JP4904638B2 - Method and apparatus for generating three-dimensional shape data and computer program - Google Patents

Description

【００９１】
として求められる。
すなわち、格子点ＴＰｘと最近点ｒi との間の距離は、格子点ＴＰｘと最近点ｒi とを結ぶ線の長さ（つまりそれらの間の距離）に、その線と最近点ｒi における法線とのなす角の余弦をかけた値として求められる。つまり、格子点ＴＰｘまでの最近点ＭＤにおける法線方向の距離として求められる。距離の符号は、法線方向がマイナスである。求めた距離について、各最近点ｒi の信頼性に応じた重み（荷重）を与えて平均する。これによって符号付き距離が求まる。
【００９２】
重みｗi は、最近点ｒi における法線ベクトルｎi と、格子点ＴＰｘを始点とし最近点ｒi を終点とするベクトルとがなす角の余弦である。つまり、その角が大きいほど、信頼性は小さくなると考えられるので、重みｗi を小さくする。
【００９３】
求めた符号付き距離を、その格子点ＴＰｘの属性値として格納する。
すべての格子点ＴＰに対して処理を行ったかどうかがチェックされる（＃９９）。まだ処理を行っていない格子点ＴＰがある場合は、ステップ＃９２以降を繰り返す。すべての格子点ＴＰについての処理を終えた場合には終了する。これによって、レンジデータＤＲに対応した、図１１のボリュームテーブルＴＬｓと同様なボリュームテーブルＴＬｒが完成する。
〔ボリュームデータの統合〕
さて、ステップ＃２および３において、シルエット画像ＦＳを用いたボリュームデータＤＶと、レンジデータＤＲを用いたボリュームデータＤＶとが生成された。ステップ＃４では、これらのボリュームデータＤＶを統合する。ボリュームデータＤＶの統合に際しては、各ボリュームデータＤＶの互いに対応するボクセルＶＸの２つの属性値に基づいて、当該ボクセルＶＸの統合された属性値を求める。なお、シルエット画像ＦＳによるボリュームデータＤＶの属性値を、シルエット属性値ｄｓ、レンジデータＤＲによるボリュームデータＤＶの属性値を、レンジ属性値ｄｒ、統合されたボリュームデータＤＶの属性値を統合属性値ｄｔと記載することがある。
【００９４】
図２６は格子点ＴＰの属性値と物体の表面ＨＭとの位置関係を示す図、図２７は２つの属性値の統合方法を示す図である。
図２６に示すように、物体の表面ＨＭに対し、属性値の値に応じて、外部（遠）、外部（近）、内部（近）、内部（遠）の４つに分類する。
【００９５】
すなわち、属性値が「マイナス無限大」の場合に外部（遠）、属性値が「プラス無限大」の場合に内部（遠）、属性値が負であって「マイナス無限大」でない場合に外部（近）、属性値が正であって「プラス無限大」でない場合に内部（近）となる。これは、シルエット画像ＦＳおよびレンジデータＤＲのいずれによるボリュームデータＤＶに対しても適用される。
【００９６】
図２７に示すように、シルエット属性値ｄｓが外部（遠）である場合に、レンジ属性値ｄｒの内容に係わらず、統合属性値ｄｔは外部（遠）を示す「マイナス無限大」である。シルエット属性値ｄｓおよびレンジ属性値ｄｒが共に内部（遠）である場合に、統合属性値ｄｔは内部（遠）を示す「プラス無限大」である。シルエット属性値ｄｓが内部（遠）で且つレンジ属性値ｄｒが外部（遠）である場合には、これは実際にはあり得ないが、統合属性値ｄｔは外部（遠）を示す「マイナス無限大」である。
【００９７】
それ以外で、一方が外部（遠）または内部（遠）で他方が外部（近）または内部（近）であった場合には、外部（近）または内部（近）であった方の属性値が用いられる。両方が外部（近）または内部（近）であった場合には、両方の属性値が混合される。混合によって、統合属性値ｄｔは次の（３）式により計算される。
【００９８】
ｄｔ＝ｗｘｄｒ＋（１−ｗｘ）ｄｓ ……（３）
但し、ｗｘは格子点ＴＰｘにおけるレンジ属性値ｄｒの重みを表す。つまり、レンジ属性値ｄｒの重みｗｘに応じた比で、レンジ属性値ｄｒとシルエット属性値ｄｓとが混合される。荷重ｗｘの値は、物体の形状などに応じて決定される。このように、重みに応じた適当な比で加算されることにより、シルエット画像ＦＳによるボリュームデータＤＶとレンジデータＤＲによるボリュームデータＤＶとの境界部分が滑らかに接続される。得られた統合属性値ｄｔは、その格子点ＴＰの属性値として格納される。このようにして、すべての格子点ＴＰについて属性値が求められる。これによって統合処理が完了する。
【００９９】
このように、シルエット画像ＦＳによるボリュームデータＤＶとレンジデータＤＲによるボリュームデータＤＶとを統合することにより、それぞれの欠点を補って欠落のない精度のよい３次元モデルＭＬを生成することができる。
【０１００】
したがって、物体に凹部領域や低反射率の部分があった場合でも、できるだけ少ないメモリ容量で且つ少ない演算時間で、正確に３次元形状を復元することができる。
【０１０１】
上に述べた実施形態においては、属性値をボクセルの頂点にセットしたが、属性値をセットするのは必ずしもボクセルの頂点である必要はない。例えば、ボクセルの重心であってもよい。
【０１０２】
また、境界と交差するボクセルについてのみ、頂点と境界との距離値をセットするようにしているが、全ボクセルについて行ってもよい。但し、実施例のように交差するボクセルについてのみ演算する方が、演算時間が短縮される。物体の内部を正、外部を負として表現したが、逆であってもよい。
【０１０３】
第４の実施形態では、キューブを８分割する例について説明したが、８分割に限らず、ｘｙｚそれぞれの方向に３分割し、つまりキューブを２７分割してもよいし、それ以上に分割してもよい。
【０１０４】
上に述べた実施形態において、３次元データ生成装置１の全体または各部の構成、処理の内容および順序、ボクセルＶＸの個数、属性値の桁数などは、本発明の趣旨に沿って適宜変更することができる。
【０１０５】
【発明の効果】
本発明によると、凹部領域や低反射率の部分があった場合でも、できるだけ少ないメモリ容量で且つ少ない演算時間で正確に３次元形状を復元することができる。
【図面の簡単な説明】
【図１】本実施形態に係る３次元データ生成装置のブロック図である。
【図２】３次元データ生成装置による３次元モデルの生成処理の流れを示すフローチャートである。
【図３】３次元データ生成装置による変換処理を示すフローチャートである。
【図４】変換処理の変形例を示すフローチャートである。
【図５】第２の実施形態の変換処理を示すフローチャートである。
【図６】図５のステップ＃３４の属性値の設定処理のサブルーチンを示すフローチャートである。
【図７】シルエット画像をボリュームデータに投影した状態を説明するための図である。
【図８】図７に示すシルエット画像の一部を拡大して示す図である。
【図９】ボリュームデータを水平面で切断した状態を示す図である。
【図１０】境界ボクセルを拡大して示す図である。
【図１１】ボリュームデータを格納したボリュームテーブルの例を示す図である。
【図１２】第３の実施形態の変換処理を示すフローチャートである。
【図１３】図１２のステップ＃５４の属性値の設定処理のサブルーチンを示すフローチャートである。
【図１４】第４の実施形態の変換処理を示すフローチャートである。
【図１５】図１４のステップ＃７５の交差判定処理のサブルーチンを示すフローチャートである。
【図１６】８分木表現の原理を説明するための図である。
【図１７】８分木表現の原理を説明するための図である。
【図１８】第５の実施形態の変換処理を示すフローチャートである。
【図１９】スペースカービングを説明するための図である。
【図２０】ボリュームデータを水平面で切断した状態を示す図である。
【図２１】境界ボクセルを拡大して示す図である。
【図２２】最近点を求める方法を説明するための図である。
【図２３】最近点を求める方法を説明するための図である。
【図２４】最近点を求める方法を説明するための図である。
【図２５】最近点を求める方法を説明するための図である。
【図２６】格子点の属性値と物体の表面との位置関係を示す図である。
【図２７】２つの属性値の統合方法を示す図である。
【符号の説明】
１ ３次元データ生成装置
１０ 装置本体（第１の変換手段、第２の変換手段、統合手段）
１１ 磁気ディスク装置
ＤＶ ボリュームデータ
ＶＸ ボクセル
ＦＴ 画像
ＦＳ シルエット画像
ＣＤ ＣＤ−ＲＯＭ（記録媒体）
ＦＤ フロッピィディスク（記録媒体）
ＴＬ ボリュームテーブル（ボリュームデータ）
ＰＲ モデリングプログラム（コンピュータプログラム）[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method and apparatus for generating three-dimensional shape data, and a computer program.
[0002]
[Prior art]
Conventionally, a silhouette method (Shape From Silhouette) is known as one of methods for generating three-dimensional shape data of an object.
[0003]
In the silhouette method, the three-dimensional shape of an object is restored from a plurality of images obtained by photographing (or imaging) the object from different viewpoints. That is, in the silhouette method, based on images obtained by photographing an object from a plurality of line-of-sight methods, an object existence region that reflects the shape of the object viewed from each line-of-sight direction is described as a set of voxels in the three-dimensional image space. Then, a set of voxels in the object existence area viewed from all directions is obtained as an object in the three-dimensional space.
[0004]
[Problems to be solved by the invention]
However, in the silhouette method, a depressed portion or a depressed portion of the object does not appear as a silhouette image, and thus the shape of such a recessed region cannot be restored.
