JPH10320553A - Shape recognition device, shape recognition method, and recording medium recording computer program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a shape recognition device capable of being miniaturized by eliminating especially the need of constitution for rotating an object to be a three-dimensional shape recognition object and obtaining the three- dimensional shape data on the object more easily than before. SOLUTION: In this shape recognition device, a CPU 10 measures the distances of respective distance measuring points set on the image pickup surface of an image pickup device 21 and an object, and obtains the respective ones as distance measurement data. Also, the CPU 10 obtains the two-dimensional image data of the object by making the image pickup device 21 pick up the image of the object. Then, the CPU 10 calculates the two-dimensional contour line vector data and two-dimensional boundary vector data of the object based on the two-dimensional image data on the object. Then, the CPU 10 converts the vector data into three-dimensional vector data based on the respective distance measurement data. Thus, the three-dimensional, modeling data on the object are prepared.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a shape recognition device for recognizing a three-dimensional shape of an object.

[0002]

2. Description of the Related Art Conventionally, as one method for recognizing a three-dimensional shape (three-dimensional shape) of an object, a light cutting method has been used. FIG. 13 is a diagram illustrating an example of a three-dimensional shape recognition device using a light-section method. In FIG. 13, a shape recognition device includes an object A that is a shape recognition target placed on a turntable 33, a light projection device that emits a slit-shaped laser beam S on the surface of the object A, and an imaging surface. An imaging device (imaging CCD) 34 (only an imaging surface is shown in FIG. 13) and an unillustrated processor device (computer) connected to the imaging device 34 are disposed.

[0005] When performing shape recognition of the object A, the shape recognition device performs the following series of operations. That is, when the slit-shaped laser light S is emitted from the light projecting device 32, a high-luminance line L is generated on the surface of the object along the surface shape. Here, in the example shown in FIG. 13, since the object A is a sphere, an arc-shaped high luminance line L is generated. When the high-luminance line L is imaged by the imaging device 34, the processor calculates the spatial coordinate data of each point on the high-luminance line L as surface shape data of the object A and stores the data in a storage device (not shown). .

The above-described series of operations are performed a plurality of times by rotating the turntable 33 to change the position of the high-intensity line L on the surface of the object A, and the surface shape data of each object A in each case is obtained. It is stored in a storage device (not shown).
Then, a processor device (not shown) combines the respective surface shape data with each other while referring to the position of the high-luminance line L with respect to the object A, and performs a process of interpolating between the respective surface shape data, thereby obtaining three-dimensional shape data of the object A. Is created. Thus, the three-dimensional shape of the object A is recognized.

[0005]

However, the conventional shape recognition apparatus has the following problems. That is, in the shape recognition device using the light cutting method, the position of the high luminance line L generated on the surface of the object A needs to be changed as described above. A configuration for rotating A, or a configuration for rotating the light projecting device 32 and the imaging device 34 around the object A in a circular locus is required.

However, when the turntable 33 is provided in the shape recognition device, the shape of the turntable 33 and the performance of the motor for rotating the turntable 33 limit the object A that can be placed on the turntable 33. In some cases, the object A that can use the shape recognition device is limited. On the other hand, when the shape recognizing device is provided with a configuration in which the light projecting device 32 and the image capturing device 34 circulate, the size of the shape recognizing device increases and the diameter of the circle around which the light projecting device 32 and the image capturing device 34 circulate. In some cases, the larger object A cannot be used.

The surface shape data obtained by the above-described series of operations of the shape recognition device is spatial coordinate data of each point on the high-intensity line L. It was necessary to combine the surface shape data. In particular, when the shape of the object A is complicated, it may not be possible to obtain proper three-dimensional shape data by interpolation processing between the surface shape data, and thus it is necessary to obtain more surface shape data. . Therefore, it takes a lot of processing and time to create three-dimensional shape data by a shape recognition device using the light section method.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-described problems, and achieves downsizing by not particularly requiring a configuration for rotating an object to be recognized as a three-dimensional shape or a configuration for rotating a shape recognition device around the object. It is another object of the present invention to provide a shape recognition device that can obtain three-dimensional shape data of an object more easily than conventional ones.

[0009]

The present invention adopts the following constitution in order to solve the above-mentioned problems. That is, the invention according to claim 1 is an apparatus for recognizing a three-dimensional shape of an object, wherein an image pickup means for acquiring two-dimensional image data of the object, and a distance between the object and the image pickup means is measured. It is characterized by comprising distance measuring means for acquiring as data, and converting means for converting two-dimensional image data of the object into three-dimensional image data of the object based on the distance measuring data.

According to the first aspect of the present invention, the imaging means acquires two-dimensional image data of the object. Further, the distance measuring means acquires the distance between the object and the imaging means as distance measurement data. Then, the conversion means converts the two-dimensional image data of the object into three-dimensional image data based on the distance measurement data. Thereby, the three-dimensional shape of the object is recognized.

Here, examples of the image pickup means include a CCD camera and a digital camera. In addition, the distance measuring unit preferably has a configuration in which, for example, a light source and an imaging unit are arranged at a predetermined distance to an object to be recognized, and the distance between the object and the imaging unit is measured using the principle of triangulation. But,
A method to measure the time until the light projected from the light source returns and find the distance corresponding to that time, or to measure the phase difference between the light projected from the light source to the object and the reflected light from the object and calculate the phase difference The distance between the object and the image pickup means may be measured by a method for calculating the distance corresponding to the distance.

