JP6282157B2 - Shape recognition device - Google Patents

Description

  The present invention relates to a shape recognition device used for evaluating the shape of the surface of an object.

  Distortion (surface distortion) may occur on the surface of a vehicle door panel or the like. As an apparatus used for quantitatively evaluating the generated distortion, the applicant of the present application disclosed in Patent Document 1 a shape recognition apparatus that recognizes the shape of the surface of an object (evaluation object) to be evaluated. is suggesting.

  In this shape recognition device, three-dimensional shape data representing the surface of the evaluation object is acquired, base shape data consisting of a curvature group that roughly reflects the shape of the surface of the evaluation object from the three-dimensional shape data, and the surface of the surface Detailed shape data including a group of curvatures reflecting the shape in detail is generated. Then, with reference to the base shape data, a deviation amount between the base shape data and the detailed shape data in the same region is generated as a distortion value. Based on this strain value, the surface strain of the evaluation object can be quantitatively evaluated.

JP 2010-210576 A

  However, when a large distortion over a wide range occurs on the surface of the evaluation object, the base shape data becomes data including the large distortion, and in the evaluation based on the distortion value generated from the base shape data and the detailed shape data, In some cases, the surface distortion of the evaluation object cannot be recognized.

  An object of the present invention is to provide a shape recognition device that can generate a distortion value that can more accurately recognize a distortion of the surface of an object.

  In order to achieve the above object, a shape recognition apparatus according to the present invention generates a base shape data composed of a group of curvatures reflecting a relatively rough shape from three-dimensional shape data representing the shape of the surface of an object. On the design of the shape data generation means, the detailed shape data generation means for generating detailed shape data consisting of a curvature group reflecting the shape of the surface of the object relatively in detail from the three-dimensional shape data, Design shape data acquisition means for acquiring design shape data comprising a group of curvatures, and a predetermined region on the surface of the object, acquired from the design shape data within an allowable range of a predetermined width based on the curvature acquired from the base shape data. When curvature is included, the difference between the curvature obtained from the base shape data and the curvature obtained from the detailed shape data is generated as a distortion value, and the design shape is within the allowable range. Not contain a curvature that is obtained from the chromatography data, and a strain value generating means for generating a difference between the curvature obtained from the curvature and detailed shape data obtained from the design shape data as distortion values.

  According to this configuration, from the three-dimensional shape data representing the shape of the surface of the object, from the base shape data composed of a curvature group reflecting the shape relatively coarsely, and the curvature group reflecting the shape relatively in detail The detailed shape data is generated. On the other hand, design shape data including a design curvature group on the surface of the object is acquired. Then, the base shape data is compared with the design shape data for a predetermined region on the surface of the object, and the curvature obtained from the design shape data is allowed to have a predetermined width based on the curvature obtained from the base shape data. It is determined whether it falls within the range.

  When the curvature acquired from the design shape data is included in the allowable range, a difference between the curvature acquired from the base shape data and the curvature acquired from the detailed shape data is generated as a distortion value. Thereby, similarly to the conventional shape recognition apparatus, the distortion of the surface of the object can be quantitatively evaluated based on the generated distortion value.

  On the other hand, when a large distortion over a wide range occurs on the surface of the object, the large distortion is included in the base shape data, and the base shape data greatly deviates from the design shape data. In this case, when the difference between the curvature acquired from the base shape data and the curvature acquired from the detailed shape data is generated as the distortion value, the distortion value does not reflect a large distortion generated on the surface of the object. Therefore, based on the strain value, the surface distortion of the object cannot be quantitatively evaluated.

  Therefore, when the curvature acquired from the design shape data is not included in the allowable range of the predetermined width based on the curvature acquired from the base shape data, the curvature acquired from the design shape data and the detailed shape data are acquired. The difference from the calculated curvature is generated as the distortion value. Thereby, since the distortion over a wide range of the surface of the object is reflected in the distortion value, the distortion of the surface of the object can be more accurately and quantitatively evaluated based on the distortion value.

  ADVANTAGE OF THE INVENTION According to this invention, the distortion value which can recognize the distortion of the surface of an object more accurately can be produced | generated.