[0005]
On the other hand, as another method for restoring the three-dimensional shape of the object, there is a method using a range sensor (three-dimensional measurement apparatus) that measures the object in a non-contact manner by a light cutting method or the like. However, when the range sensor is used, there is a problem that data is lost when the reflectance of the surface of the object is extremely low, and the three-dimensional shape of the portion cannot be restored.
[0006]
The present invention has been made in view of the above-described problems, and it is an object of the present invention to accurately restore a three-dimensional shape with as little memory capacity as possible and with a small calculation time even when there is a recessed region or a low reflectance portion. And
[0007]
  According to one aspect of the present invention, there is provided a method for generating three-dimensional shape data for an object, wherein a first method is performed by converting a plurality of images of the object with different viewpoints into volume data using a silhouette method. A first step of generating volume data of the second, a second step of generating second volume data by converting the three-dimensional data obtained by three-dimensional measurement of the object into volume data, the first volume A third step of integrating the data and the second volume data into one volume data;The volume data is composed of a plurality of voxels having lattice points at discrete coordinate positions in the coordinate system representing the volume, and values corresponding to the distance from the surface of the object at the lattice points of the voxels. Is stored as an attribute value, and for the attribute value, the outside (far) indicating that the grid point is outside the object, and the surface of the object is more external than the outside (far) when the grid point is outside the object. Outside (near) indicating that the grid point is inside the object (far), and that the grid point is inside the object and closer to the surface of the object than the inside (far) In the third step, for the lattice point of each voxel, the attribute value by the first volume data and the second volume data are used for the lattice points of the voxels. Obtaining an integrated attribute value is an integrated attribute value based on the sexual value.
[0008]
Preferably, in the first step, when converting to volume data using the silhouette method,
(A) obtaining a value corresponding to the distance to the boundary of the visual volume determined by the contour of the image of the object and the viewpoint of the image at each discrete coordinate position in the coordinate system representing the volume;
(B) For each coordinate position, a value corresponding to the distance from the object surface is determined based on a plurality of values obtained for a plurality of images of the object, and a step of holding the value in association with each coordinate position is executed.
[0009]
In the second step, when converting the three-dimensional data into volume data,
(A) obtaining a value corresponding to the distance to the object surface represented by the three-dimensional data of the object at each discrete coordinate position in the coordinate system representing the volume;
(B) A step of holding the obtained value in association with each coordinate position is executed.
[0010]
By inversely transforming the generated volume data, the three-dimensional shape of the object is restored.
According to the present invention, an accurate three-dimensional model without missing is generated by making up for each defect due to a difference in the method of generating three-dimensional shape data.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a block diagram of a three-dimensional data generation apparatus 1 according to this embodiment.
In FIG. 1, the three-dimensional data generation device 1 includes a device main body 10, a magnetic disk device 11, a medium drive device 12, a display device 13, a keyboard 14, a mouse 15, and the like.
[0012]
The apparatus main body 10 includes a CPU, a RAM, a ROM, a video RAM, an input / output port, various controllers, and the like. Various functions described below are realized by the CPU executing programs stored in the RAM and ROM.
[0013]
The magnetic disk device 11 includes an OS (Operating System), a modeling program PR for generating a three-dimensional model ML, other programs, input three-dimensional data (three-dimensional shape data) DT, and images (two-dimensional image data). ) FT, volume data DV, generated three-dimensional model ML, and other data are stored. These programs and data are loaded into the RAM of the apparatus main body 10 at appropriate times.