The conversion means can be configured as a function realized by executing a predetermined control program by the CPU, for example. The conversion means calculates, for example, two-dimensional contour vector data of the object and two-dimensional boundary vector data of a surface forming the surface of the object based on the two-dimensional image data of the object. It is preferable that the contour line data and the boundary vector data are converted into a three-dimensional vector based on the distance measurement data to obtain three-dimensional image data of the object (claim 3).

According to a second aspect of the present invention, there is provided an apparatus for recognizing a three-dimensional shape of an object, wherein the imaging means has an imaging surface composed of a plurality of pixels and acquires two-dimensional image data of the object for each pixel. And distance measuring means for acquiring distances between the object and a plurality of distance measuring points set on the imaging surface in association with the respective pixels as distance measurement data; and a two-dimensional image of each of the objects. Conversion means for converting the data into three-dimensional image data of the object based on each distance measurement data.

Further, the shape recognition device according to the present invention may have editing means for editing the three-dimensional image data of the object converted by the conversion means, and based on the three-dimensional image data of the object, A means for creating texture data and color palette data may be provided.

According to a fourth aspect of the present invention, there is provided a shape recognition method for recognizing a three-dimensional shape of an object, the method comprising: obtaining two-dimensional image data of the object by imaging the object; Acquiring a distance from the position as distance measurement data, and converting the two-dimensional image data of the object into three-dimensional image data of the object based on the distance measurement data. I do.

A fifth aspect of the present invention is a shape recognizing method for recognizing a three-dimensional shape of an object, wherein two-dimensional image data of the object for each pixel is obtained by an image pickup means having an image pickup surface comprising a plurality of pixels. Acquiring, as a distance measurement data, the distance between the object and the distance measurement points of the plurality of objects set on the imaging surface in association with each of the pixels, respectively; and Converting the image data into three-dimensional image data of the object based on each distance measurement data.

According to a sixth aspect of the present invention, the three-dimensional image data of the object according to the fourth or fifth aspect is obtained by converting the two-dimensional contour vector data of the object based on the two-dimensional image data of the object, The two-dimensional boundary vector data of the surface forming the surface of the object is calculated, and the contour line vector data and the boundary vector data are obtained by three-dimensional vectorization based on the distance measurement data. Things.

According to a seventh aspect of the present invention, there is provided a recording medium storing a computer program for executing a process of recognizing a three-dimensional shape of an object, wherein the computer is provided with a two-dimensional image of the object obtained by imaging the object. Reading the image data, reading the distance measurement data obtained by measuring the distance between the object and the imaging position, and converting the two-dimensional image data of the object based on the distance measurement data. And a step of converting the data into three-dimensional image data of the object.

Here, the recording medium includes, for example, a ROM,
Various storage devices such as RAM, CD-ROM, magneto-optical disk, floppy disk, hard disk, PD, and magnetic tape are included.

The invention according to claim 8 is a recording medium on which a computer program for executing a process of recognizing a three-dimensional shape of an object is recorded. The recording medium is obtained by an image pickup means having an image pickup surface comprising a plurality of pixels. Reading the two-dimensional image data of the object for each pixel, and measuring the distance between the object and the distance measurement points of the plurality of objects set on the imaging surface in association with the pixels. A computer-readable recording medium storing a computer program for executing the steps of reading the distance measurement data and converting the two-dimensional image data of each object into three-dimensional image data of the object based on each distance measurement data. It is.

According to a ninth aspect of the present invention, the step of converting the two-dimensional image data of the object according to the seventh or eighth aspect into three-dimensional image data of the object is performed based on the two-dimensional image data of the object. Two-dimensional contour vector data of the object,
And calculating two-dimensional boundary vector data of a plane forming the surface of the object, and converting these contour line data and boundary vector data into three-dimensional vectors based on the distance measurement data. , Specified.

[0022]

Embodiments of the present invention will be described below with reference to the drawings. <Configuration of Shape Recognition Apparatus> First, the configuration of the shape recognition apparatus according to the present embodiment will be described. FIG. 1 is a diagram illustrating a configuration example of a shape recognition device. In FIG. 1, the shape recognition device includes:
Control device 1 and imaging device connected to control device 1
(Imaging CCD) 21, infrared light projector 22, turntable 2
3, a lighting device 24, a display 25, an actuator 26, a keyboard 27, and a mouse 28.

Here, the control device 1 includes a CPU 10 and a bus (control bus, address bus,
And an imaging device driver 11a, an A / D conversion circuit 11b, an infrared light projecting device driver 12, a turntable device driver 13, a lighting device driver 14, a display circuit 15, an actuator connected to each other via a data bus B. Driver 1
6, a ROM 17, and a RAM 18.

The imaging CCD device 21 is connected to the imaging CCD device driver 11a and the A / D conversion circuit 11b, and the infrared light projecting device 22 is connected to the infrared light projecting device driver 12. Further, the turntable device 23 is connected to the turntable device driver 13, and the lighting device 24 is connected to the lighting device driver 14. The display 25
Is connected to the display circuit 15, and the actuator 26 is
It is connected to the actuator driver 16. The keyboard 27 and the mouse 28 are each directly connected to the bus B.

