JPH0814829A - Object measurement device and method - Google Patents

Abstract

PURPOSE:To accurately and quickly calculate the three-dimensional position of an object, even when the characteristic point thereof is unavailable, by providing a calculation means or the like for computing the unknown of the dislocation and displacement angle of the object relative to a reference position. CONSTITUTION:The reference coordinate data of an attention point on an object assumed to be at the preset reference position is extracted with the first extraction means 7. Also, the detected coordinate data of an actual object attention point detected with an object detection means 3 is extracted with the second data extraction means 8. Then, a sight line vector is established to give an equation for finding the amount of the dislocation or the like of the object relative to the reference position, on the basis of the detected coordinate data and the reference coordinate data. Then, the equation obtained depending on the sight line vector is solved, and the unknown of a positional conversion factor regarding one of the reference coordinate data of the attention point and the detected coordinate data is calculated, thereby finding the amount of the dislocation of the object as well as the displacement angle thereof relative to the reference position.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an object measuring apparatus and method for measuring the three-dimensional position of an object having an arbitrary shape.

[0002]

2. Description of the Related Art Conventionally, as shown in, for example, Japanese Unexamined Patent Publication No. 2-38804, model data of an object assumed to be at a preset reference position and an object arranged at an actual measurement position. In the object measuring device for measuring the amount of displacement and the displacement angle of the object, the object arranged at the measuring position is photographed by the camera, and in order to match the photographed object image with the model data, on the camera screen. In, the installation state of the object is measured by repeating the operation of moving the model data in parallel and rotating the model data.

[0003]

In the conventional device described in the above publication, a corner portion of an object is set as a feature point, and an image of this feature point matches the feature point of the object assumed to be at the reference position. By repeating the above operation on the image, it is configured to measure the amount of displacement of the object, etc., so that the position of the object can be measured when the corner portion that is the feature point cannot be found. There is a problem that you cannot do it. Moreover, it takes time to search for a point on the screen at which the feature point of the object assumed to be at the reference position and the captured image of the object arranged at the measurement position are exactly coincident with each other. There is a problem that it is difficult to accurately measure the amount of displacement and the displacement angle.

The present invention has been made to solve the above problems, and an object measuring device and an object measuring device capable of accurately and quickly measuring the three-dimensional position of an object even when there are no characteristic points. It is intended to provide a way.

[0005]

The invention according to claim 1 is
Input means for inputting the edge line as data of a three-dimensional curve defined by a parameter assuming that an object exists at a preset reference position, and a position on the edge line according to the input data A first data extracting means for extracting the reference coordinate data of the target point, an object detecting means for detecting the edge line of the object arranged at the measurement position, and the above-mentioned edge line on the basis of the detection data of the object detecting means. The second data extracting means for extracting the detected coordinate data of the point of interest set in (1) and one of the two coordinate data extracted by the first and second data extracting means are composed of the positional deviation amount and the unknown number of the displacement angle. Position conversion means for position conversion using position conversion parameters, and object detection means based on the coordinate data of the point of interest converted by this position conversion means. First vector setting means for setting a line-of-sight direction vector from the point of view to the target point after position conversion, and a line of sight from the detected point of view of the object detecting means to the target point on the edge line based on the other of the coordinate data. Second vector setting means for setting a direction vector, equation setting means for setting an equation under the constraint condition that the directions of both line-of-sight direction vectors set by the first and second vector setting means match, and this equation And a calculation means for calculating the unknown amount of the positional deviation amount and the displacement angle with respect to the reference position of the object.

According to a second aspect of the present invention, in the object measuring device according to the first aspect, the irradiation means for irradiating the edge line of the object arranged at the measurement position with the slit light, and the irradiation means. An object detecting unit including a camera that captures an image of an edge-line-shaped point of interest is provided.

According to a third aspect of the present invention, the edge line of the object is represented as a three-dimensional curve defined by a parameter based on the data set assuming that the object exists at a preset reference position. After that, while extracting the reference coordinate data of the point of interest located on this edge line, the edge line of the object arranged at the measurement position is detected by the object detection means, and based on this detection data, it is detected on the edge line. After extracting the detected coordinate data of the target point to be located, one of the detected coordinate data of the target point and the reference coordinate data is subjected to position conversion using a position conversion parameter consisting of an unknown number of displacement and displacement angle, If the first line-of-sight direction vector from the detection viewpoint of the object detecting means to the target point after position conversion is set based on the coordinate data after position conversion, Moni, after setting the second sight line direction vector from the detected viewpoint towards the target point on the edge line of the object detecting means on the basis of the other of the two coordinate data,
An equation is set under the constraint that the directions of the first and second line-of-sight direction vectors match, and by solving this equation, the amount of displacement and the unknowns of the displacement angle with respect to the reference position of the object are calculated. It is a thing.

According to a fourth aspect of the invention, in the object measuring method according to the third aspect, it is assumed that the edge line detection data of the object at the measurement position detected by the object detecting means and the reference position are present. Based on the virtual data obtained when the edge line of the object is detected by the object detecting means, the initial value of the unknown number used in the numerical solution of the nonlinear equation is set.

According to a fifth aspect of the present invention, in the object measuring method according to the third aspect, a plurality of detection viewpoints of an object detecting means for detecting an edge line of an object arranged at a measurement position are set, It is configured to detect edge lines of an object from multiple directions.