It is a figure which shows the external appearance structure of the shape evaluation system with which the shape recognition apparatus which concerns on one Embodiment of this invention was integrated. It is a block diagram which shows the electric constitution of a shape evaluation system. It is a perspective view which shows an example of the object which has an evaluation object surface, and shows X-axis, Y-axis, and Z-axis of a three-dimensional orthogonal coordinate system collectively. It is the figure which looked at the object shown by FIG. 3 from the X-axis direction. It is a figure for demonstrating extraction of the three-dimensional data by a base shape data generation part. It is a figure which shows notionally that the value of Z coordinate corresponding to each value of Y coordinate in three-dimensional data becomes a mutually different value when the evaluation object surface is curved. It is a figure for demonstrating the curvature production | generation method by a curvature production | generation part. It is a flowchart which shows the flow of the distortion value production | generation process which a distortion value production | generation part performs. It is a figure which shows an example of the curve which base shape data, detailed shape data, and design shape data represent. It is a figure which shows the other example of the curve which base shape data, detailed shape data, and design shape data represent.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

  FIG. 1 is a diagram showing an external configuration of a shape evaluation system 1 in which a shape recognition device according to an embodiment of the present invention is incorporated.

  The shape evaluation system 1 is a system for evaluating the shape (distortion) of the surface (hereinafter referred to as “evaluation target surface”) A of an object. The shape evaluation system 1 shoots an evaluation target surface A and generates a camera 2 that generates image data of the surface, a robot 3 that holds the camera 2, and a shooting control device 4 that controls the operations of the camera 2 and the robot 3. And an image data processing device 5 that generates 3D image data of the evaluation target surface A from the image data generated by the camera 2, and the evaluation target surface based on the 3D image data generated by the image data processing device 5. And a shape recognition device 6 that outputs an evaluation image corresponding to the shape of A to the display D.

  The camera 2 is composed of a camera including a solid-state image sensor such as a CMOS image sensor.

  The robot 3 includes an articulated arm 7, and the camera 2 is held at the tip of the articulated arm 7.

  FIG. 2 is a block diagram showing an electrical configuration of the shape evaluation system 1.

  The photographing control device 4, the image data processing device 5, and the shape recognition device 6 are all configured to include a microcomputer, and substantially include a plurality of function processing units that are realized by software by program processing. Note that some functions of the function processing unit may be realized by hardware such as a logic circuit.

  The imaging control device 4 includes a camera control unit 11 that controls the camera 2 and a robot control unit 12 that controls the robot 3 as function processing units. The camera 2 and the robot 3 are controlled so that the entire evaluation target surface A is photographed by the camera 2 and image data of the entire evaluation target surface A is generated.

  The image data processing device 5 represents the shape of the evaluation target surface A based on the image data acquisition unit 21 that acquires image data from the camera 2 and the image data acquired by the image data acquisition unit 21 as a function processing unit. And a three-dimensional shape data generation unit 22 that generates three-dimensional shape data. Since the processing by the three-dimensional shape data generation unit 22 is publicly known (see, for example, Japanese Patent Application Laid-Open Nos. 2002-257528 and 2004-317495), the description thereof is omitted.

  The shape recognition device 6 includes, as function processing units, a data acquisition unit 31, a data conversion unit 32, a noise removal unit 33, an excluded area setting unit 34, a base shape data generation unit 35, a detailed shape data generation unit 36, and a curvature generation unit 37. A design shape data acquisition unit 38, a distortion value generation unit 39, a measurement target shape generation unit 40, and an evaluation image generation unit 41.

  FIG. 3 is a perspective view illustrating an example of an object having the evaluation target surface A, and illustrates the X axis, the Y axis, and the Z axis of the three-dimensional orthogonal coordinate system. FIG. 4 is a view of the object shown in FIG. 3 as viewed from the X-axis direction.

  In the following, as shown in FIG. 3 and FIG. 4, an example in which the object having the evaluation target surface A is a door panel DP is taken as an example. The door panel DP is formed with a seating surface DH1 and an opening DH2 on which a door handle (not shown) is mounted. The seating surface DH1 is recessed in a substantially square shape from the periphery thereof, and the peripheral edge of the seating surface DH1 changes in shape with a large curvature in a relatively narrow region. The opening DH2 is formed in the seating surface DH1.