The ROM 17 stores various computer programs such as a basic program such as an operation system of the control device 1 and a control program of the shape recognition device, and data used when executing the various programs. The RAM 18 is a work area of the CPU 10, and various programs and data stored in the ROM 17 are loaded as appropriate. Further, the RAM 18 stores the CPU 1
The data processed by 0 is stored. The display circuit 15 also stores image data converted by the A / D conversion circuit 11a based on a command from the CPU 10 and RAM.
The image data stored in 18 is converted into an RGB signal.
The display 25 is a CRT that displays a video (image) on its screen based on the RGB signals transferred from the display circuit 15. Furthermore, the keyboard 27 and the mouse 28 are input devices for the user of the shape recognition device to input instructions, data, and the like.

FIG. 2 is a diagram showing a configuration for measuring a distance from the image pickup device 21 to an object (object) B whose shape is to be recognized. FIG. 2 shows the shape recognition device and the object B
Are shown in plan view from above.

In FIG. 2, the turntable 23 is disposed at a predetermined distance from each of the image pickup device 21 and the infrared light projecting device 22. The rotary table device 23 includes a rotary shaft (not shown) arranged perpendicular to the paper surface, a rotary table 23a attached to the rotary shaft, and a driving device (not shown) for driving a rotary shaft (not shown). Consists of On the upper surface of the turntable 23a, an object B to be recognized as a three-dimensional shape is placed.
The object B according to the present embodiment has a shape obtained by removing a rectangular solid having a smaller volume than the rectangular solid from the rectangular solid (FIG. 6).
reference).

The turntable driver 13 shown in FIG.
A control signal is supplied to a drive device (not shown) of the turntable 23 according to the instruction data from the PU 10 to rotate a rotation shaft (not shown). Accordingly, the turntable 23a rotates by a predetermined rotation angle (for example, 180 °), and accordingly, the object B rotates by the same angle as the rotation angle of the turntable 23a.

As shown in FIG. 2, the infrared light projector 22 includes an LED 22a, an XY table 22b, and a light projecting lens 22c. XY table 22b
Has a plane 40 that is arranged to face the object B, and this plane 40 is in the left-right direction on the paper surface of FIG. 2 (Y in FIG. 2).
Direction) and a direction perpendicular to the paper surface (X direction in FIG. 2).

The LED 22a is provided in a fixed state on the above-described plane 40, and moves as the XY table 22b moves in the X or Y direction. This L
The ED 22a irradiates the object B with infrared rays for measuring the distance between the object B and the imaging surface 41 of the imaging CCD 21a.

The light projecting lens 22c is disposed on an optical path between the LED 22a and the object B, and collects infrared light emitted from the LED 22a and emits the infrared light to the object B. The light projecting lens 22c and the LED 22a are arranged at a predetermined distance from each other, and the focal point of the infrared light emitted from the light emitting end face of the light projecting lens 22c is located near the object B.

The infrared light projector driver 1 shown in FIG.
2 moves the XY table 22b in the X direction or the Y direction according to the instruction data from the CPU 10. Accordingly, the infrared LED 22a moves in the X direction or the Y direction. Then, the angle theta ₁ formed by the center axis and the optical axis of the projection lens 22c of the emitted infrared light from LED22a shown in FIG. 2 is changed.

Further, as shown in FIG.
On the optical path between the object a and the object B, an illuminating device 24 is disposed so as to be able to advance and retreat by an actuator (not shown). The illumination device 24 is configured to emit infrared light while the shape recognition device irradiates infrared light.
Retreat outside the optical path between the LED 22a and the object B. On the other hand, while the imaging device 21 performs imaging of the object B, the illumination device 24 advances on the optical path between the LED 22a and the object B, and emits white light to the object B for the imaging CCD 21a to image the object B. And irradiate.

Further, as shown in FIG.
1 is arranged at a predetermined distance from the infrared light projector 22 along the X direction of the XY table 22b. The imaging device 21 includes an imaging CCD 21a, an imaging lens 21b,
And a visible light cut filter 26a.

Here, the imaging CCD 21a is formed in a flat plate shape having a rectangular imaging surface 41, and the imaging surface 41 is arranged facing the object B. FIG. 3 is a configuration diagram of the imaging surface 41. 3, an imaging surface 41 is composed of a plurality of CCD elements (pixels) 31 arranged in a matrix, and forms a CCD element forming an upper left corner of the imaging surface 41 in FIG. 3 (an upper right corner of the imaging surface 41 in FIG. 2). 31 is the origin
It is set as a CCD element 31 forming (0,0).
In addition, the imaging surface 41 is the imaging surface 4 in the horizontal direction of the paper surface of FIG.
1 is set as the X direction, and the vertical direction of the paper is
1 is set as the Y direction.

Further, as shown in FIG.
According to the arrangement of the CCD elements 31, a distance measuring point 32 for measuring the distance between the object B and the imaging CCD 21a is set for each of the four adjacent CCD elements 31 (four rectangular CCD elements 31). Have been. The correspondence between the CCD elements 31 and the distance measuring points 32 is stored in the RAM 18 as distance measuring point data when the CPU 10 executes the control program.

In this embodiment, a display buffer area for storing image data is formed in the RAM 18 and has the same number of storage areas as the CCD elements 31. Image data corresponding to the output of each CCD element 31 is stored in each storage area of the display buffer area. Although a color CCD is used as the imaging CCD 21a according to the present embodiment, a monochrome CCD may be used. Also, the CCD element 31 corresponding to one distance measuring point 32
Can be set as appropriate.

