A) Two-dimensional correction processing (b) Calibration data correction The two-dimensional correction processing does not correct the calibration data itself, but apparently matches the registered calibration data with the input image. Is a process for correcting. In the image that matches the registered calibration data, the position of the actual reference mark and the position of the standard reference mark match. Therefore, as shown in FIG. 11, the position of the actual reference mark 600 and the reference reference mark 550 are shown.
It is only necessary to calculate the difference between the positions and to perform the parallel movement processing on the image input by this difference. Then, the three-dimensional data calculation process may be performed based on the image subjected to the parallel movement process. In addition,
The difference between the position of the actual reference mark 600 and the position of the standard reference mark 550 does not necessarily reflect the translation error of the position of the digital camera 10b.
In many cases, the rotational movement error of the digital camera 10b is reflected. However, since the difference is almost negligible on the image, an error caused by either of them can be treated as a translation error.

On the other hand, the calibration data correction is
The position of the actual reference mark 600 and the reference reference mark 5 described above.
The difference in the position of 50 is considered to be caused by the rotational movement error of the digital camera 10b, and the calibration data itself is changed.

Specifically, as shown in FIG. 12, the position of the actual reference mark 600 on the CCD when the focal length f of the digital camera 10b is normalized to 1.0 is X.
s, and the position of the standard reference mark 550 is Xr,

2−θ−θ ′ = arctan (Xr) −arcta
n (Xs) is an approximate value of the rotation angle error. After correcting the calibration data by this angle, the three-dimensional data calculation process may be performed.

Further, by disposing a plurality of reference marks 16 as shown in FIG. 8 or FIG. 9, it is possible to correct the error caused by the forward and backward movement of the digital camera 10b. If the digital camera 10b is not moved back and forth, the distance between the plurality of reference marks is constant. However, when the distance between the plurality of reference marks becomes large, it can be determined that this is because the position of the digital camera 10b is close to the rotary base 10e side. Similarly to the above, this correction is also performed by (a) correcting the input image so as to match the registered calibration data by image reduction processing, and (b) obtaining the moving distance of the digital camera 10b and correcting the calibration data itself. It can be corrected by two methods.

FIG. 13 schematically shows the method (b). When the position (focus position) of the digital camera 10b changes from P to P ', the distance between the reference mark 16 and the digital camera 10b is L', the distance on the image between the reference marks 16 is W ', If the registered reference distances are L and W respectively,

## EQU3 ## L '= W.L / W' The moving distance can be calculated with.

The digital camera 10b and the turntable 10
The reference mark 16 is used so that the positional relationship of e changes slightly.
The positional relationship between the rotary table 10e and the rotary table 10e also changes. However,
Regarding the error that affects the three-dimensional calculation, generally, the amount of parallel movement is extremely small, and in reality, the rotation error of the digital camera 10b is largely caused. Therefore, the reference mark 1
It can be considered that the slight deviation of the position of 6 itself has almost no effect.

The position of the actual reference mark in the correction process using the reference mark 16 may be detected for all input images or for a specific image. In particular, if it is possible to assume that the position of the camera does not change during the input work, it is preferable to actually detect the position of the reference mark only for a specific image from the viewpoint of shortening the calculation time.

<Third Embodiment> In the above-described embodiment,
The method of correcting the calibration data by using the reference mark 16 has been described. However, the calibration data correction can be performed without using the reference mark 16.

When the target object 100 is placed on the rotary table 10e and the data of the entire circumference is photographed at regular steps,
A specific point on the object moves circularly in three dimensions. Therefore, an elliptical orbit is drawn on the input image. Furthermore, after extracting the silhouette of the object, it can be considered that the movement of the center of gravity on the image also draws an elliptical orbit as shown in FIG. Then, the center of the ellipse (this is called Ce)
Stands on the straight line obtained by projecting the rotation axis of the turntable 10e onto the digital camera image. The center of this ellipse can be obtained, for example, as the average position of all the centers of gravity of the silhouette of the target object 100.

Therefore, after the silhouette extraction processing, Ce is calculated as described above and compared with the projected straight line (Lc) of the rotation axis calculated from the registered calibration data. Ce
Is present on Lc, calibration data correction is unnecessary. On the other hand, as shown in FIG.
If is out of Lc, the two-dimensional correction or the calibration data correction as already described in the second embodiment may be performed correspondingly.

Note that this processing only corrects the rotation error of the digital camera 10b in the direction perpendicular to the rotation axis of the turntable 10e (usually in the horizontal direction in the image). The rotation error of the digital camera 10b may include an error in the direction along the rotation axis. However, in the three-dimensional data calculation, the influence of the error in the direction perpendicular to the rotation axis is large, while the influence of the error in the direction along the rotation axis is extremely small. Therefore, practically, it is often sufficient to correct the error in the direction perpendicular to the rotation axis.