According to a sixth aspect of the present invention, in the object measuring method according to the third aspect, the edge lines of the object arranged at the measurement position are respectively detected by a plurality of object detecting means, and based on the detected data. The detection coordinate data of the point of interest located on the edge line is extracted.

According to a seventh aspect of the present invention, in the object measuring method according to the third aspect, an end plate is provided which projects an edge line of the object disposed at a measurement position, and the object detecting means detects the edge line of the object and the edge line of the object. Detect that mirror image,
The detection coordinate data of the target point located on the edge line is extracted based on this detection data.

According to an eighth aspect of the present invention, in the object measuring method according to the third aspect, the edge line is located on an xy plane of a coordinate system that constitutes a reference position of an object having a circular edge line. Thus, the reference coordinate system of the object is set.

According to a ninth aspect of the present invention, in the object measuring method according to the eighth aspect, in addition to the point of interest set on the edge line, the center of gravity of a circle forming the edge line is set as a point of interest and detection is performed. The line-of-sight direction vector from the detection viewpoint of the means toward the center of gravity of the circle is set.

[0014]

According to the first aspect of the present invention, by setting the target point at an arbitrary position on the edge line of the object, the target point on the object which is assumed to be at the preset reference position is detected. The reference coordinate data is extracted by the first data extracting means, and the detected coordinate data of the target point of the actual object detected by the object detecting means is the second.
A line-of-sight direction vector for setting an equation for obtaining the amount of displacement of the object with respect to the reference position is set based on the detected coordinate data and the reference coordinate data extracted by the data extracting means. Then, by solving the equation set according to this line-of-sight belt and calculating the unknown number of position conversion parameters used when performing position conversion of either the reference coordinate data or the detected coordinate data of the point of interest, the reference of the object The amount of displacement and the displacement angle with respect to the position will be calculated.

According to the second aspect of the present invention, the change in brightness that occurs at the intersection of the slit light irradiated on the object by the irradiation means and the edge Latin of the object is photographed by the camera, so that the edge line is displayed. The coordinate data of the point of interest can be easily and accurately extracted.

According to the third aspect of the invention, the point of interest is set at an arbitrary position on the edge line of the object, and the reference coordinate data of the point of interest at the preset reference position and the object detecting means are used. A line-of-sight direction vector for setting an equation for obtaining the amount of displacement of the object with respect to the reference position is set based on the detected detected object coordinate data of the actual object. Then, by solving the equation set based on the line-of-sight direction vector to calculate the unknown number of position conversion parameters used when performing position conversion of one of the reference coordinate data or the detected coordinate data of the point of interest, The amount of displacement and the displacement angle with respect to the reference position will be calculated.

According to the invention described in claim 4, when the deviation amount between the detection position of the object arranged at the measurement position and the preset reference position of the object is large, the deviation amounts of both positions are determined. By solving the above equations by using the correspondingly set initial values of unknowns, the solution of the above unknowns can be quickly obtained by the numerical solution of the nonlinear equation, while preventing incorrect solutions from being obtained. Become.

According to the fifth aspect of the invention, the three-dimensional shape of the object arranged at the measurement position is accurately detected from a plurality of angles, and the line of sight for setting the above equation based on the detected data. The direction vector will be set appropriately.

According to the invention of claim 6, the three-dimensional shape of the object arranged at the measurement position is accurately detected by the plurality of object detecting means, and the above equation is set based on the detected data. Will be set appropriately.

According to the invention described in claim 7, the single object detecting means simultaneously detects the object arranged at the measuring position and the mirror image of the object projected on the end plate, and the detected data is detected. Based on this, the three-dimensional shape of the object can be accurately grasped.

According to the invention described in claim 8, based on the reference coordinate data and the detected coordinate data of the point of interest set on the circumference forming the edge line of the object, the positional deviation amount with respect to the reference position of the object and The calculation for obtaining the displacement angle can be easily performed.

According to the invention described in claim 9, the reference coordinate data and the detected coordinate data of the point of interest set on the circumference forming the edge line of the object, and the point of interest set at the center of gravity of the circle. Based on the reference coordinate data and the detected coordinate data, the positional deviation amount and the displacement angle of the object with respect to the reference position can be accurately calculated.

[0023]

1 shows an embodiment of an object measuring device according to the present invention. This object measuring device measures the position of an object 1 made of a work or the like conveyed by a conveyance line of an automobile, and as indicated by a virtual line, a reference coordinate system A provided at a preset processing position. An input means 2 for inputting three-dimensional data of an edge line of the object 1 on the assumption that the object 1 is accurately conveyed to the reference position of
Based on the object detection means 3 for detecting the edge line of the object 1'actually conveyed to the processing position as shown by the solid line, the input data of the input means 2 and the detection data of the object detection means 3, the reference Actual object 1'for position
Has a measuring means 4 for measuring in what state it is installed.

The input means 2 comprises a keyboard or the like for inputting data in which the edge line of the object 1 consisting of a curve or a straight line is expressed as a three-dimensional curve, and is installed at the reference position based on the design data of the object 1 or the like. The edge line of the formed object 1 is configured to be input as a three-dimensional curve defined by a preset parameter. That is, when the above parameter is set to α, the three-dimensional curve of the edge line is represented as x coordinate fx (α), y coordinate fy (α), and z coordinate fz (α). Become.

The object detecting means 3 irradiates the object 1'conveyed to the measurement position, which is the machining position of the work, with a plurality of slit lights at a predetermined interval, and the irradiation means 5 irradiates the object 1 '. And a camera 6 for capturing an image of the object 1 ', which is an image of a point of interest formed by an intersection of the slit light emitted by the irradiating means 5 and an edge line of the object 1'. Thus, the change in luminance occurring at the intersection is detected and the image data is output to the measuring means 4.