  The data acquisition unit 31 acquires the three-dimensional shape data generated by the image data processing device 5 (three-dimensional shape data generation unit 22).

  The data conversion unit 32 sets the three-dimensional shape data acquired by the data acquisition unit 31 in a predetermined three-dimensional orthogonal coordinate system in a grid pattern on the evaluation target surface A with a pitch of less than several millimeters. The data is converted into data representing the height at each virtual point (hereinafter, this data is referred to as “three-dimensional data”). The predetermined three-dimensional orthogonal coordinate system is a coordinate system in which the front-rear direction, the up-down direction, and the vehicle width direction of the door panel DP when the door panel DP is attached to the vehicle are set to the X-axis direction, the Y-axis direction, and the Z-axis direction, respectively. is there. The X axis, Y axis, and Z axis shown in FIG. 3 are coordinate axes in the predetermined three-dimensional orthogonal coordinate system.

  In the three-dimensional orthogonal coordinate system of the three-dimensional shape data generated by the image data processing device 5, the X-axis direction, the Y-axis direction, and the Z-axis direction do not necessarily coincide with the front-rear direction, the vertical direction, and the vehicle width direction of the door panel DP, respectively. Is not limited. Further, the three-dimensional shape data generated by the image data processing device 5 is data representing the height at each virtual point set in a grid pattern at a predetermined pitch on the evaluation target surface A. The pitch is not necessarily an appropriate value for processing by the shape recognition device 6.

  Therefore, the data conversion unit 32 converts the three-dimensional shape data acquired by the data acquisition unit 31 into three-dimensional data. At the time of this conversion, the data conversion unit 32 performs an interpolation process by a nearest neighbor method or the like as necessary.

  The noise removing unit 33 removes noise included in the three-dimensional data by filtering. For the filter process, for example, a smoothing filter such as a low-pass filter, a moving average filter, or a weighted average filter is used.

  Based on the three-dimensional data from which the noise has been removed, the excluded area setting unit 34 checks whether or not there is a site whose height is greatly changed in a relatively narrow area on the evaluation target surface A. When such a part exists, the part or the area surrounding the part is set as an excluded area. For example, since the height of the peripheral portion of the seat surface DH1 of the door panel DP changes greatly in a relatively narrow region, the peripheral portion of the seat surface DH1 or the region surrounding the peripheral portion is set as the exclusion region.

  FIG. 5 and FIG. 6 are diagrams for explaining the contents of processing by the base shape data generation unit 35.

  The base shape data generation unit 35 cuts out three-dimensional data having a set width from the entire three-dimensional data from which noise has been removed, and generates two-dimensional data reflecting the surface shape of the cut-out three-dimensional data. Base shape data composed of a curvature group acquired from two-dimensional data is generated.

  Specifically, the base shape data generation unit 35 first cuts out three-dimensional data of an area having a set width (see FIG. 5) in the Y-axis direction from the entire three-dimensional data from which noise has been removed. (Extracted).

  Then, as shown in FIG. 5, a predetermined range of a first set interval D1 (for example, about 30 mm) is set in the X-axis direction, and the predetermined range is determined by the pitch of the virtual point group (between adjacent virtual points). The three-dimensional data of the virtual point group included in the predetermined range is extracted and the moving average of the Z coordinate value of the extracted three-dimensional data is calculated while being shifted in the X-axis direction at intervals. Thereby, the value of the Z coordinate is generated at a pitch equal to the pitch of the virtual point group in the X-axis direction. At this time, if a point (exclusion point) in the exclusion region exists within the predetermined range, the three-dimensional data of the exclusion point is excluded (excluded from a distortion value generation target described later).

  As shown in FIGS. 3 and 4, when the evaluation target surface A is curved, as shown in FIG. 6, the Y coordinate values Y1, Y2,. The values of the Z coordinates corresponding to Yn are different from each other. Therefore, in the base shape data generation unit 35, next, with respect to the three-dimensional data, an angle θ formed by a normal point of the evaluation target surface A at the virtual point and the Z-axis direction corresponding to the three-dimensional data. A rotation process is performed to rotate the curve center (see FIG. 4). Thereby, the virtual point corresponding to the three-dimensional data is located on the XZ plane, the conversion of the three-dimensional data into the two-dimensional data is achieved, and the two-dimensional data reflecting the surface shape of the cut-out three-dimensional data is obtained. can get.