[0014]
The modeling program PR includes programs for initialization processing, attribute value setting processing, attribute value determination processing, boundary determination processing, division processing, integration processing, inverse transformation processing, mapping processing, and other processing. .
[0015]
The medium drive device 12 accesses a semiconductor memory HM such as a CD-ROM (CD), a floppy disk FD, a magneto-optical disk, a compact flash, and other recording media, and reads / writes data or programs. An appropriate drive device is used according to the type of the recording medium. The modeling program PR described above can also be installed from these recording media. The three-dimensional data DT and the image FT can also be input via a recording medium.
[0016]
On the display surface HG of the display device 13, the various data described above, the three-dimensional data DT, the image FT, the image of the processing process by the modeling program PR, the generated three-dimensional model ML, and other data or images are displayed. Is done.
[0017]
The keyboard 14 and the mouse 15 are used for the user to make various designations for the image FT and the three-dimensional data DT displayed on the display device 13, and input various data or commands to the apparatus main body 10. Used to give
[0018]
Although not shown, the apparatus main body 10 can be connected to a digital camera for photographing an object as a subject with various lines of sight or photographing from various viewpoints and inputting an image FT. The silhouette image FS is generated by cutting out the outline of the object that is the subject from the image FT. Based on the silhouette image FS, the three-dimensional data DT can be generated by the silhouette method.
[0019]
It is also possible to generate 3D data DT by performing 3D reconstruction based on two images with parallax.
The apparatus main body 10 can be connected to a three-dimensional input device (three-dimensional measurement device) for photographing an object and inputting the three-dimensional data DT. Such a three-dimensional input device measures the three-dimensional data DT of an object in a non-contact manner by, for example, a light cutting method. Further, instead of the three-dimensional data DT, data that is the basis for generating the three-dimensional data DT may be output from the three-dimensional input device, and the three-dimensional data DT may be obtained by calculation by the apparatus body 10.
[0020]
The three-dimensional data DT obtained in this way, or data in the process of generating the three-dimensional data DT, represents the three-dimensional shape of the object in the boundary representation format.
In the present embodiment, the three-dimensional shape of the object expressed in the boundary expression format is converted into volume data (Volume Data) DV in which each voxel has a multivalued attribute value d. A voxel is a small cube composed of unit lattices by dividing a three-dimensional space into small unit lattices.
[0021]
Further, a plurality of volume data DV converted from a plurality of three-dimensional shapes having different generation methods are also integrated into one volume data DV.
The final volume data DV is again converted back to a shape representation in the boundary representation format. At that time, for example, a known zero isosurface extraction method or the like can be used. That is, a point corresponding to a zero isosurface on an edge connecting adjacent vertices (lattice points) is calculated, and a triangular polygon obtained by connecting these points can be converted into a polygon mesh.
[0022]
A texture image is also pasted on the three-dimensional shape obtained by the inverse transformation. In this way, the three-dimensional model ML is generated.
The three-dimensional data generation apparatus 1 can be configured using a personal computer or a workstation. The programs and data described above can be received and acquired via the network NW.
[0023]
FIG. 2 is a flowchart showing a flow of processing for generating the three-dimensional model ML by the three-dimensional data generating apparatus 1.
In FIG. 2, an appropriately sized volume (volume data DV) is prepared and initialized (# 1). As the volume, for example, a volume of about 50 × 50 × 50 voxels or 100 × 100 × 100 voxels is used. The volume does not have to be a cube. Also, the coordinates and posture of the center of the voxel in the coordinate system are set.
[0024]
The volume can store two types of multivalued data as attribute values d at each vertex of the voxel. That is, an area for storing two attribute values ds and dr is provided for each vertex. This may be that there are two volumes capable of storing multivalued data as attribute values at each vertex of the voxel. In this case, one can be described as a first volume and the other as a second volume.
[0025]
In this specification, “volume” may be referred to as “volume data”. In this case, “volume data” means a three-dimensional area that is configured to have a predetermined shape by voxels and that can store data at each vertex of the voxel.
[0026]
Also, the voxel vertices are common to neighboring voxels. Therefore, the vertices that are not in the peripheral edge are the vertices of eight voxels. The vertex coincides with the grid point of the volume data. Therefore, “vertex” may be described as “grid point”.
[0027]
Now, returning to FIG. 2, the conversion process is performed using the silhouette image, and the first volume data DV is generated (# 2). Conversion processing is performed using the range data, and second volume data DV is generated (# 3). Note that the range data is three-dimensional shape data obtained by measuring an object using a three-dimensional measuring device called a range finder or the like. The three-dimensional shape data reconstructed from a plurality of images with parallax by the stereo method is also included in the range data referred to here.
[0028]
These first and second volume data DV are integrated (# 4). The integrated volume data DV is inversely transformed into a shape representation in the boundary representation format (# 5). Texture mapping is performed as necessary (# 6).
[0029]
Hereinafter, the conversion processing to volume data DV and the integration processing of a plurality of volume data DV will be described in detail. First, the conversion process will be described.
[Conversion processing according to the first embodiment]
FIG. 3 is a flowchart showing conversion processing by the three-dimensional data generation apparatus 1, and FIG. 4 is a flowchart showing a modification of the conversion processing. The conversion processing shown in these flowcharts is applied to conversion processing using range data, other three-dimensional shape data, and a silhouette image.
[0030]
In FIG. 3, first, three-dimensional data DT representing the three-dimensional shape of an object is prepared. Alternatively, instead of the three-dimensional data DT, a plurality of images FT that is a source for generating the three-dimensional data DT is prepared. Then, the alignment of the three-dimensional data DT and the volume data DV is performed (# 11). Normally, the size and position of the three-dimensional data DT are determined so that the three-dimensional data DT is exactly contained in the volume data DV.
[0031]
For each vertex of the voxel constituting the volume, a value corresponding to the distance from each vertex to the boundary expressing the three-dimensional shape, that is, the surface of the object is obtained (# 12). The obtained value is held as the attribute value of each vertex (# 13).
[0032]
In this way, multi-value attribute values are stored at the vertices of each voxel, that is, lattice points, and volume data DV composed of lattice point groups having multi-value attribute values is generated.
[0033]
In FIG. 4, voxels that intersect the surface of the object are extracted (# 22). Such a voxel may be referred to as a “boundary voxel”. A value corresponding to the distance from the vertex to the surface of the object is obtained for the boundary voxel (# 23). The obtained value is held as the attribute value of each vertex in the same manner as above (# 24).
[0034]
When a certain vertex is a common vertex between the boundary voxel and other voxels, the vertex is treated as a vertex of the boundary voxel.
[Conversion processing according to the second embodiment]
Based on the second embodiment, the conversion processing will be described in more detail. Note that the second and third embodiments are mainly applied to conversion processing using a silhouette image.