In the measuring means 4, as shown in FIG.
Based on the input data of the input means 2, a first data extracting means 7 for setting at least 6 points of interest at arbitrary positions on the edge line of the object 1 and extracting the reference coordinate data thereof, and the object detecting means. A second data extracting means 8 for extracting detected coordinate data of 6 points of interest located on the edge line based on the detection data of 3, and a reference of each point of interest extracted by the first data extracting means 7. Position conversion means 9 for converting the position of the coordinate data using a position conversion parameter described later is provided.

In the measuring means 4, the point of interest after the position conversion is performed from the detection viewpoint of the object detecting means 3, that is, the focus of the camera 6, based on the coordinate data of the position of interest converted by the position converting means 9. The first vector setting means 10 for setting a line-of-sight direction vector toward the eye and the detected point of the object detection means 3 based on the detected coordinate data of the noted points extracted by the second data extraction means 7 The second vector setting means 11 for setting the line-of-sight direction vector toward the eye and the constraint condition that the outer product of both the line-of-sight direction vectors set by the first and second vector setting means 10, 11 is 0. An equation setting means 12 for setting, and a calculation means 13 for calculating the amount of displacement and the displacement angle of the object with respect to the reference position by solving this equation. It is provided.

An object measuring method for measuring the amount of displacement and the displacement angle of the object conveyed to the processing position with respect to the reference position by using the object measuring device having the above-mentioned structure will be described with reference to FIG. First, assuming that the object 1 such as an automobile hood is at a reference position of a reference coordinate system A set in advance, the data of the edge line formed of the outer peripheral line of the object 1 is used as a parameter in the reference coordinate system A. It is input by the input means 2 as the data of the three-dimensional curve defined by α.

Next, based on the data of the three-dimensional curve, the reference coordinate data of the six points of interest a to f, which are set at arbitrary positions on the edge line of the object 1, are set as the first reference data.
After being extracted by the data extracting means 7, each of the extracted reference coordinate data is indicated by a broken line in the position converting means 9 using position conversion parameters (X, Y, Z, φ, θ, ψ) of unknown numbers. Convert to position. Then, based on the respective reference coordinate data after the position conversion, the x-direction component and the y-direction of the line-of-sight direction vector T extending from the detection viewpoint B of the object detecting means 3 toward the respective attention points a to f after the position conversion. The component and the z-direction component are sequentially set by the first vector setting means 10.

After detecting the intersections a'to f'of the edge line of the object 1'conveyed to the measuring position and the slit light by the object detecting means 3, the edge line is detected based on the detection data. The detected coordinate data of the 6 points of interest set above are extracted by the second data extracting means 7. Then, based on the detected coordinate data, the x-direction component, the y-direction component and the z-direction component of the line-of-sight direction vector V extending from the detection viewpoint B of the object detection means 3 toward the points of interest a ′ to f ′ are described above. The second vector setting means 11 sets each.

Next, the first line-of-sight direction vector T set by the first and second vector setting means 10 and 11 described above.
Based on the constraint condition (V × T = 0) that the outer product with the second line-of-sight direction vector V becomes 0, two simultaneous equations (f ₁ = vy · tz−vz · ty = 0 per point) , F ₂ =
(vz · ty−vx · tz = 0) is set, and this is executed for each target point, thereby setting a total of 12 simultaneous equations in the equation setting means 12.

After that, the simultaneous equations are solved by the calculating means 13 to set the unknown parameter parameter α and the unknown parameter position conversion parameter (X, Y, Z, φ, θ, ψ) set for each target point. To calculate. The position conversion parameters X, Y, and Z are the position of the object 1 assumed to be at the reference position and the actually detected object 1 ', respectively.
, Φ, θ, ψ are values indicating displacement angles between the installation angle of the object 1 which is assumed to be at the reference position and the installation angle of the object 1 ′ actually detected. By calculating these values (X, Y, Z, φ, θ, ψ), it is possible to accurately measure the installation state of the object 1 transported to the measurement position.

That is, the reference coordinate data (x, y, z) of the target point a located on the edge line of the object 1 assumed to be at the reference position is the coordinate data x on the x axis.
Is a parameter, the value is extracted as a value shown in the following Expression 1.

[0034]

[Equation 1]

Next, the reference coordinate data (x, y, z) is subjected to position conversion using the position conversion parameters (X, Y, Z, φ, θ, ψ), and after the conversion shown in the following equation 2, The coordinate data (x ', y', z ') of is obtained.

[0036]

[Equation 2]

Then, the coordinate data (x ',
y ′, z ′) and the data (c ₁ , c ₂ , c ₃ ) indicating the focus coordinates of the camera 6 with respect to the basic coordinate system A, from the detection viewpoint B of the detection means 3 after the conversion. Point a
The x-direction component t ₁ , the y-direction component t _2, and the z-direction component t ₃ of the line-of-sight direction vector T directed toward are set to the values shown in the following Expression ₃ .