  The detailed shape data generation unit 36 cuts out three-dimensional data having a set width from the entire three-dimensional data from which noise has been removed, and generates two-dimensional data reflecting the surface shape of the cut-out three-dimensional data. Detailed shape data including a group of curvatures acquired from two-dimensional data is generated.

  Specifically, the detailed shape data generation unit 36 first cuts out three-dimensional data of a region having a set width (see FIG. 5) in the Y-axis direction from the entire three-dimensional data from which noise has been removed. .

  Then, as shown in FIG. 5, a predetermined range of the second set interval D2 is set in the X-axis direction, and the predetermined range is set in the X-axis direction at the pitch of the virtual point group (interval between adjacent virtual points). While shifting, the extraction of the three-dimensional data of the virtual point group included in the predetermined range and the calculation of the moving average of the Z coordinate values of the extracted three-dimensional data are performed. Thereby, the value of the Z coordinate is generated at a pitch equal to the pitch of the virtual point group in the X-axis direction. The second setting interval D2 is smaller than the first setting interval D1, and is, for example, about 1/10 to 1/2 of the first setting interval D1. When an exclusion point exists in the predetermined range, the three-dimensional data of the exclusion point is excluded.

  Next, the detailed shape data generation unit 36 converts the three-dimensional data into two-dimensional data. Since this conversion method is the same as the conversion method performed by the base shape data generation unit 35, the description thereof is omitted.

  FIG. 7 is a diagram for explaining a curvature generation method by the curvature generation unit 37.

  The curvature generation unit 37 generates curvatures in curves represented by the two-dimensional data generated by the base shape data generation unit 35 and the detailed shape data generation unit 36, respectively.

  Specifically, the two-dimensional data generated by the base shape data generation unit 35 and the detailed shape data generation unit 36 includes a plurality of pieces shifted in the X-axis direction at a virtual point pitch set on the evaluation target surface A. The value of the Z coordinate at the point. In the curvature generating unit 37, for example, a point group of three or more points adjacent to each other in the X-axis direction is set, and a quadratic curve passing through the vicinity of the point group is set as an approximate curve F with the vicinity of the center of the point group as a vertex. Is done. Then, a curvature at the vertex of the approximate curve F is obtained, and the curvature is generated as a curvature at a point located near the vertex of the point group.

  When the curvature is generated by the curvature generation unit 37 for the two-dimensional data, the base shape data generation unit 35 replaces the value of each Z coordinate included in the two-dimensional data with the corresponding curvature. Thereby, base shape data including a curvature group including curvatures at a plurality of points arranged at a predetermined pitch in the X-axis direction is generated.

  When the curvature is generated by the curvature generation unit 37 for the two-dimensional data, the detailed shape data generation unit 36 replaces the value of each Z coordinate included in the two-dimensional data with the corresponding curvature. Thereby, detailed shape data including a curvature group including curvatures at a plurality of points arranged at a predetermined pitch in the X-axis direction is generated.

  The design shape data acquisition unit 38 reads and acquires design shape data from the memory of the shape recognition device 6, for example. The design shape data is created in advance and stored in the memory of the shape recognition device 6. The design shape data is obtained from the design data of the door panel DP having the evaluation target surface A, and the curvature at each virtual point set in a lattice shape with a pitch of less than several millimeters on the evaluation target surface A is obtained. It is created by associating curvatures at virtual points.

  FIG. 8 is a flowchart showing the flow of the distortion value generation process executed by the distortion value generation unit 39. FIG. 9 is a diagram illustrating an example of curves represented by base shape data, detailed shape data, and design shape data. FIG. 10 is a diagram showing another example of curves represented by base shape data, detailed shape data, and design shape data.

  The distortion value generation unit 39 acquires each curvature (hereinafter referred to as “base curvature”) included in the base shape data, and sets a certain allowable width for each base curvature (step S1).