[0035]
FIG. 5 is a flowchart showing the conversion processing of the second embodiment, FIG. 6 is a flowchart showing the attribute value setting processing subroutine of step # 34 in FIG. 5, and FIG. 7 is a state in which the silhouette image FS is projected onto the volume data DV. FIG. 8 is an enlarged view showing a part of the silhouette image FS shown in FIG. 7, FIG. 9 is a view showing a state in which the volume data DV is cut along a horizontal plane, and FIG. 10 is an enlarged view of boundary voxels. FIG. 11 is a diagram showing an example of a volume table TLs storing volume data DV.
[0036]
In FIG. 5, first, initial values are set at all grid points TP for the set volume data DV. As a result, the volume data DV is initialized (# 31). As the initial value of the lattice point TP, for example, “plus infinity” meaning the inside of the object is set. In this case, “plus infinity” is the first specific value in the present invention. When the initialization is completed, all of the volume data DV shown in FIG. 10 have initial values.
[0037]
Next, one image FT is input (# 32). At that time, the camera parameters when the image FT is taken are also input. Camera parameters include camera internal matrix such as focal length and external matrix such as viewpoint position. A projection matrix including these may be input. Based on the camera parameters, the viewpoint position and the projection direction when the image FT is projected onto the volume data DV are determined.
[0038]
When the image FT is input in step # 32, the image FT is stored in the magnetic disk device 11. The image FT stored in the magnetic disk device 11 is automatically read into the RAM by a program, and the subsequent processing is added. However, the image input in step # 32 may be used to mean that the image FT stored in the magnetic disk device 11 is read into the RAM. In that case, a large number of images FT may be stored in advance in the magnetic disk device 11, and one image FT designated in step # 32 may be read onto the RAM. Further, the input of an image can also be used to designate one image FT to be processed from among a large number of images FT stored in the magnetic disk device 11.
[0039]
The silhouette image FS is generated by cutting out the contour of the object that is the subject from the input image FT (# 33). The silhouette image FS only needs to be known in outline, so a monochrome image is sufficient. The silhouette image can be generated automatically or manually by a known method.
[0040]
The volume data DV, that is, the attribute value of each vertex (grid point) TP of the voxel VX is obtained and set (# 34). The attribute value is obtained as a signed distance between each vertex TP and the surface of the object. That is, for example, as shown in FIGS. 7 and 8, the distance Ls from the boundary SF of the visual volume VV determined by the contour (shielding contour) by the silhouette image FS and the viewpoint VP to the vertex TP is set to the inner side of the visual volume VV. The direction toward is taken as + (plus). Details will be described later.
[0041]
When an image FT that needs to be processed remains, the next one image is input (Yes in # 35, # 32). This is repeated until the processing for all necessary images FT is completed (# 35). Even when the processing has been completed for all scheduled images FT, it is possible to perform steps # 32 to # 34 by adding an image FT as necessary.
[0042]
Instead of inputting the image FT at step # 32, it is also possible to input a silhouette image FS. In that case, step # 33 is unnecessary.
In FIG. 6, when setting the attribute value, first, attention is paid to one lattice point TP constituting the volume (# 41). The attribute value of the focused lattice point TP is checked (# 42). If the attribute value is “minus infinity”, the process proceeds to step # 45. That is, that the attribute value is “minus infinity” means that the lattice point TP is outside the object. Since the grid point TP outside the object is cut off, there is no need to obtain an attribute value beyond that. In this case, “minus infinity” is the second specific value in the present invention.
[0043]
At the time of processing the first image, since “plus infinity” is set as the initial value for all grid points, the process proceeds to step # 43. At the time of processing the second and subsequent images, either “plus infinity”, “minus infinity”, or “signed distance ds” described later is set as an attribute value by the processing of the previous image. If it is other than “minus infinity”, ie, “plus infinity” or “signed distance ds”, the process proceeds to step # 43.
[0044]
If the attribute value is not “minus infinity”, the signed distance ds is calculated in step # 43. There are various methods for calculating the signed distance. Next, three methods from the first to the third will be described.
(First method)
The focused lattice point TP and the lattice point adjacent to the lattice point TP (that is, the vertex for the same voxel VX) are each projected onto the image FT toward the viewpoint VP. When projected points are connected to each other on the image FT, a signed distance is calculated in a case where a side formed thereby intersects with a contour (shielding contour).
[0045]
That is, as shown in FIG. 7, when a certain voxel VX is on the boundary SF of the visual volume VV, the voxel VX is a boundary voxel VXs. A signed distance is calculated for the vertex TP of the boundary voxel VXs.
[0046]
In FIG. 9, voxels VX that intersect the boundary SF of the visual volume VV are shown as intermediate voxels as boundary voxels VXs. In FIG. 10, the distance from each vertex TP of the boundary voxel VXs to the boundary SF of the visual volume VV is obtained. The obtained distance or a numerical value corresponding to the distance is shown as an attribute value of each vertex TP.
[0047]
The unit or scale of the distance may be set appropriately. The maximum value of the distance is the length of the diagonal line of the voxel VX. Therefore, the distance may be normalized based on the length of the diagonal line. For the vertex TP inside the visual volume VV, the sign of the distance is set to a positive value as it is, and for the vertex TP located outside, a negative value is added to the distance.
[0048]
For example, when 8-bit data is used as the attribute value, the most significant bit is a sign bit and the lower 7 bits indicate a distance. As a result, values of −127 to +128 can be expressed, so that −127 corresponds to “minus infinity” indicating the outside, +128 corresponds to “plus infinity” indicating the interior, and −126 ˜ + 127 corresponds to the signed distance. As the attribute value, data of 12 bits, 16 bits, or other number of bits can be used.
[0049]
Therefore, when the lattice point TP is on the boundary SF of the visual volume VV, the attribute value is zero. The attribute value increases as the lattice point TP moves into the visual volume VV, and the attribute value decreases as it moves outward.
[0050]
If the side does not intersect the occlusion outline, that is, if it is not the boundary voxel VXs, the lattice point TP exists inside or outside the view volume VV. If it is outside, the grid point TP is a point to be cut out, so the attribute value is set to “minus infinity”. If it is inside, the attribute value remains “plus infinity”.
(Second method)
The target lattice point TP is projected on the image FT toward the viewpoint VP. It is checked whether or not there is a shielding outline in a region inside a circle having a certain radius from the center with the projected point as the center. Calculate signed distance if occluded contour exists. When there is no shielding contour in the region and the lattice point TP is outside the visual volume VV, the attribute value is set to “minus infinity”.
.