The imaging lens 21b is arranged on the optical path between the object B and the imaging CCD 21a, as shown in FIG. The imaging lens 21b and the above-described light projecting lens 22c are arranged at a predetermined distance on the same plane orthogonal to the center axis of the imaging lens 21b and the center axis of the light projecting lens 22c, respectively.

The light reflected from the object B is incident on the incident end face of the imaging lens 21b. The focal position of the imaging lens 21b from the exit end face is set on the imaging surface 41 of the imaging CCD 21a. Also, this imaging lens 21b
The angle θ ₂ between the optical axis of the light emitted from and the center axis of the imaging lens 21b changes according to the change of the angle θ ₁ accompanying the movement of the LED 22a described above. Such an imaging CC
D21a converts the image incident on the imaging surface 41 into the CCD element 31
A / D conversion circuit 11b shown in FIG.
Transfer to The A / D conversion circuit 11b includes the imaging CCD 21
The two-dimensional image data of the object B is obtained by performing an analog-to-digital conversion of the output from a.

The visible light cut filter 26a is connected to the CPU 1
By driving the actuator 26 receiving a command from 0, the actuator 26 can move forward and backward on the optical path between the imaging CCD 21a and the imaging lens 21b. That is, when the object B is irradiated with the infrared light, the visible light cut filter 26a advances on the optical path between the imaging lens 21b and the imaging CCD 21a, and reflects the reflected light of the object B emitted from the imaging lens 21b. CCD2 removes visible light components from infrared light and captures only infrared light components
1a.

The CPU 10 is started when the power of the shape recognition device is turned on, and controls the entire shape recognition device by executing a control program stored in the ROM 17. More specifically, the CPU 10 performs a process of inputting instruction data for moving the visible light cut filter 26 a back and forth to the actuator device driver 16.
A process of inputting instruction data for capturing an image of the object B into the device 1a, a process for image data output from the A / D conversion circuit 11b, and the illumination device 2 via the illumination device driver 14.
4, the process of moving the visible light cut filter 26a through the actuator driver 16,
A process of reading and writing various data to and from M18, a process of transferring image data to the display circuit 15, a process of receiving input signals from the keyboard 27 and the mouse 28, and the like are performed. <Operation of Shape Recognition Apparatus> Next, the operation of the above-described shape recognition apparatus will be described. 4 and 5 are flowcharts showing the processing of the CPU 10 when the shape recognition device operates.
The shape recognition device starts operation when a power source (not shown) is turned on.

First, the CPU 10 of the shape recognition device is activated. The CPU 10 controls each part (RA) of the control device 1 based on the operating system stored in the ROM 17.
M18, each driver, the A / D conversion circuit 11a, etc.) are initialized. When this initialization is completed, the CPU 10
Stores the control program stored in the ROM 17 in the RAM 1
8 and execute the control program. As a result, for example, the characters “Start recognition processing of object B? [Y / N]” are displayed on the display screen of the display 25.

At this time, the user of the shape recognition device places the object B on the turntable 23a and, for example,
When the "Y" key is pressed, a shape recognition processing command for the object B is input to the CPU 10 via the keyboard 27 and the bus B. Thereby, the CPU 10 starts the processing shown in FIG. However, at the time of this start, the illumination device 24 is retracted from the optical path between the light projecting lens 22c and the object B, and the visible light cut filter 26a is located between the imaging CCD 21a and the imaging lens 21b. It is assumed that the robot has been retracted from the optical path.

In FIG. 4, in S001, the CPU 40
Input predetermined instruction data to the actuator device driver 16. By the process of S001, the actuator 26 is driven and the visible light cut filter 26
a is inserted into the optical path between the imaging CCD 21a and the imaging lens 21b.

In the next step S 002, the CPU 10 inputs the image capturing start instruction data to the image capturing device 21. This S00
By the processing of 2, the LED 22a of the infrared light projector 22
Emits infrared light. This infrared light is transmitted through the projection lens 2
The surface of the object B is irradiated via 2c. Then, the reflected light of the infrared light reflected by the object B is
1b and imaging C through the visible light cut filter 26a.
The light enters the CD 21a. Thereby, the imaging CCD 21
a captures an image of the object B. The output of the imaging CCD 21a (the output of each CCD element 31) is supplied to the A / D conversion circuit 21a.
Is converted into image data of the object B (luminance data of each CCD element 31). Then, the CPU 10 stores each image data as the first image data in the display buffer area of the RAM 18.

In the next S003, the CPU 10
The distance measuring point 32 is specified based on the first image data stored in 18. That is, the CPU 10 calculates the luminance distribution of each CCD element 31 (pixel) from the first image data, and takes the average value of the luminance data of each CCD element 31 to obtain the threshold value of the luminance of each CCD element 31. Is calculated.

Subsequently, the CPU 10 compares the luminance data of each CCD element 31 constituting the first image data with the calculated threshold value, and outputs the luminance data equal to or higher than the threshold value.
The D element 31 is detected. At this time, since the brightness data of the CCD element 31 on which the reflected light of the object B is incident has a value equal to or greater than the threshold value, an approximate contour of the object B projected on the imaging surface 41 of the imaging CCD 21a is obtained. It will be.