<Fourth Embodiment> In the third embodiment, the method of correcting the calibration data by using the information of the silhouette image has been described. However, when the extracted silhouette is noisy, such as when an unnecessary object is reflected in the background, the method of the third embodiment may not always work. In this case, calibration correction using voxel generation processing is possible.

Now, consider a case where the voxel generating process is performed based on the registered calibration data including an error in the direction perpendicular to the rotation axis of the rotary table 10e. In this case, due to the influence of the error, the generated three-dimensional data has a shape thinner than the actual object.

In FIG. 16, two viewpoints a and b opposite to each other by 180 degrees are shown when the error is not included in the direction perpendicular to the rotation axis of the rotary base 10e (A) and when the error is included (B). The silhouette image obtained from is shown. When the error is included in the direction perpendicular to the rotation axis, the object 100 moves by the amount of the error at the viewpoint b, and the silhouette image also shifts. Therefore, when the voxels of the target object 100 are extracted from the silhouette images obtained at the viewpoints a and b, the shape 800 actually obtained rather than the true shape 700 is proportional to the error of the rotation axis as shown in FIG. Becomes smaller. Therefore, the error is corrected by performing the following processing.

(1) Voxel generation is performed based on the registered calibration data.

(2) A silhouette image (referred to as a re-silhouette image or a second silhouette image) is generated by projecting the generated voxel data on a digital camera image.

(3) As shown in FIG. 18, the silhouette image 800 obtained by the silhouette extraction processing is compared with the re-silhouette image (second silhouette image) 900 obtained in (2).

(4) In step (3) above, the difference between the silhouette image 800 and the re-silhouette image 900 (the re-silhouette image does not become larger but is the same or smaller) is obtained.

The difference obtained in step (4) corresponds to the error of the calibration data. As described in the second embodiment, the two-dimensional correction or the calibration data correction may be performed for this error.

In the correction processing described above, it is not necessary to use all input images, and processing can be performed based on several sheets (for example, 6 sheets).

Although the embodiment of the present invention has been described above, the present invention is not limited to this, and other modifications are possible. For example, the background board 10d can be replaced with a plurality of colors. Generally, with the simple silhouette method, it is difficult to input an object having the same color as the background color. However, as shown in Japanese Patent Application No. 11-322098, by inputting using a plurality of background plates having different hues, it is possible to input an object of any color.
Further, by adding white as a background color, (1) silhouette extraction for creating shape data is performed based on an image using a colored (for example, blue) background plate (2) texture acquisition processing Can be selectively used based on an image using a white background plate.

When an image is input using a colored background plate, the background color may be reflected in the peripheral portion of the object depending on the material of the object. When texture acquisition processing is performed based on such an image, the quality of texture information of the finally obtained three-dimensional data may be insufficient. On the other hand, if texture acquisition is performed based on a white background image,
It is possible to minimize the adverse effect that the background color is reflected in the peripheral portion of the object as described above.

A lighting device may be added to the input section 10 shown in FIG.

Further, the digital camera 10 shown in FIG.
It is also possible to provide a plurality of latch positions in the rotation range of the column 10c that supports b so that the digital camera 10b can be fixed at a plurality of positions. The same applies to the background plate 10d.

[0100]

As described above, according to the present invention, a three-dimensional data input device having high portability can be realized, and the positional relationship between the camera and the turntable can be obtained at the time of inputting all-round data. It is possible to reduce or eliminate the calibration process.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an apparatus according to an embodiment.

FIG. 2 is a configuration diagram of an input unit in FIG.

3 is a development explanatory diagram of a background portion in FIG. 1. FIG.

FIG. 4 is an explanatory diagram of silhouette extraction.

FIG. 5 is an explanatory diagram of voxel generation.

FIG. 6 is an explanatory diagram of a voting method.

FIG. 7 is an explanatory diagram of conversion from voxel data to polygon data.

FIG. 8 is an explanatory diagram of a reference mark.

FIG. 9 is an explanatory diagram of another reference mark.

FIG. 10 is an explanatory diagram of an input image when a reference mark is used.

FIG. 11 is an explanatory diagram of a positional relationship between a standard mark and a reference mark.

FIG. 12 is an explanatory diagram (No. 1) of calibration data correction based on a rotation error of the digital camera.

FIG. 13 is an explanatory diagram of correction of calibration data based on a front-back movement error of the digital camera.

FIG. 14 is an explanatory diagram (part 2) of calibration data correction based on a rotation error of the digital camera.

FIG. 15 is an explanatory diagram showing the relationship between the center of the ellipse in FIG. 14 and the rotation axis of the rotary table.

FIG. 16 is a calibration data correction explanatory diagram (part 3) based on the rotation error of the digital camera using the voxel generation processing.