(Third method)
The shielding contour is sampled at regular intervals, and a straight line connecting the sampling point and the viewpoint VP is obtained. These straight lines are lines of sight passing through the shielding contour. Using a line of sight instead of the boundary SF of the visual volume VV, a signed distance is calculated three-dimensionally. That is, for the vertex TP of the boundary voxel VXs, the distance from the vertex TP to the line of sight is calculated. When the focused lattice point TP is outside the visual volume VV, the attribute value is set to “minus infinity”.
[0051]
In step # 44, an attribute value is set at the grid point TP. When setting an attribute value, only a value smaller than an already set attribute value is set. If a new value larger than the already set attribute value is obtained, the new attribute value is ignored.
[0052]
In other words, if “minus infinity” is newly obtained as the attribute value, the attribute value becomes “minus infinity” whatever the previously set attribute value. In this way, the outside of the object is cut off.
[0053]
In addition, when the attribute value is a signed distance for both old and new, an average value thereof may be obtained and used as a new attribute value d.
It is checked whether or not processing has been performed for all grid points TP (# 45). If there is a grid point TP that has not yet been processed, step # 41 and subsequent steps are repeated. When the processing for all the lattice points TP is completed, the process returns. As a result, the volume table TLs as shown in FIG. 11 is completed.
[0054]
According to the second embodiment, since processing is performed for each image FT, the image can be deleted when the processing of the image is completed, and the memory capacity to be used can be reduced accordingly.
[Conversion processing according to the third embodiment]
In the conversion process of the second embodiment, the conversion process has progressed by adding the image FT. However, it is also possible to input (read) a plurality of images FT at a time and perform the process. . Next, an example is shown as a third embodiment.
[0055]
FIG. 12 is a flowchart showing the conversion process of the third embodiment, and FIG. 13 is a flowchart showing a subroutine of the attribute value setting process in step # 54 of FIG.
[0056]
In FIG. 12, step # 51 is the same as step # 31 of FIG. In step # 52, all the images FT taken of the object are input at a time. At that time, the camera parameters when each image FT is photographed are input. For each image FT, a silhouette image FS is generated as in step # 33 (# 53). Then, an attribute value is set (# 54).
[0057]
In FIG. 13, when setting an attribute value, first, attention is paid to one grid point TP as in step # 41 (# 61). It is checked at what position the focused lattice point TP is present with respect to the visual volume VV (# 62).
[0058]
That is, when the lattice point TP exists near the boundary SF of the visual volume VV, the attribute value is “BORDER”, and when the lattice point TP exists inside the visual volume VV, the attribute value is “INSIDE”. If it exists outside, the attribute value is “OUTSIDE”. The attribute values “INSIDE” and “OUTSIDE” are the first specific value and the second specific value in the present invention. Next, two checking methods will be described.
(First method)
The target grid point TP and the adjacent grid point are projected on all the images FT. In each projected image FT, when the two projected points are connected on each image FT, it is determined whether or not a side formed thereby intersects the contour (shielding contour). If there is at least one intersecting image FT, it is determined that the lattice point TP exists in the vicinity of the boundary SF of the visual volume VV, and the attribute value is assumed to be “BORDER”. If there is no intersecting image FT and there is at least one image where the projection point of the grid point TP exists outside the shielding contour, it is determined that the grid point TP exists outside the subject, and the attribute value is temporarily set. “OUTSIDE”. In other cases, it is determined that the grid point TP exists inside the subject, and the attribute value is temporarily set to “INSIDE”.
(Second method)
The focused lattice point TP is projected on all the images FT. It is checked whether or not there is a shielding outline in a region inside a circle having a certain radius from the center with each projected point as the center. If there is even one image that is determined to include the occlusion contour, it is determined that the image exists near the boundary SF of the visual volume VV. If there is no image determined to include the occlusion outline and there is one image in which the projection point of the lattice point TP exists outside the occlusion outline, it is determined that the image exists outside the subject. In other cases, it is determined that the object exists inside the subject.
[0059]
In this way, a temporary attribute value is determined for one lattice point TP in consideration of all images FT.
When the temporary attribute value is “OUTSIDE” or “INSIDE”, the attribute value is, for example, “minus infinity” or “plus infinity”, respectively. If the provisional attribute value is “BORDER”, the signed distance is calculated (# 63).
[0060]
The calculation method of the signed distance is basically the same as described in the explanation of Step # 43. However, here, for one grid point TP, the signed distance is determined in consideration of all the images FT at the same time.
[0061]
That is, for example, the signed distance is calculated as follows.
(First method)
The target lattice point TP is projected on all the images FT. From all the projected images FT, the one having the shortest distance from the projection point to the shielding contour is selected. For the selected image FT, a line of sight passing through a point on the shielding contour is obtained. The distance from the grid point TP to the line of sight is obtained. An attribute value is added to the obtained distance with a positive or negative sign.
(Second method)
For all images FT, the occlusion contour is sampled at regular intervals, and the signed distance is calculated three-dimensionally using the line of sight that passes through the occlusion contour, as in the third method in step # 34.
[0062]
Next, in step # 64, an attribute value is set at the grid point TP as in step # 44. However, here, the final attribute value is set for each lattice point TP.
[0063]
In step # 65, as in step # 45, it is checked whether or not processing has been performed for all grid points TP.
According to the third embodiment, since the signed distance is calculated only for the lattice point TP whose provisional attribute value is “BORDER”, the calculation amount of the signed distance is greatly reduced. Therefore, the processing speed is fast.
[Conversion processing according to the fourth embodiment]
Next, a conversion process using octree representation will be described as a fourth embodiment.
[0064]
FIG. 14 is a flowchart showing the conversion process of the fourth embodiment, FIG. 15 is a flowchart showing a subroutine of the intersection determination process in step # 75 of FIG. 14, and FIGS. 16 and 17 are diagrams for explaining the principle of the octree representation. FIG.
[0065]
As shown in FIGS. 16 and 17, in the octree representation, a cube larger than the target object is defined, and this is defined as a root cube (Root-Cube) RC. When the root cube is divided into two equal parts along the x, y, and z directions, eight cubes having a volume of 1/8 are generated. By repeating such division recursively to an arbitrary level, data of an octree is generated. The octree representation itself is known.
[0066]
In FIG. 14, first, a root cube is set (# 71). When setting the root cube, enter the coordinates and size of its center. Also, “CO” meaning the inside of the object is set as an initial value for all the vertices of the root cube. Further, the attribute of the root cube is set to “GRAY” indicating that it intersects the occlusion outline, and the level is set to “0”. A cube whose attribute is “GRAY” corresponds to a boundary cube in the present invention.