Subsequently, the CPU 10 reads out the distance measurement point data from the RAM 18 and acquires one or more distance measurement points 32 corresponding to the detected pixel based on the distance measurement point data. And CPU1
0 is the distance measurement point data by associating the acquired position information of each distance measurement point 32 with the position information of each CCD element 31 surrounded by each distance measurement point 32 (the position information of a pixel); Number each ranging point data to RAM18
Respectively. At this time, the CPU 10 specifies the ranging point 32 in the data with the smallest number among the ranging point data.

In the next S004, the CPU 10
The LED 22a is moved so that the infrared light impinges on the distance measuring point 32 specified in Step 3. That is, the CPU 10
The distance measuring point data of the specified distance measuring point 32 is read from the AM 18, and instruction data corresponding to the distance measuring point data is given to the infrared light projector driver 12. Then, the infrared light projector driver 12 moves the XY table 22b to arrange the LEDs 22a at appropriate positions. As a result, the reflected light of the infrared light emitted from the LED 22 a
The light enters the CD element 31.

In the next step S005, the CPU 10 determines the distance from the object B to the imaging plane 41 (each distance measuring point 32) based on the position of the infrared light LED 22a and the position of the distance measuring point 32 specified in S003. Measure. That is, CP
U10 is the LED 22 moved by the processing of S004.
a, and based on the position information of the LED 22a, the angle between the optical axis of the infrared light emitted from the LED 22a to the surface of the object B and the central axis of the light projecting lens 22c (see FIG. 2). θ ₁ ) is calculated. The CPU 10 also controls the CCD element 31 at the distance measuring point 32 where the infrared reflected light is incident.
An angle (θ ₂ ) between the optical axis of the reflected light incident on the CCD element 31 and the central axis of the imaging lens 21b is calculated based on the (pixel) position data.

Subsequently, the CPU 10 determines the angle θ ₁ , the angle θ _2, and distance data in the horizontal direction (the left-right direction (Y direction) on the plane of FIG. 2) from the light projecting lens 22 c to the imaging lens 21 b. Then, the distance from the surface of the object B to the corresponding ranging point 32 is calculated. That is, the CPU 10
Is the distance between the object B and the distance measuring point 32 using the principle of triangulation.
Measure the distance between Then, the CPU 10 associates each of the calculated distances with the number of the distance measurement point data to obtain distance measurement data, and stores the distance measurement data in the RAM 18.

In the next step S006, the CPU 10 determines whether or not distance measurement has been performed for all the distance measurement points 32. That is, the CPU 10 determines whether or not there is distance measurement point data having a number larger than the distance measurement point data number measured in S005. At this time, the CPU 10
When it is determined that there is distance measurement point data with a large number
In (S006; YES), the process proceeds to S007.

On the other hand, when the CPU 10 determines that there is no ranging point data with a large number (S006;
(NO), the process returns to S003, and the processes of S003 to S006 are repeated until the determination of YES is made in S006. However, S003 to S00 after the second round
The processing of No. 6 is performed on the ranging point data having the next largest number to the number of the ranging point data subjected to the previous processing of S003 to S006 among the plurality of ranging point data acquired in S003. Done.

If the processing has proceeded to S007, the CPU
10 inputs to the actuator driver 16 instruction data for taking out the visible light cut filter 26 a from the imaging device 21. Thereby, the visible light cut filter 26a
Is driven by the actuator 26, the imaging CCD 2
Retreat from the optical path between 1a and the imaging lens 21b.
At this time, the CPU 10 controls the infrared light projector driver 12
And the illumination device driver 14. Thereby, the light emission of the LED 22a is stopped.

In the next step S008, the CPU 10 inputs a release command to the imaging device driver 11b and the illumination device driver 14. Then, the illumination device 24 is turned on
The laser beam advances on the optical path between 2c and the object B, and irradiates the object B with white light. Thereby, the reflected light of the white light by the object B is transmitted through the imaging lens 21b to the imaging CCD 21a.
Incident on the image pickup surface 41. The output from the imaging CCD 21a (the output of each CCD element 31 (pixel)) is converted from analog to digital in the A / D conversion circuit 11a, and becomes two-dimensional image data of the object B for one screen. CPU10
Uses this image data as the second image data in the RAM 18
In the display buffer area.

In the next S009, the CPU 10 sets
The respective distance measurement data obtained in step 5 is associated with pixel position data. That is, the CPU 10 detects a corresponding pixel from the second image data based on the position information of the ranging point 32 included in each ranging point data. Then, CP
U10 associates the distance measurement data of each distance measurement point 32 with each detected pixel.

In the next S010, the CPU 10 sets the object B
Trace the outline of. That is, the CPU 10 calculates a threshold from the luminance distribution by calculating a luminance distribution from the second image data and taking an average value. continue,
The CPU 10 starts from the origin (0, 0) of the image data,
The luminance value of each pixel is compared with the threshold value in the order of (0, 1) (0, n) (see FIG. 3). At this time, the CPU 10
In the 0th row, the first detected pixel and the last detected pixel are detected (pixels detected in this processing are referred to as “contour pixels”), and RAM 18 is used as contour data.
To be stored. The CPU 10 similarly performs the above-described processing for the first to n-th rows.

Subsequently, the CPU 10 determines the origin of the image data.
Starting from (0,0), (1,0), (2,0) ... (n,
The luminance value of each pixel is compared with the threshold value in the order of 0). Also in this case, the CPU 10 detects the first detected pixel and the last detected pixel in the 0th column (the pixel detected in this processing is referred to as “outline pixel”), and the outline data is obtained. And stored in the RAM 18. CP
U10 similarly performs the above-described processing for the first to n-th columns.