  Thereafter, the distortion value generation unit 39 extracts design shape data in a region corresponding to the base shape data from the design shape data acquired by the design shape data acquisition unit 38. Then, for each base curvature included in the base shape data, the distortion value generation unit 39 acquires the curvature of the point corresponding to the base curvature (hereinafter referred to as “design curvature”) from the design shape data, and each of them. It is determined whether or not the design curvature is included within the allowable range (allowable range) set to the corresponding base curvature (step S2).

  When the design curvature is included within the allowable range (YES in step S2), the distortion value generation unit 39 acquires the curvature of the point corresponding to the base curvature (hereinafter referred to as “detail curvature”) from the detailed shape data. The difference between the base curvature and the detailed curvature is calculated, and the difference is generated as a distortion value (step S3).

  On the other hand, when the design curvature is not included in the allowable range (NO in step S2), the distortion value generation unit 39 acquires the detailed curvature of the point corresponding to the base curvature from the detailed shape data, and the detailed curvature and the design curvature And the difference is generated as a distortion value (step S4).

  When the design curvature is within the allowable range, as shown in FIG. 9, the curve represented by the base shape data approximates the curve represented by the design shape data, so that the difference between the base curvature and the detailed curvature is distorted. By setting the value, the distortion in the evaluation target surface A can be accurately and quantitatively evaluated from the distortion value.

  On the other hand, when the design curvature is not included in the allowable range, the curve represented by the base shape data greatly deviates from the curve represented by the design shape data, as shown in FIG. For example, when a wide range of distortion has occurred on the evaluation target surface A, the curve represented by the base shape data has a portion N greatly deviated from the curve represented by the design shape data. In this case, if the difference between the base curvature and the detailed curvature in the portion N is a distortion value, the distortion value does not reflect the distortion in the evaluation target surface A. Therefore, for a portion N greatly deviating from the curve represented by the design shape data in the curve represented by the base shape data, the difference between the detailed curvature and the design curvature is taken as the distortion value. Thereby, since the distortion over a wide range of the evaluation target surface A is reflected in the distortion value, the distortion in the evaluation target surface A can be quantitatively evaluated with higher accuracy based on the distortion value.

  The measurement target shape generation unit 40 illustrated in FIG. 2 generates shape image data indicating an outline indicating the outer edge of the evaluation target surface A based on the three-dimensional data.

  The evaluation image generation unit 41 acquires shape image data from the measurement target shape generation unit 40, divides the internal region of the shape image data into areas including the same distortion value, and the hue corresponding to the distortion value for each area. And the density is set, and the image obtained by coloring the internal region of the shape image data with the hue and density set for each area is displayed on the display D.

  By displaying this image on the display D, the evaluation operator can visually grasp the region where the distortion exists, the direction of the distortion, and the degree of the distortion.

  In the above-described embodiment, the case where the object having the evaluation target surface A is the door panel DP is taken as an example. However, the object having the evaluation target surface A is not limited to the door panel DP.

  The present invention can also be implemented in other forms, and various modifications can be made to the above-described configuration within the scope of the matters described in the claims.

6 Shape Recognition Device 35 Base Shape Data Generation Unit (Base Shape Data Generation Unit)
36 Detailed shape data generation unit (detailed shape data generation means)
38 Design shape data acquisition unit (design shape data acquisition means)
39 Strain value generator (Strain value generator)

Claims (1)

Base shape data generating means for generating base shape data consisting of a group of curvatures reflecting the shape of the object from the three-dimensional shape data representing the shape of the surface of the object;
Detailed shape data generating means for generating detailed shape data consisting of a group of curvatures reflecting the shape of the surface of the object relatively in detail from the three-dimensional shape data;
Design shape data acquisition means for acquiring design shape data consisting of a design curvature group of the surface of the object;
When the curvature acquired from the design shape data is included in an allowable range of a predetermined width with respect to the curvature acquired from the base shape data, with respect to the predetermined region on the surface of the object, the surface is acquired from the base shape data. When the difference between the curvature and the curvature acquired from the detailed shape data is generated as a distortion value, and the curvature acquired from the design shape data is not included in the allowable range, the curvature acquired from the design shape data; A shape recognition device comprising: a distortion value generating means for generating a difference from a curvature acquired from the detailed shape data as a distortion value.

Images