FIG. 17 is a diagram for explaining the difference between the true shape and the actual shape in FIG.

FIG. 18 is a diagram illustrating a difference between a silhouette image and a re-silhouette image.

[Explanation of symbols]

10 input unit, 12 control unit, 14 data processing unit.

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) H04N 1/393 H04N 1/393 5C076 // H04N 5/225 5/225 Z F term (reference) 5B047 AA07 BA03 BB04 BC14 BC23 CB22 DC09 5B050 BA04 BA07 BA09 EA02 EA05 EA12 EA13 EA19 EA28 5B057 CA08 CA13 CA16 CB06 CB08 CB13 CB16 CD02 CD05 CF05 DA07 DB03 DC05 5C022 AA13 A27 A17 A17 A17 AB17 A17 AB17 A17 AB17 A17 AB17 A17 AB17 AB17 AB17 AB17 AB17 AB17 AB23 AB17 AB03

Claims (14)

[Claims] 1. A device for inputting three-dimensional data of an object, wherein a rotating means for rotationally driving the object and a photographing means for photographing the rotating object have the same positional relationship with each other. A portable three-dimensional data input device, which is provided on a table. 2. The apparatus according to claim 1, further comprising: on the base, a straight line connecting the photographing means and the rotating means,
A portable three-dimensional data input device, characterized in that a background plate is provided behind the object viewed from the photographing means. 3. The apparatus according to claim 1, wherein the photographing means is fixed to a support rotatably provided on the base, and the support supports the support at the surface of the base. The portable three-dimensional data input device is characterized in that it can be rotated substantially vertically, and the column can be rotated substantially parallel to the surface of the base when not in use. 4. The apparatus according to claim 2, wherein the background plate is rotatably provided on the base, and when used, the background plate is substantially vertical to the base. A portable three-dimensional data input device, wherein the background plate can be rotated, and when not in use, the background plate can be rotated substantially parallel to the surface of the base. 5. The apparatus according to claim 1, further comprising: a reference marker unit that is within an image capturing range of the image capturing unit and is fixed with respect to the rotating unit. Characteristic portable 3D data input device. 6. A three-dimensional model generation device comprising the three-dimensional data input device according to claim 1, wherein an image acquired by a photographing means of the three-dimensional data input device is processed to obtain a cubic image. And a processing unit for generating original data, wherein the processing unit generates the three-dimensional data based on pre-stored positional relationship data of the photographing unit and the rotating unit. apparatus. 7. A three-dimensional model input device comprising the three-dimensional data input device according to claim 5, wherein the image acquired by the photographing means of the three-dimensional data input device is processed to generate three-dimensional data. Means for storing the positional relationship data of the photographing means and the rotating means. 8. The apparatus according to claim 7, wherein the processing unit is a unit that detects the reference marker portion image from the image, a position of the reference marker portion image, and a theoretical value corresponding to the positional relationship data. A means for performing parallel movement processing of the image based on a difference between the reference position of the marker and the reference position, and a stereo model generation device. 9. The apparatus according to claim 7, wherein the processing means detects the reference marker part image from the image, the position of the reference marker part image, and a theoretical value corresponding to the positional relationship data. A means for enlarging or reducing the image based on the difference between the marker and the reference position, and a stereo model generation device. 10. The apparatus according to claim 7, wherein the processing means detects the reference marker portion image from the image, a position of the reference marker portion image, and a theoretical value corresponding to the positional relationship data. A means for correcting the positional relationship data based on a difference between the marker and the reference position, and a stereo model generation device. 11. The apparatus according to claim 6, wherein the processing means includes means for extracting silhouette information of the object from the image, and a plurality of silhouette information pieces extracted from a plurality of images having different photographing angles. Means for performing parallel movement processing of the image based on a position, and a three-dimensional model generation device. 12. The apparatus according to claim 6, wherein the processing unit includes means for extracting silhouette information of the object from the image, and a plurality of pieces of silhouette information extracted from a plurality of images having different shooting angles. A means for correcting the positional relationship data based on the position, and a stereo model generation device. 13. The apparatus according to claim 6, wherein the processing unit extracts the first silhouette information of the object from the image, and the plurality of the plurality of images extracted from a plurality of images with different photographing angles. 1 means for generating three-dimensional shape data of the object based on the silhouette information, a means for extracting second silhouette information based on the three-dimensional shape data, a difference between the first silhouette information and the second silhouette information Means for performing parallel translation processing of the image based on the above. 14. The apparatus according to claim 6, wherein the processing unit extracts the first silhouette information of the object from the image, and the plurality of first images extracted from a plurality of images with different photographing angles. 1 means for generating three-dimensional shape data of the object based on the silhouette information, a means for extracting second silhouette information based on the three-dimensional shape data, a difference between the first silhouette information and the second silhouette information And a means for correcting the positional relationship data based on the above.