In the next S011, the CPU 10 executes
Based on the contour data obtained at 0, two-dimensional contour vector data of the object B is calculated. That is, the CPU 10
The position information of each contour pixel included in each contour data is substituted into a so-called HOUGH equation (HOUGH conversion is performed). Subsequently, the CPU 10 sets H for each contour pixel.
The calculation results of the OUGH equation are compared respectively. And C
The PU 10 calculates two-dimensional outline vector data of the object B having the start pixel of the vector and the end pixel of the vector, respectively, based on the comparison result and the outline data.

In the next S012, the CPU 10 executes
Each contour line vector data calculated in 1 is stored in the RAM 18. As shown in FIG. 5, in the next S013,
The CPU 10 traces the boundary of the luminance (the boundary of the surface forming the surface of the object B) toward the inside of each contour pixel detected in S010. That is, the pixel surrounded by each contour pixel is compared with the threshold value detected in S010,
A pixel having a luminance value equal to or higher than the threshold value is detected (a pixel detected by this processing is referred to as a “boundary pixel”). Subsequently, the CPU 10 stores the detected pixels in the RAM 18 as boundary data.

In the next S014, the CPU 10 executes
Based on the boundary data obtained in step 3, two-dimensional boundary vector data of the object B is calculated in the same manner as in S011. That is, the CPU 10 substitutes the position information of the boundary pixel included in each boundary data into the HOUGH equation.
(Perform HOGHH conversion). Subsequently, the CPU 10 compares the calculation results of the HOUGH equation for each boundary pixel. Subsequently, the CPU 10 calculates two-dimensional boundary vector data of the object B having the start pixel of the vector and the end pixel of the vector based on the comparison result and the boundary data.

In the next S015, the CPU 10 executes
Each boundary vector data calculated in step 4 is stored in the RAM 18. In the next S016, the CPU 10 executes S005
The two-dimensional boundary vector data obtained in S014 is converted into a three-dimensional vector based on the distance measurement data obtained in. That is, the CPU 10 calculates the distance measurement points 32 corresponding to the start pixel or the end pixel of each two-dimensional boundary vector data. Subsequently, the CPU 10 sets each of the distance measurement data corresponding to each of the calculated distance measurement points 32 as the position of the start point pixel or the end point pixel corresponding to the distance measurement point 32, thereby obtaining three-dimensional boundary vector data. Convert.

As described above, the CPU 10 converts the two-dimensional boundary vector data (two-dimensional coordinates) into three-dimensional boundary vector data (three-dimensional coordinates) based on the corresponding distance measurement data. Then, the CPU 10 stores each conversion result in each storage area of the corresponding display buffer area.

In S017, the CPU 10 creates modeling data of the object B using the two-dimensional contour vector data of the object B obtained in S011. That is, CPU1
0 detects, at regular intervals, pixels located along the two-dimensional contour vector among the pixels surrounded by the contour vector, based on the two-dimensional contour vector data. Subsequently, the CPU 10 converts each of the detected pixels and the contour line vector into three-dimensional vector data based on the distance measurement data.

That is, the CPU 10 converts the two-dimensional contour vector data (two-dimensional coordinates) into three-dimensional contour vector data (three-dimensional coordinates) based on the corresponding distance measurement data using the same method as in S016. ). And CP
U10 stores the converted three-dimensional contour line vector data in each storage area of the corresponding display buffer area. Thus, the display buffer area of the RAM 18 stores the three-dimensional vector data of the object B forming the modeling data.

In the next S018, the CPU 10 executes
An instruction to display the modeling data obtained in step 7 on the display 25 is given to the display circuit 15. Thus, a three-dimensional image of the object B is displayed on the display 25.

In the next S019, the CPU 10 causes the display 25 to display an operation instruction of the user of the shape recognition device. That is, the CPU 10 displays, for example, a “M” key to edit the modeling data, an “S” key to perform the segmentation processing, and Press the "E" key. "

In the next step S020, the CPU 10 accepts that any one of the "M", "S" or "E" keys on the keyboard 27 is pressed. If any one of these keys is pressed, the CPU 10 proceeds to S021.
Proceed to. In the next step S021, the CPU 10 determines whether or not the “M” key has been pressed. At this time, if it is determined that the "M" key has been pressed, the CPU 10 advances the process to S022, and performs a modeling data editing process. For example, the keyboard 27 or mouse 28
In accordance with an instruction input from, processing for correcting the modeling data is performed. Then, when the process of S022 ends, the CPU 10 returns the process to S019. On the other hand, when it is determined that a key other than “M” has been pressed (S0
21; NO), the CPU 10 advances the processing to S023.

If the processing has proceeded to S023, the CPU
Step 10 determines whether the "S" key has been pressed. At this time, when it is determined that the “S” key has been pressed (S0
23; YES), the CPU 10 advances the processing to S024. On the other hand, if it is determined that a key other than the “S” key has been pressed (S023; NO), the CPU 10
Advances the process to S026.

When the processing has proceeded to S024, the CPU
Numeral 10 subdivides each surface surrounded by the contour line vector and the boundary vector. The CPU 10 extracts pixels from the portion surrounded by the contour line vector and the boundary vector at regular intervals in the column direction and the row direction of the image data. CPU1
0 adds the corresponding distance measurement data to the extracted pixel.