[0038]

(Equation 3)

The detected coordinate data (u, v) extracted from the pixel coordinates of the point of interest a ′ located on the edge line of the object 1 ′ arranged at the measurement position and the calibration parameter of the camera 6 ( c _{11 to} c ₃₃ ), the x-coordinate v ₁ , the y-direction component v ₂ and the z-direction component v ₃ of the line-of-sight direction vector V from the detection viewpoint B toward the point of interest a ′ are the values shown in the following formula 4. Is set to. The calibration parameters (c _{11 to} c ₃₃ ) are defined as a unique relationship between the point of interest a and the pixels of the camera 6 depending on the lens center of the camera 6, the direction of its optical axis, the angle of view, and the magnification. It is a given value.

[0040]

[Equation 4]

Taking the outer product of the line-of-sight direction vector V shown in the above equation 4 and the line-of-sight direction vector T shown in the above equation 3 and setting the value to 0, the following equation is obtained.

F ₁ = v ₂ t ₃ −v ₃ t ₂ = v ₂ {−SθCψ · x + SθSψ · f (x) + Cθ · g (x) + Z−c ₃ } −v ₃ {(SφCθCψ + CφSψ) x− (SφCθSψ -CφCψ) f (x) + Y -c 2} = 0 f 2 = v 3 t 1 -v 1 t 3 = v 3 {(CφCθCψ-SφSψ) x + (CφCθSψ + SφCψ) f (x) + CφSθ · g (x) + X _{-c 1} -v 1 {-SθCψ ·} x + SθCψ · f (x) + Cθ · g (x) + Z-c 3} = 0 then, the equation is set for each target point, the nonlinear equations numerical solution, e.g. By solving by the Newton-Raphson method or the like, the unknown parameter x set for each target point and the unknown position conversion parameter (X,
Y, Z, φ, θ, ψ) is calculated. This Newton-Raphson method uses a specific approximation Xi,
After Yi ... is set, and this is applied to the functions f ₁ , f ₂ ... And the partial differential value thereof, the calculation process of obtaining approximate values Xi + 1, Yi + 1 ... which are closer to the solution of the unknown number is repeated. , Is a numerical solution method that uses the value that has reached convergence as the solution. In the Newton-Raphson method, the partial differential values of the functions f ₁ and f ₂ are as follows.

∂f ₁ / ∂X = 0 ∂f ₁ / ∂Y = −v ₃ ∂f ₁ / ∂Z = v ₂ ∂f ₁ / ∂φ = −v ₃ {(CφCθCψ−SφSψ) x + (CφCθSψ + SφCψ ) F (x) + CφSθ ・ g (x)} ∂f ₁ / ∂θ = v ₂ {-CθCψ ・ x + CθSψ ・ f (x) -Sθ ・ g (x)} -v ₃ {-SφSθCψ ・ x + SφSθSψ ・ f ( x) + SφCθ · g (x)} ∂f ₁ / ∂ψ = v ₂ {SθSψ · x + SθCψ · f (x)} − v ₃ {(−SφCθSψ + CφCψ) x− (SφCθCψ + CφSψ) f (x)} ∂f _{1 / ∂x = v 2 {-SθCψ +} SθSψ · f'(x) + Cθ · g'(x)} -v 3 {SφCθCψ + CφSψ- (SφCθSψ-CφSψ) · f'(x) + SφSθ · g'(x)} ∂f ₂ / ∂X = V ₃ ∂f ₂ / ∂Y = 0 ∂f ₂ / ∂Z = −v ₁ ∂f ₂ / ∂φ = v ₃ {(−SφCθCψ−CφSψ) x − (− SφCθSψ + CφCψ) f ( x) − φSθ · g (x)} ∂f 2 / ∂θ = v 3 {-CφSθCψ · x + CφSθSψ · f (x) + CφCθ · g (x)} - v 1 {-CφCψ · x + CθSψ · f (x) -Sθ · g (x)} ∂f ₂ / ∂ψ = v ₃ {(−CφCθSψ−SφCψ) x− (CφCθCψ−SφSψ) f (x)} − v ₁ {SθSψ · x + SθCψ · f (x)} ∂f ₂ / ∂ x = v ₃ {CφCθCψ-SφSψ− (CφCθSψ + SθCψ) ・ f ′ (x) + CφSθ · g ′ (x)} − v ₁ {−SθCψ + SθSψ · f ′ (x) + Cθ · g ′ (x)} of the above nonlinear equation In the numerical solution, the object detecting means 3
From the detection data of the edge line of the object 1 ′ at the measurement position detected by the virtual data on the assumption that the edge line of the object 1 at the reference position is detected by the object detecting means 1, An unknown number position conversion parameter (X, Y, Z, φ, used to solve the above simultaneous equations based on the position and shape of 1 ′ is investigated.
Estimate and set the initial values of θ, ψ). That is, in the iterative calculation by the Newton-Rapson method, the calculation is started by giving an appropriate initial value as an approximate value of the unknown number, and thus the closer the initial value to the true value is, the accuracy of the calculation and the calculation time. It is advantageous in terms of shortening. Therefore, it is desirable to roughly check the positional deviation amount and the like as described above and set the value as the initial value of the approximate value.

The edge line of the object 1 which is assumed to be at the reference position as described above is input as the data of the three-dimensional curve defined by the parameter, and the attention point corresponding to the number of unknowns is input on the edge line. , The reference coordinate data of this point of interest is extracted according to the input data, and then the reference coordinate data of this point of interest is used with unknown position conversion parameters (X, Y, Z, φ, θ, ψ). Then, the position is converted, and the line-of-sight direction vector T toward the point of interest after the position conversion is set, and the edge line of the object 1'arranged at the measurement position is detected by the object detecting means 3 to obtain the detection data. Since the detection coordinate data of the point of interest located on the edge line is extracted based on the configuration, the line-of-sight direction vector V toward the point of interest is set, so that the object to be measured Even when a feature point cannot be found because there is no corner portion at, the equation is set based on the line-of-sight direction vectors T and V, and the installation state of the object 1 is obtained by solving the equation. be able to.