[0067]
Next, all necessary images are input for the image FT obtained by photographing the object that is the subject (# 72). At that time, the camera parameters when each image FT is photographed are input.
[0068]
In step # 73, a silhouette image FS is generated as in step # 33.
In step # 74, the cube (root cube) is divided. The division here is performed only for cubes having the attribute “GRAY”. The division is 8 divisions. Then raise the level by one.
[0069]
In step # 75, a cube intersection determination is performed. Here, each divided cube is projected onto the image FT, and the presence or absence of an intersection with the shielding contour is determined. Based on the determination result, the attributes of the cube are determined.
[0070]
Then, steps # 74 and 75 are repeated until a predetermined level is reached (# 76). Or, when there are no more cubes with the attribute “GRAY”, the process ends there. In that case, in the subsequent processing, a cube on the boundary or a cube whose vertex is on the boundary is used as a cube whose attribute is “GRAY”.
[0071]
When the processing is completed, the attribute value of each vertex is obtained and set for the cube having the attribute “GRAY” (# 77). As a method for obtaining the attribute value for each vertex, the method described in the second embodiment can be used.
[0072]
In FIG. 15, attention is paid to one of the divided cubes (# 81). The cube is projected on all the images FT, and it is determined whether or not there is an intersection with the occlusion contour in each of them (# 82).
[0073]
As a result of the determination, if there is even one image FT that intersects the occlusion outline, the attribute of the cube is set to “GRAY”. In all the images, when the projected cube exists within the occlusion outline, the cube attribute is set to “WHITE” representing the inside of the object. In all the images, when the projected cube exists outside the occlusion outline, it is set to “BLACK” representing the outside of the object (# 83).
[0074]
“GRAY”, “WHITE”, and “BLACK” here correspond to “BORDER”, “INSIDE”, and “OUTSIDE” described above, respectively.
If processing has been performed for all of the eight divided cubes (Yes in # 84), the process returns.
[0075]
Also in the fourth embodiment, since the signed distance is calculated only for the lattice point TP of the cube whose attribute is “GRAY”, the calculation amount of the signed distance is greatly reduced, and the processing speed is high.
[0076]
As described above, according to the conversion processes of the first to fourth embodiments, multivalued data can be given to the lattice points TP, and a highly accurate three-dimensional shape can be expressed by a small number of voxels VX. it can. Therefore, the three-dimensional shape of the object can be expressed with high accuracy with a small memory capacity.
[0077]
The conversion processes of the second to fourth embodiments described above are mainly applied to silhouette images. Next, the conversion process mainly applied to the range data will be described in detail.
[Conversion processing according to the fifth embodiment]
18 is a flowchart showing the conversion process of the fifth embodiment, FIG. 19 is a diagram for explaining space carving, FIG. 20 is a diagram showing a state in which the volume data DV is cut along a horizontal plane, and FIG. 21 is an enlarged view of boundary voxels. FIGS. 22 to 25 are diagrams for explaining a method of obtaining the nearest point MD.
[0078]
Here, conversion processing is performed on a plurality of range data DR. Such a plurality of range data can be obtained, for example, by measuring an object in multiple times from different positions around the object. It is assumed that the range data has been aligned with the volume data DV.
[0079]
In FIG. 18, space carving is performed (# 91). Space carving is a process of cutting out unnecessary parts that are not objects from the volume data DV.
[0080]
That is, as shown in FIG. 19, among the voxels VX of the volume data DV, “minus infinity” is set as the attribute value of the vertex TP for the voxel VX outside the range data DR. At that time, a line of sight including the range data DR is hypothesized from the viewpoint VP to the range data DR, and the voxel VX inside the line of sight and outside the range data DR is the voxel VX to be cut off. To do. However, as in the case of the above embodiment, the vertex TP in the vicinity of the range data DR is excluded. In FIG. 19, white voxels are displayed as external voxels VX.
[0081]
Attention is paid to one lattice point TPx (# 92). The attribute value of the focused lattice point TPx is checked (# 93). If the attribute value is not “minus infinity”, the process proceeds to step # 94.
[0082]
For the lattice point TPx, the nearest point MD is obtained (# 94). The closest point MD in the i-th range data DR is set as the closest point ri. The method for obtaining the nearest point MD is as follows.
[0083]
As shown in FIG. 22, a perpendicular is drawn from the lattice point TPx for each of the plurality of range data DR. When a plurality of perpendicular lines can be drawn for one range data DR, the shortest perpendicular line is selected. The droop point is the latest points MD1 and MD2. However, it is a condition that the lengths LM1 and LM2 of the perpendicular are shorter than the predetermined length α. That is, when the perpendicular lengths LM1 and LM2 are both greater than the predetermined length α, the nearest point MD does not exist.
[0084]
If the nearest point MD does not exist (No in # 95), the attribute value of the lattice point TPx is set to “plus infinity” (# 96). If one or more nearest points MD exist (Yes in # 95), the nearest point MD closest to the grid point TPx is set as the nearest point rmin (# 97).
[0085]
Here, the nearest point MD is selected from the three-dimensional polygon data DP constituting the range data DR.
That is, in FIGS. 23 and 25, the nearest point MD1 on the range data DR with respect to the lattice point TPx coincides with the three-dimensional point existing as the polygon data DP of the range data DR. Therefore, the coordinates of the nearest point MD1 coincide with the coordinates of the polygon data DP. Since the coordinates of the polygon data DP are known, the coordinates of the nearest point MD1 can be obtained very easily. However, the nearest point MD1 is not truly the closest point to the range data DR. Therefore, when calculating the signed distance later, the correction is performed by applying the cosine of the angle formed with the normal.
[0086]
On the other hand, as shown in FIG. 24, when an arbitrary point PMT on the polygon mesh PM is set as the nearest point MD, a true nearest point with respect to the range data DR can be obtained. However, in this case, it is necessary to obtain the coordinates of the nearest point MD1 by calculation from the coordinates of the surrounding polygon data DP.
[0087]
Returning to FIG. 18, in step # 98, a signed distance from the grid point TPx to the range data DR is obtained by a weighted average of the distances from the grid point TPx to one or more nearest points MD. The method for obtaining the signed distance is as follows.
[0088]
That is, only the nearest point MD within a certain distance from the nearest point rmin is extracted from the nearest points MD. The nearest point rmin is also included in this. That is, the nearest point MD far from the nearest point rmin by a certain distance is excluded. For the extracted nearest point MD, the distance from the grid point TPx is obtained. It is obtained by weighted average of all obtained distances.