At S025, the CPU 10 stores the image data of the object B subdivided by the processing at S024 into the RAM 18 as texture data and color pallet data.

In S024, the CPU 10 causes the display 25 to display, for example, a text character of “Do you want to end the process [Y / N]” and presses the “Y” or “N” key from the keyboard 27? Wait for When the “Y” key is pressed, the processing of the CPU 10 ends and the operation of the shape recognition device ends. On the other hand, when the “N” key is pressed, C
The PU 10 advances the process to S027.

In S027, the CPU 10 displays, for example, “Do you want to rotate the turntable? [180
-359 / N] ", and waits for the user to press the numbers 180-359 and the return key or the" N "key. And
If the “N” key is pressed (S027; NO), C
The PU 10 returns the process to S019. In contrast, 18
If any of the numbers 0 to 359 and the return key are pressed, the CPU 10 advances the processing to S028.

If the processing has proceeded to S028, the CPU
Numeral 10 inputs a command to the turntable device driver 23 to rotate the turntable 23 by a rotation angle corresponding to the input number. Then, the CPU 10 returns the processing to S001.
At this time, for example, when a numeral 180 is input from the keyboard 27, the turntable 23 rotates 180 ° and the object B rotates 180 °. Thereafter, S001 to S02
When the processing in step S6 is performed and the data is combined with the data obtained in the previous processing in steps S001 to 026, the RAM 28 stores a complete three-dimensional modeling data of the object B. <Operation of Embodiment> Next, a process until modeling data, texture data, and the like of the object B are created by the shape recognition device according to the present embodiment will be described.

As described above, prior to the shape recognition of the object B, the user places the object B on the turntable 23a of the shape recognition device and starts the three-dimensional shape recognition via the keyboard 27 or the mouse 28. Enter the instruction data. Then
The CPU 10 goes through the processes of S001 and S002 in the drawing,
The loop from S003 to S006 is repeated. here,
The CPU 10 acquires distance measurement data between the object B and each distance measurement point 32.

All distance measuring points 3 set on object B
When the acquisition of the distance measurement data for the
Performs the processing of S007 and S008, and acquires image data of the object B. FIG. 6 is a display example of image data displayed on the display 25 after the processing of S008 is completed.

Next, the CPU 10 executes S009 to S012.
Are performed, and the contour vector data of the object B is calculated based on the acquired image data. FIG. 7 shows outline vector data displayed on the display device 25. Subsequently, the CPU 10 performs the processing from S013 to S015 to calculate the boundary vector data. FIG.
It is a screen display example of the outline vector data and the boundary vector data displayed on the display 25 after the processing of 015 is completed.

Subsequently, the CPU 10 determines in steps S016 to S01
8 to create modeling data based on the contour vector data. FIG. 8 is a screen display example of the modeling data displayed on the display 25 after the processing of S018 is completed.

Further, the CPU 10 sets "M" in S020.
If the key is pressed, the modeling data is edited (S022). Further, the CPU 10 executes S02
4, the processing of S025 is performed, and each surface surrounded by the contour line vector and the boundary vector is subdivided to create texture data. FIG. 9 is a screen display example showing the modeling data and the texture data displayed on the display 25 after the processing of S022 and S025 is completed.

As described above, according to the present embodiment, the object B
Image data for each CCD element 31 (pixel) for one screen
By acquiring (two-dimensional image data), acquiring the distance measurement data of the distance measurement point 32 corresponding to each pixel, and performing three-dimensional conversion on each two-dimensional image data based on each distance measurement data, The modeling data of the object B is created. That is, the three-dimensional shape of the object B is recognized.

Therefore, it is possible to recognize the three-dimensional shape of the object B without rotating the object B or rotating the imaging device 21 or the like as in the related art. Further, in the present embodiment, since the turntable device 23 is provided, the complete three-dimensional shape of the object B can be recognized by, for example, rotating the object by 180 °. Therefore, the number of times the object B is rotated can be significantly reduced, and the number of processes (operations) of the shape recognition device can be simplified as compared with the shape recognition device using the light section method. Can be planned.

By the way, the shape recognition device according to the present invention
As described above, since a configuration for rotating the object B (the turntable device 23) and the like are not particularly required, as shown in FIG. 11, for example, a digital camera K1 and a cable C and an interface (see FIG. (Not shown) and a notebook personal computer (personal computer) K2 connected thereto. In this case, for example, as shown in FIG.
The digital camera K1 includes an infrared light projecting device 22, a lighting device 24, and an actuator 26, and the personal computer K2 includes a control device 1, a liquid crystal display 25, a keyboard 27, and a mouse 28.

The two-dimensional image data of the object B (the imaging CCD 21a)
Output) and the distance (distance measurement data) between the object B and the distance measurement point 32, and these data are transferred to the personal computer K2, and transferred from the digital camera K1 by the control device 1 (CPU 10) of the personal computer K2. 3D image data of the object B (modeling data) based on the obtained data
Is configured to be created. With such a configuration, the size of the shape recognition device can be reduced, so that its portability can be improved and modeling data of the object B can be easily created. Note that in the shape recognition device shown in FIG. 11, data transfer between the digital camera K1 and the personal computer K2 may be wireless.