That is, the above-mentioned bidirectional gaze direction vectors T, V
By setting the equation with the constraint that the outer product becomes 0 when the directions of are the same, and solving the equation to calculate the position conversion parameters (X, Y, Z, φ, θ, ψ) , An unknown number required to overlap the actual position of the object 1 by shifting the installation angle of the object 1 at the reference position and changing the installation angle, and based on this value the position of the object 1 with respect to the reference position The amount of deviation and the displacement angle can be obtained quickly and accurately.

Therefore, for example, regardless of the shape of the object 1 transported to the processing position in an automobile assembly line, the installation state of the object 1 transported to the processing position is accurately calculated and the installation state is assembled. By teaching the use robot, it is possible to accurately execute the work of automatically assembling the parts with respect to the object 1.

It should be noted that the number of points of interest is equal to the number of position conversion parameters (X, Y, Z, φ, θ, ψ) of the unknowns, that is, it is sufficient to set only 6 points, but it is more than 6 points. You may set many. In this case, the measurement accuracy can be improved by obtaining the position conversion parameter using the least squares method.

The irradiation means 5 for irradiating the edge line of the object 1'arranged at the measurement position with the slit light as described above and the image of the point of interest on the edge line irradiated by the irradiation means 5 are displayed. In the case where the object detecting means 3 including the camera 6 for photographing is provided, and the camera 6 photographs the change in luminance generated at the intersection of the slit light irradiated by the irradiating means 5 and the edge line, It is possible to easily and accurately extract the coordinates of the intersections according to the captured image.

In place of the object detecting means 3 including the irradiating means 5 and the camera 6, the edge line of the object is detected by irradiating the object with ultrasonic waves and detecting the reflection state thereof, or the object and its surroundings. The edge line of the object may be detected by detecting the temperature difference between

In the numerical solution of the non-linear equation, the edge line detection data of the object 1'at the measurement position detected by the object detecting means 3 and the edge line of the object 1 assumed to be at the reference position are calculated. Object detection means 3
Based on the virtual data obtained when detected by, the initial values of the position conversion parameters (X, Y, Z, φ, θ, ψ) of unknowns used for solving the simultaneous equations are estimated and set. According to the configuration described above, it is possible to prevent the above numerical calculation from being performed based on the initial value set at a position far from the actual object, so it is possible to effectively improve the calculation accuracy and shorten the calculation time. You can

In the above embodiment, the edge line of the object 1'is collectively detected by the single camera 6 and the detection viewpoint B of the object detecting means 3 is set to one position. May be set at a plurality of positions, and the edge lines of the object 1 ′ may be detected from multiple directions. For example, as shown in FIG. 4, a plurality of cameras 6 are arranged on the edge line of the object 1 ′, and the plurality of cameras 6 are configured to simultaneously detect the corresponding edge lines,
Alternatively, a moving means for moving the camera 6 may be provided, and the installation position of one camera 6 may be moved so that each edge line is sequentially photographed from multiple directions.

When the edge lines of the object 1 are detected from multiple directions as described above, the edge lines of the object 1'are collectively photographed by one camera 6. As described above, the image of the object 1 ′ becomes relatively small, or the object 1 ′ in the axial direction of the camera 6
It is possible to accurately recognize the three-dimensional shape of the object 1 ′ and appropriately set the line-of-sight direction vector V without causing a problem such that the displacement of No. 1 is not accurately detected.

Further, as shown in FIG. 5, an end plate 14 is installed on the side of the object 1 ′, and the edge line of the object 1 ′ and the end plate 14 are detected by a single object detecting means 3 composed of one camera 6. It is also possible to simultaneously detect the mirror image 1 ″ shown in FIG. 2 and extract the detected coordinate data of the target point set at an arbitrary position on the edge line based on the detected data. In the case of the above configuration, the displacement of the object 1'with respect to the axial direction of the camera 6 can be accurately detected with a simple configuration, and the three-dimensional shape of the object 1'can be properly grasped.

Next, an object 1 having a circular edge line
A specific example of the object measuring method for measuring the installation state of the 'will be described. First, as shown in FIG. 6, while the circular edge line is located on the xy plane of the reference coordinate system A, the reference coordinate system A is set so that the center of gravity of the circle is located on the z axis. The edge line data is input as a three-dimensional curve defined by using the angle α with the x-axis as a parameter.

If the edge line is a perfect circle, it is not necessary to obtain the displacement angle in the circumferential direction, so the position conversion parameter ψ is set to 0 and the number of points of interest is set to 5. Then, the attention point a set on the edge line
The reference coordinate data (x, y, z) is extracted as a value shown in the following Expression 5 based on the data of the three-dimensional curve. In the following Expression 5, r is the radius of the circle.

[0056]

(Equation 5)

Then, the reference coordinate data (x, y,
z) is subjected to position conversion using the position conversion parameters (X, Y, Z, φ, θ) to obtain the coordinate data (x ′, y ′, z ′) after conversion shown in Equation 6 below. Will be done.