[0089]
That is, the signed distance of the grid point TPx is dr (x)
dr (x) = Σwi · [ni · (ri−TPx)] / Σwi (1)
Where ri is the closest point in the i-th range data
ni is the normal vector of the nearest point ri
wi is the weight of the nearest point ri
[0090]
[Expression 1]

[0091]
As required.
That is, the distance between the lattice point TPx and the nearest point ri is the length of the line connecting the lattice point TPx and the nearest point ri (that is, the distance between them), and the normal line at that point and the nearest point ri. Is obtained by multiplying the cosine of the angle formed by That is, the distance in the normal direction at the nearest point MD to the lattice point TPx is obtained. The sign of the distance is negative in the normal direction. The obtained distance is averaged by giving a weight (load) according to the reliability of each nearest point ri. Thereby, the signed distance is obtained.
[0092]
The weight w i is the cosine of the angle formed by the normal vector ni at the nearest point ri and the vector starting from the lattice point TPx and ending at the nearest point ri. That is, since the reliability is considered to decrease as the angle increases, the weight wi is decreased.
[0093]
The obtained signed distance is stored as an attribute value of the lattice point TPx.
It is checked whether processing has been performed for all grid points TP (# 99). If there is a grid point TP that has not been processed yet, step # 92 and subsequent steps are repeated. When the processes for all the lattice points TP are completed, the process ends. As a result, a volume table TLr corresponding to the range data DR and similar to the volume table TLs of FIG. 11 is completed.
[Integration of volume data]
In steps # 2 and # 3, volume data DV using the silhouette image FS and volume data DV using the range data DR are generated. In step # 4, these volume data DV are integrated. When integrating the volume data DV, the integrated attribute value of the voxel VX is obtained based on the two attribute values of the voxel VX corresponding to each other of the volume data DV. Note that the attribute value of the volume data DV by the silhouette image FS is the silhouette attribute value ds, the attribute value of the volume data DV by the range data DR is the range attribute value dr, and the attribute value of the integrated volume data DV is the integrated attribute value dt. May be described.
[0094]
FIG. 26 is a diagram showing a positional relationship between the attribute value of the lattice point TP and the surface HM of the object, and FIG.
As shown in FIG. 26, the surface HM of the object is classified into four types: outside (far), outside (near), inside (near), and inside (far) according to the value of the attribute value.
[0095]
That is, external (far) when the attribute value is “minus infinity”, internal (far) when the attribute value is “plus infinity”, and external when the attribute value is negative and not “minus infinity” (Near) If the attribute value is positive and not “plus infinity”, it is internal (near). This is applied to the volume data DV by either the silhouette image FS or the range data DR.
[0096]
As shown in FIG. 27, when the silhouette attribute value ds is outside (far), the integrated attribute value dt is “minus infinity” indicating outside (far) regardless of the contents of the range attribute value dr. When the silhouette attribute value ds and the range attribute value dr are both inside (far), the integrated attribute value dt is “plus infinity” indicating the inside (far). If the silhouette attribute value ds is inside (far) and the range attribute value dr is outside (far), this is not possible, but the integrated attribute value dt is “minus infinity” indicating the outside (far). "Large".
[0097]
Otherwise, if one is outside (far) or inside (far) and the other is outside (near) or inside (near), the attribute value that was outside (near) or inside (near) Is used. If both are external (near) or internal (near), both attribute values are mixed. By mixing, the integrated attribute value dt is calculated by the following equation (3).
[0098]
dt = wxdr + (1-wx) ds (3)
However, wx represents the weight of the range attribute value dr at the lattice point TPx. That is, the range attribute value dr and the silhouette attribute value ds are mixed at a ratio corresponding to the weight wx of the range attribute value dr. The value of the load wx is determined according to the shape of the object. In this way, by adding at an appropriate ratio according to the weight, the boundary portion between the volume data DV by the silhouette image FS and the volume data DV by the range data DR is smoothly connected. The obtained integrated attribute value dt is stored as the attribute value of the lattice point TP. In this way, attribute values are obtained for all grid points TP. This completes the integration process.
[0099]
In this way, by integrating the volume data DV based on the silhouette image FS and the volume data DV based on the range data DR, it is possible to generate a highly accurate three-dimensional model ML that compensates for the respective defects.
[0100]
Therefore, even when the object has a concave region or a low reflectance portion, the three-dimensional shape can be accurately restored with as little memory capacity as possible and with a short calculation time.
[0101]
In the embodiment described above, the attribute value is set at the vertex of the voxel. However, it is not always necessary to set the attribute value at the vertex of the voxel. For example, it may be the center of gravity of the voxel.
[0102]
In addition, the distance value between the vertex and the boundary is set only for the voxel that intersects the boundary, but may be performed for all the voxels. However, the calculation time is shortened by calculating only the intersecting voxels as in the embodiment. Although the inside of the object is expressed as positive and the outside is expressed as negative, it may be reversed.
[0103]
In the fourth embodiment, an example in which a cube is divided into eight parts has been described. However, the present invention is not limited to eight parts, but is divided into three parts in each of xyz directions, that is, the cube may be divided into 27 parts or more. Also good.
[0104]
In the embodiment described above, the configuration of the whole or each part of the three-dimensional data generation device 1, the contents and order of processing, the number of voxels VX, the number of digits of attribute values, and the like are changed as appropriate in accordance with the spirit of the present invention. be able to.
[0105]
【The invention's effect】
According to the present invention, a three-dimensional shape can be accurately restored with as little memory capacity as possible and with a short calculation time even when there is a recessed area or a portion with low reflectance.
[Brief description of the drawings]
FIG. 1 is a block diagram of a three-dimensional data generation apparatus according to an embodiment.
FIG. 2 is a flowchart showing a flow of processing for generating a three-dimensional model by the three-dimensional data generating apparatus.
FIG. 3 is a flowchart showing conversion processing by the three-dimensional data generation apparatus.
FIG. 4 is a flowchart showing a modification of the conversion process.
FIG. 5 is a flowchart illustrating a conversion process according to the second embodiment.
FIG. 6 is a flowchart showing a subroutine of attribute value setting processing in step # 34 of FIG. 5;
FIG. 7 is a diagram for explaining a state in which a silhouette image is projected onto volume data.
8 is an enlarged view showing a part of the silhouette image shown in FIG.
FIG. 9 is a diagram illustrating a state in which volume data is cut along a horizontal plane.
FIG. 10 is an enlarged view showing boundary voxels.
FIG. 11 is a diagram showing an example of a volume table storing volume data.
FIG. 12 is a flowchart illustrating conversion processing according to the third embodiment.
FIG. 13 is a flowchart showing a subroutine of attribute value setting processing in step # 54 of FIG. 12;
FIG. 14 is a flowchart illustrating a conversion process according to the fourth embodiment.