The shape recognizing device according to the present invention can also be configured by, for example, storing the above-described control program in a storage device of a mobile computer (for example, an electronic organizer) equipped with a CCD camera. In this case, for example, it is preferable that the processing until the above-described modeling data is created is performed on a mobile computer. Then, the modeling data created by the mobile computer is transferred to a host computer such as a personal computer or a workstation, and modeling software or CAD executed by the host computer is transferred.
Editing or subdivision of the modeling data may be performed on software. In this way, the portability of the shape recognition device can be further improved.

[0085]

According to the present invention, a shape recognition apparatus, a shape recognition method,
According to the recording medium on which the computer program is recorded, the configuration for rotating the object to be recognized as the three-dimensional shape or the configuration for rotating the shape recognition device around the object is not particularly required, and the size can be reduced. 3D shape data of an object can be easily obtained.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration example of a shape recognition device according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram of a configuration for performing distance measurement of a shape recognition device.

FIG. 3 is a diagram showing an imaging surface of an imaging CCD.

FIG. 4 is a flowchart showing processing of a CPU of the shape recognition device.

FIG. 5 is a flowchart showing processing of a CPU of the shape recognition device.

FIG. 6 is a display screen example of a display.

FIG. 7 is a screen display example of a display.

FIG. 8 is a display example of a display screen.

FIG. 9 is a display screen example of a display.

FIG. 10 is a display screen example of a display.

FIG. 11 is a diagram showing a modification of the shape recognition device.

FIG. 12 is a diagram showing a modification of the shape recognition device.

FIG. 13 is an explanatory diagram of a conventional shape recognition device.

[Explanation of symbols]

 B Object K1 Digital Camera K2 Personal Computer 1 Controller 10 CPU (Conversion Unit) 17 ROM 18 RAM 21 Imaging Device (Imaging Unit) 22 Infrared Light Projector (Ranging Unit)

Claims (9)

[Claims] 1. An apparatus for recognizing a three-dimensional shape of an object, comprising: imaging means for acquiring two-dimensional image data of the object; and acquiring a distance between the object and the imaging means as distance measurement data. A shape recognition apparatus comprising: a distance measuring unit; and a converting unit that converts two-dimensional image data of the object into three-dimensional image data of the object based on the distance measurement data. 2. An apparatus for recognizing a three-dimensional shape of an object, comprising: an image pickup means having an image pickup surface composed of a plurality of pixels and acquiring two-dimensional image data of the object for each pixel; Distance measuring means for acquiring distances between a plurality of distance measuring points of the objects set on the imaging surface and the object as distance measuring data; and measuring two-dimensional image data of each of the objects by distance measuring. Conversion means for converting each of the data into three-dimensional image data of the object based on the data. 3. The conversion means calculates two-dimensional contour vector data of the object and two-dimensional boundary vector data of a surface forming the surface of the object based on the two-dimensional image data of the object. 3. The method according to claim 1, further comprising obtaining three-dimensional image data of the object by converting these contour vector data and boundary vector data into three-dimensional vectors based on the distance measurement data. The shape recognition device described in the above. 4. A method for recognizing a three-dimensional shape of an object, comprising: acquiring two-dimensional image data of the object by imaging the object; and determining a distance between the object and the imaging position. A shape recognition method, comprising: acquiring as distance measurement data; and converting two-dimensional image data of the object into three-dimensional image data of the object based on the distance measurement data. 5. A method for recognizing a three-dimensional shape of an object, comprising: acquiring two-dimensional image data of the object for each pixel by an imaging unit having an imaging surface including a plurality of pixels; Obtaining distance measurement points of the plurality of objects set on the imaging surface in association with pixels as distance measurement data, respectively; and obtaining two-dimensional image data of the respective objects as distance measurement data. Respectively converting the three-dimensional image data into three-dimensional image data of the object on the basis of 6. The three-dimensional image data of the object is based on the two-dimensional image data of the object, the two-dimensional contour vector data of the object, and the two-dimensional surface data of the object.
The shape according to claim 4 or 5, wherein the shape is obtained by calculating three-dimensional boundary vector data and converting these outline vector data and boundary vector data into three-dimensional vectors based on the distance measurement data. Recognition method. 7. A recording medium on which a computer program for executing a process of recognizing a three-dimensional shape of an object is recorded.
Reading two-dimensional image data; reading distance measurement data obtained by measuring a distance between the object and the imaging position; and obtaining two-dimensional image data of the object based on the distance measurement data. Converting the image data into three-dimensional image data of the object. 8. A recording medium on which a computer program for executing a process of recognizing a three-dimensional shape of an object is recorded. The computer stores, for each pixel, an object obtained by an image pickup means having an image pickup surface including a plurality of pixels. Reading the two-dimensional image data respectively, and measuring the distance between the object and the distance measurement points of a plurality of objects set on the imaging surface in association with each of the pixels. A recording medium storing a computer program for executing a reading step and a step of converting the two-dimensional image data of each object into three-dimensional image data of the object based on each distance measurement data. 9. The method according to claim 9, wherein the two-dimensional image data of the object is
Converting the image data into two-dimensional image data;
Calculating two-dimensional contour vector data of the object and two-dimensional boundary vector data of a surface forming the surface of the object based on the two-dimensional image data; 9. A recording medium storing a computer program according to claim 7, further comprising a step of performing three-dimensional vectorization based on the distance measurement data.