[0058]

(Equation 6)

Then, the coordinate data (x
′, Y ′, z ′) and coordinate data (c ₁ , c ₂ , c ₃ ) of the focus of the camera 6 with respect to the basic coordinate system A, from the detection viewpoint B of the detection means 3 after the conversion. The x-direction component t ₁ , the y-direction component t _2, and the z-direction component t ₃ of the line-of-sight direction vector T toward the target point are set to the values shown in the following Expression 7.

[0060]

(Equation 7)

Further, the error of the object 1'disposed at the measurement position
Extracted from the pixel coordinates of the point of interest a'located on the edge line
Detected coordinate data (u, v) and the carry of the camera 6
Blation parameter (c₁₁~ C₃₃) And based on
Line-of-sight vector from the detection viewpoint B to the point of interest a '
X-direction component v of V₁, Y-direction component v₂And z-direction component v ₃
Is set to the value shown in Equation 8 below.

[0062]

(Equation 8)

If the outer product of the line-of-sight direction vector V shown in the above equation 8 and the line-of-sight direction vector T shown in the above equation 7 is taken and the value is set to 0, the following equation is obtained.

F ₁ = v ₂ t ₃ −v ₃ t ₂ = v ₂ (−r · SθCα + Z−c ₁ ) −v ₃ (r · SφCθCα + r · CφSα + Y−c ₂ ) = 0 f ₂ = v ₃ t _{1 -v 1 t 3 = v 3 (} r · CφCθCα-r · SφSα + X-c 1) -v 1 (-r · SθCα + Z-c 3) = 0 and is set by the number corresponding to the number corresponding to the unknown By setting the above equation for each attention point and solving this equation by Newton-Raphson method,
The parameter α of each equation and the position conversion parameters (X, Y, Z, φ, θ) of the unknowns are calculated. In the Newton-Raphson method, the partial differential values of the functions f ₁ and f ₂ are as follows.

∂f ₁ / ∂φ = −v ₃ (r · CφCθCα−r · SφSα) ∂f ₁ / ∂θ = v ₂ (−r · CθCα) −v ₃ (−r · SφSθCα) ∂f ₁ / ∂X = 0 ∂f ₁ / ∂Y = −v ₃ ∂f ₁ / ∂Z = v ₂ ∂f ₁ / ∂α = v ₂ (r · SθSα) −v ₃ (−r · SφCθSα + r · CφCα) ∂f ₂ / ∂φ = v ₃ (−r · SφSθCα−r · CφSα) ∂f ₂ / ∂θ = v ₂ (−r · CφSθCα) −v ₁ (−r · CθCα) ∂f ₂ / ∂X = v ₃ ∂f ₂ / ∂Y = 0 0 ∂f ₂ / ∂Z = -v ₁ ∂f ₂ / ∂α = v ₃ (-r ・ CφCθSα-r ・ SφCα) -v ₁ (r ・ SθSα) Circular shape as above After the position conversion of the object 1 having the edge lines of (3) using the position conversion parameters (X, Y, Z, φ, θ) of unknowns, the number of points of interest corresponding to the number of unknowns are set on the edge lines. Responsibility And the unknown is solved by a numerical solution of a nonlinear equation such as the Newton-Raphson method, whereby the object 1'is installed in any state with respect to the reference position consisting of the origin of the reference coordinate system A. Since it is configured to measure whether
Even when it is difficult to detect the entire edge line, the installation state can be accurately and quickly measured.

That is, as a method of measuring a circular edge line, it is possible to three-dimensionally view the center of gravity of a circle and recognize its position, but this method cannot detect the entire edge line. Since it is difficult to accurately recognize the position of the center of gravity, the applicable range is limited. However, the measurement is performed based on the coordinates of the point of interest set on the circular edge line as described above. In such a case, since the point of interest can be set at an arbitrary position on the edge line, the object 1
The amount of displacement and the displacement angle of ′ with respect to the reference position can be accurately and quickly obtained.

Therefore, even when an object having a circular edge line such as an oil injection hole or a bottle hole is moved from the reference position to the processing position of the assembly line, the oil injection hole, the bolt hole, etc. It is possible to accurately recognize the position of [1] and properly perform automatic oil injection work, automatic bolt tightening work, and the like.

Further, when the reference coordinate system A is set so that the circular edge line is located on the xy plane of the reference coordinate system A as described above, the edge which is assumed to be at this reference position Since the three-dimensional curve representing the line can be defined by a simple function, the unknown value can be easily obtained. Especially when the center of gravity of the circle is arranged on the z-axis of the reference coordinate system A, the three-dimensional curve representing the edge line of the basic coordinate system A can be defined by a simpler function. , The unknown value can be calculated more easily.

If the barycentric position of the circle can be accurately obtained, the barycentric position of the circle is set as the target point in addition to the target point set on the circular edge line, and the object detecting means is set. The line-of-sight direction vector from the detection viewpoint B of 3 toward the center of gravity of the circle may be set. In such a configuration, the object 1 ′ based on the line-of-sight direction vector set toward the point of interest set on the edge line and the line-of-sight direction vector set toward the point of interest formed of the center of gravity of the circle. The installation state of can be measured more accurately.

In place of the above Newton-Raphson,
The above simultaneous equations may be solved by using a numerical solution method of a non-linear equation including a bisection method, a scissors method and a secant method.

In the above embodiment, the object measuring apparatus and the object measuring method according to the present invention are used as the position recognizing means of the work conveyed in the automobile assembly line. However, it is used in a factory or the like. The object measuring device and the object measuring method described above can also be used in order to make the automatic traveling robot for transporting the parts recognize the traveling position. That is, by providing an object made of a sign having a three-dimensional shape at a preset reference position in a factory or the like, and detecting the sign object by the object detection means and recognizing the position, the relative movement of the automatic traveling robot The position may be recognized.