FIG. 15 is a flowchart showing a subroutine of intersection determination processing in step # 75 of FIG.
FIG. 16 is a diagram for explaining the principle of octree representation;
FIG. 17 is a diagram for explaining the principle of octree representation;
FIG. 18 is a flowchart illustrating conversion processing according to the fifth embodiment.
FIG. 19 is a diagram for explaining space carving.
FIG. 20 is a diagram illustrating a state in which volume data is cut along a horizontal plane.
FIG. 21 is an enlarged view showing boundary voxels.
FIG. 22 is a diagram for explaining a method for obtaining a closest point;
FIG. 23 is a diagram for explaining a method for obtaining a closest point;
FIG. 24 is a diagram for explaining a method for obtaining a closest point;
FIG. 25 is a diagram for explaining a method for obtaining a closest point;
FIG. 26 is a diagram illustrating a positional relationship between an attribute value of a lattice point and the surface of an object.
FIG. 27 is a diagram illustrating a method for integrating two attribute values.
[Explanation of symbols]
1 3D data generator
10 Device body (first conversion means, second conversion means, integration means)
11 Magnetic disk unit
DV volume data
VX Voxel
FT image
FS silhouette image
CD CD-ROM (recording medium)
FD floppy disk (recording medium)
TL volume table (volume data)
PR modeling program (computer program)

Claims (8)

A method of generating three-dimensional shape data about an object,
A first step of generating first volume data by converting a plurality of images of the object from different viewpoints into volume data using a silhouette method;
A second step of generating second volume data by converting three-dimensional data obtained by three-dimensional measurement of the object into volume data;
A third step of integrating the first volume data and the second volume data into one volume data;
The volume data is composed of a plurality of voxels having lattice points at discrete coordinate positions in a coordinate system representing the volume, and a value corresponding to the distance from the surface of the object is obtained at the lattice point of each voxel. As an attribute value,
For the attribute value, the outside (far) indicating that the grid point is outside the object, and the outside (near) indicating that the grid point is outside the object and closer to the surface of the object than the outside (far). Inside (far) indicating that the grid point is inside the object, and outside (near) indicating that the grid point is inside the object and closer to the surface of the object than the inside (far). Classified into
In the third step, an integrated attribute value that is an attribute value integrated based on the attribute value based on the first volume data and the attribute value based on the second volume data on the lattice point of each voxel. Seeking
A method for generating three-dimensional shape data. When the attribute value by the first volume data is the external (far), the integrated attribute value is external (far) regardless of the attribute value by the second volume data,
When the attribute value based on the first volume data and the attribute value based on the second volume data are both internal (far), the integrated attribute value is internal (far),
When the attribute value based on the first volume data is the inside (far) and the attribute value based on the second volume data is the outside (far), the integrated attribute value is set to the outside (far). ,
Otherwise, if the attribute value according to the first volume data is different from the attribute value according to the second volume data, the integrated attribute value is external (near) or internal (near),
When both the attribute value based on the first volume data and the attribute value based on the second volume data are the same on the outside (near) or inside (near), both the attribute values are mixed. To obtain the integrated attribute value,
The method for generating three-dimensional shape data according to claim 1. When both the attribute value based on the first volume data and the attribute value based on the second volume data are the same outside (near) or inside (near), the integrated attribute value dt is formula,
dt = wx · dr + (1−wx) ds
However, wx: Weight of attribute value by second volume data at grid point
ds: attribute value based on the first volume data
dr: attribute value based on second volume data
Calculate with
The method for generating three-dimensional shape data according to claim 2. In the first step, when converting to volume data using the silhouette method,
(A) obtaining a value corresponding to the distance to the boundary of the visual volume determined by the contour of the image of the object and the viewpoint of the image at each discrete coordinate position in the coordinate system representing the volume;
(B) For each coordinate position, determining a value according to the distance from the object surface based on a plurality of values obtained for a plurality of images of the object, and holding the value in association with each coordinate position;
The method for generating three-dimensional shape data according to claim 1, wherein: In the second step, when converting the three-dimensional data into volume data,
(A) obtaining a value corresponding to the distance to the object surface represented by the three-dimensional data of the object at each discrete coordinate position in the coordinate system representing the volume;
(B) a step of holding the obtained value in association with each coordinate position;
The method for generating three-dimensional shape data according to claim 1, wherein: An apparatus for generating three-dimensional shape data about an object,
First conversion means for generating first volume data by converting a plurality of images with different viewpoints of the object into volume data using a silhouette method;
Second conversion means for generating second volume data by converting three-dimensional data obtained by three-dimensional measurement of the object into volume data; and
Integration means for integrating the first volume data and the second volume data into one volume data ;
The volume data is composed of a plurality of voxels having lattice points at discrete coordinate positions in a coordinate system representing the volume, and a value corresponding to the distance from the surface of the object is obtained at the lattice point of each voxel. As an attribute value,
In the integration means,
For the attribute value, the outside (far) indicating that the grid point is outside the object, and the outside (near) indicating that the grid point is outside the object and closer to the surface of the object than the outside (far). Inside (far) indicating that the grid point is inside the object, and outside (near) indicating that the grid point is inside the object and closer to the surface of the object than the inside (far). Classified into
For the lattice points of each voxel, an integrated attribute value that is an attribute value integrated based on the attribute value by the first volume data and the attribute value by the second volume data is obtained.
An apparatus for generating three-dimensional shape data. A computer program for generating three-dimensional shape data about an object,
A first step of generating first volume data by converting a plurality of images of the object from different viewpoints into volume data using a silhouette method;
A second step of generating second volume data by converting three-dimensional data obtained by three-dimensional measurement of the object into volume data;
A third step of integrating the first volume data and the second volume data into one volume data;
The above processing is executed by a computer, and the computer further includes:
The volume data is composed of a plurality of voxels having lattice points at discrete coordinate positions in a coordinate system representing the volume, and a value corresponding to the distance from the surface of the object is obtained at the lattice point of each voxel. Execute the process so that it is retained as an attribute value,
For the attribute value, the outside (far) indicating that the grid point is outside the object, and the outside (near) indicating that the grid point is outside the object and closer to the surface of the object than the outside (far). Inside (far) indicating that the grid point is inside the object, and outside (near) indicating that the grid point is inside the object and closer to the surface of the object than the inside (far). Process to classify
In the third step, an integrated attribute value that is an attribute value integrated based on the attribute value based on the first volume data and the attribute value based on the second volume data on the lattice point of each voxel. To execute the process for
Computer program for.   A computer-readable recording medium on which the computer program according to claim 7 is recorded.

Images