Further, as described above, the first data extracting means 7
Object 1 assumed to be at the reference position extracted by
On the edge line of the object 1 ′ at the measurement position extracted by the second data extracting means 8 instead of the structure in which the position converting means 9 performs position conversion of the reference coordinate data of the point of interest located on the edge line The detected coordinate data of the point of interest located at may be converted by the position conversion means using the position conversion parameter.

[0073]

As described above, the invention according to claim 1 has an input means for inputting an edge line of an object assumed to be at a reference position as data of a three-dimensional curve defined by a parameter. First data extracting means for extracting reference coordinate data of the point of interest set on the edge line according to the input data, and detected coordinates of the point of interest provided on the edge line of the object arranged at the measurement position. A second extracting means for extracting data is provided, and a position converting means for performing position conversion of one of the reference coordinate data and the detected coordinate data by using an unknown number of position converting parameters is provided, and the unknown position converting parameters are calculated. Since the amount of displacement and the displacement angle with respect to the reference position of the object are calculated by means of the means, there is no corner portion in the object. If it is impossible to find a feature points may also be measuring the installed state by numerical calculation to accurately recognize the object three-dimensional shape.

Therefore, for example, regardless of the shape of the object conveyed to the processing position in an automobile assembly line, the installation state of the object conveyed to the processing position is accurately obtained by calculation, and this installation state is determined by the assembling robot. It is possible to accurately perform the work of automatically assembling the parts with respect to the above objects by teaching the object, and to make the automatic traveling robot for parts transportation used in the factory etc. accurately recognize the traveling position. There is an advantage that the automatic traveling robot can accurately travel.

According to the second aspect of the present invention, the irradiation means for irradiating the edge line of the object arranged at the measurement position with the slit light and the image of the point of interest on the edge line irradiated by the irradiation means are photographed. Since the object detection means composed of a camera is provided and the change in luminance caused at the intersection of the slit light irradiated by the irradiation means and the edge line is photographed by the camera, the intersection is determined according to the photographed image. The coordinates of can be extracted easily and accurately.

In the invention according to claim 3, the edge line of the object assumed to be at the reference position is input as the data of the three-dimensional curve defined by the parameter, and the attention point set on the edge line is After extracting the reference coordinate data according to the input data, the edge line of the object arranged at the measurement position is detected by the object detecting means, and the detected coordinate data of the point of interest located on the edge line is extracted. , One of the detected coordinate data of the point of interest and the reference coordinate data is position-converted using the position conversion parameter of the unknown number, and the position conversion parameter of the unknown number is moved to the reference position of the object by a numerical solution method of a non-linear equation. Since it is configured to calculate the amount of displacement and the displacement angle, it is possible to accurately recognize the three-dimensional shape of the object regardless of the shape of the object. There is an advantage that an installation state of an object can be accurately and quickly measured.

According to a fourth aspect of the present invention, in the numerical solution of the non-linear equation, the edge line detection data of the object at the measurement position detected by the object detecting means and the edge of the object assumed to be at the reference position are detected. Based on the virtual data obtained when the line is detected by the object detecting means, since it is configured to estimate and set the initial value of the unknown number used to solve the simultaneous equations,
It is possible to prevent the occurrence of a situation in which unnecessary calculation is performed due to the above-described numerical calculation being performed based on the initial value set at a position far from the actual object, and improve the calculation accuracy. At the same time, there is an advantage that the calculation time can be shortened.

In the invention according to claim 5, the object detecting means for detecting the edge line of the object arranged at the measurement position is set to a plurality of detection viewpoints, and the edge line of the object is detected from multiple directions. Therefore, the image of the object becomes relatively small as in the case where the edge lines of the object are collectively captured by one camera, and the displacement of the object in the axial direction of the camera cannot be accurately detected. It is possible to properly recognize the three-dimensional shape of the object and appropriately set the line-of-sight direction vector used for solving the unknown number without causing a problem such as a problem.
Therefore, the amount of displacement and the displacement angle of the object with respect to the reference position can be measured more accurately.

According to the sixth aspect of the invention, since the edge lines of the object arranged at the measurement position are respectively detected by the plurality of object detecting means, the above-mentioned edge is detected based on the detection data of the object detecting means. By extracting the detected coordinate data of the point of interest set at any position on the line, the amount of displacement of the object with respect to the axial direction of the object detection means can be detected accurately and quickly, and the three-dimensional shape of the object can be properly grasped. There is an advantage that you can.

In the invention according to claim 7, an end plate is installed at a position corresponding to the edge line of the object, and the edge line of the object and the mirror image projected on the end plate are simultaneously detected by the object detecting means, and the detection data thereof is detected. Based on the above, since it is configured to extract the detected coordinate data of the point of interest set at any position on the edge line, the displacement of the object with respect to the axial direction of the camera is accurately detected by a single detection means, It is possible to properly grasp the three-dimensional shape of.

The invention according to claim 8 is x- of a reference coordinate system which constitutes a reference position of an object having a circular edge line.
Since the reference coordinate system is set so that the edge line is located on the y plane, a three-dimensional curve representing the edge line assumed to be located at the reference position can be defined by a simple function. The unknown value can be easily done. Therefore, even when an object having a circular edge line such as an oil injection hole or a bottle hole is transferred from the reference position to the processing position of the assembly line, the positions of the oil injection hole and the bolt hole, etc. should be maintained. There is an advantage that the automatic oil injection work and the automatic bolt tightening work can be accurately performed by accurately recognizing.

In addition to the point of interest set on the circular edge line, the position of the center of gravity of the circle is set as the point of interest, and the line of sight from the detection viewpoint of the object detecting means toward the center of gravity of the circle. Since it is configured to set the direction vector, the point of interest set on the edge line, and the installation state of the object based on both the line-of-sight direction vector set toward the point of interest consisting of the center of gravity of the circle, It is possible to measure more accurately.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing an overall configuration of an object measuring device according to the present invention.

FIG. 2 is a block diagram showing a configuration of a measuring unit provided in the object measuring device.

FIG. 3 is an explanatory diagram showing an embodiment of an object measuring method according to the present invention.

FIG. 4 is an explanatory diagram showing another embodiment of the object measuring method according to the present invention.

FIG. 5 is an explanatory view showing still another embodiment of the object measuring method according to the present invention.

FIG. 6 is an explanatory view showing still another embodiment of the object measuring method according to the present invention.

[Explanation of symbols]

 1 Object 2 Input Means 3 Object Detecting Means 4 Measuring Means 5 Irradiating Means 6 Cameras 7 First Data Extracting Means 8 Second Data Extracting Means 9 Position Converting Means 10 First Vector Setting Means 11 Second Vector Setting Means 12 Equation Setting Means 13 Calculation means 14 End plate

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location G06T 7/00

Claims (9)

[Claims] 1. Input means for inputting an edge line of the object as data of a three-dimensional curve defined by a parameter, assuming that an object is present at a preset reference position.
First data extracting means for extracting the reference coordinate data of the point of interest located on the edge line according to the input data, and object detecting means for detecting the edge line of the object arranged at the measurement position, Second data extracting means for extracting the detected coordinate data of the point of interest set on the edge line based on the detected data of the object detecting means, and both coordinate data extracted by the first and second data extracting means. Position conversion means for converting one of the positions using a position conversion parameter consisting of a positional displacement amount and an unknown number of displacement angles, and a detection viewpoint of the object detection means based on the coordinate data of the point of interest converted by this position conversion means. First to set the line-of-sight direction vector from the position to the point of interest after position conversion
Vector setting means, second vector setting means for setting the line-of-sight direction vector from the detection viewpoint of the object detection means to the point of interest on the edge line based on the other of the coordinate data, and the first and second vectors. Equation setting means for setting an equation under the constraint that the directions of both line-of-sight direction vectors set by the setting means coincide with each other, and by solving this equation, the unknown amount of the positional deviation amount and displacement angle with respect to the reference position of the object An object measuring device, comprising: a calculating means for calculating. 2. Object detection comprising irradiation means for irradiating an edge line of an object arranged at a measurement position with slit light, and a camera for photographing an image of a point of interest on the edge line irradiated by this irradiation means. The object measuring device according to claim 1, further comprising means. 3. An edge line of an object is represented as a three-dimensional curve defined by a parameter based on data set assuming that the object exists at a preset reference position, and the edge line is then represented. While extracting the reference coordinate data of the point of interest located above, the edge line of the object arranged at the measurement position is detected by the object detection means,
After extracting the detected coordinate data of the point of interest located on the edge line based on this detected data, one of the detected coordinate data of the point of interest and the reference coordinate data is set to a position composed of an unknown amount of displacement and displacement angle. Position conversion is performed using a conversion parameter, and a line-of-sight direction vector from the detection viewpoint of the object detecting means to the target point after position conversion is set based on the coordinate data after the position conversion, and the other of the coordinate data is the other. After setting the line-of-sight direction vector from the detection viewpoint of the object detecting means to the point of interest on the edge line based on the above, set the equation as a constraint condition that the directions of these line-of-sight direction vectors match, and set this equation It is characterized in that it is configured to calculate an unknown amount of displacement and displacement angle with respect to the reference position of the object by solving Body measurement method. 4. The virtual data obtained when the detection data of the edge line of the object at the measurement position detected by the object detection means and the edge line of the object assumed to be at the reference position are detected by the object detection means. The object measuring method according to claim 3, wherein an initial value of an unknown number used in a numerical solution of the nonlinear equation is set based on the data. 5. An object detecting means for detecting an edge line of an object arranged at a measurement position is set at a plurality of detection viewpoints, and the edge line of the object is detected from multiple directions. The object measuring method according to claim 3. 6. An edge line of an object arranged at a measurement position is detected by each of a plurality of object detecting means, and detected coordinate data of a point of interest located on the edge line is extracted based on the detected data. The object measuring method according to claim 3, wherein the object measuring method is configured. 7. A mirror plate which projects an edge line of an object arranged at a measurement position is installed, the edge line of the object and its mirror image are detected by an object detecting means, and the position on the edge line is detected based on the detection data. The object measuring method according to claim 3, wherein the detected coordinate data of the target point is extracted. 8. The reference coordinate system of an object is set such that the edge line is located on an xy plane of a coordinate system that constitutes a reference position of an object having a circular edge line. 3. The object measuring method described in 3. 9. In addition to the point of interest on the edge line, the center of gravity of the circle forming the edge line is set as the point of interest, and a line-of-sight direction vector from the detection viewpoint of the detection means to the center of gravity of the circle is set. The object measuring method according to claim 8, wherein the object measuring method is configured.