JP2007248072A - Apparatus and method for inspecting irregular color - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus and method for inspecting irregular color capable of correctly determining existence of irregular color. <P>SOLUTION: An inspection surface of an inspected object is imaged, a three-dimensional shape on the inspection surface is determined based on the imaged image data, variation A of an arbitrary sectional shape is determined based on the three-dimensional shape, variation B of the color of the surface in this sectional part is determined based on the image data, and similarity of the inflection degree of the shape variation A and color variation B is evaluated. A part (α) whose color varies similarly to the shape variation A is determined not to be irregular color, and a part (β) whose color varies without similarity to the shape variation A is determined to be irregular color. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an uneven color inspection device and an uneven color inspection method.

  Conventionally, as a method for inspecting the color unevenness of the inspection surface of the inspection object, a plurality of arbitrary pixel data are extracted by image processing from a planar image obtained by imaging the inspection surface of the inspection object, and the hue and saturation thereof are extracted. Color unevenness fog is determined from relative evaluation of brightness (see, for example, Patent Document 1).

Such a method for inspecting color unevenness by image recognition is also used in the case of inspecting the color of the painted surface after coating the outer surface of an automobile.
JP 2004-170109 A

  By the way, the painting of the outer surface of an automobile is not necessarily performed on the outer surface of the same part using the same color paint in the same place. For example, the painting is performed on the outer surfaces of two adjacent parts such as a trunk and a rear fender. Then, painting may be performed separately in different places, that is, in different lighting environments. For this reason, the painted surfaces of the two separately coated parts can have slightly different colors. If this happens, the color change may be noticeable especially at the joint of these two parts.

  However, in the conventional color image unevenness inspection, only the hue, saturation, and brightness of a plurality of pixel data are relatively evaluated. This is not taken into account until the cross-sectional shape of the inspection surface changes across the gap or step. For this reason, there has been a problem that an inspection result obtained by the color unevenness inspection apparatus differs from an appropriate evaluation result obtained by a visual inspection by a skilled inspector. Further, even in the case of the same component, it may be difficult to determine whether the color unevenness is good or not because the visible hue changes depending on the light source position and the viewpoint position.

  The present invention has been made to solve the above-described problems, and it is an object of the present invention to provide a color unevenness inspection apparatus and method that can more reliably determine the presence or absence of color unevenness.

  To achieve the above object, the present invention provides a shape acquisition means for acquiring the shape of the inspection object on the inspection surface, a color acquisition means for acquiring the same color as the part from which the shape has been acquired, and a change in the shape. A color unevenness inspecting apparatus, comprising: a judgment means for evaluating similarity of the degree of inflection of the color change and judging the presence / absence of color unevenness from the evaluation result.

  Further, the present invention for achieving the above object acquires the shape of the inspection object on the inspection surface and also acquires the color of the inspection surface, and the degree of inflection of the change in the shape and the change in the color is obtained. A color unevenness inspection method characterized by evaluating similarity and determining the presence or absence of color unevenness based on the evaluation result.

  In addition, the present invention for achieving the above object simulates a color change due to a three-dimensional shape of an inspection object using computer graphics on the inspection surface of the inspection object, and uses the color change of the simulation result as a reference. And storing the actual color change on the inspection surface from the image data obtained by capturing the inspection object with a three-dimensional camera, and obtaining the actual color change and the reference information color change. A color unevenness inspection method characterized by collating with an inflection point, evaluating similarity based on the degree of coincidence of the inflection point, and determining the presence or absence of color unevenness based on the evaluation result.

  According to the present invention, color unevenness is determined by evaluating the similarity of the degree of change in the shape of the inspection surface of the inspection object and the color of the portion, so that the appearance depending on the shape change It is possible to reliably identify whether the color is different or the color of the paint itself is uneven. In addition, according to the present invention, the present invention can be applied without problems even under conditions in which the hue, saturation, and brightness change due to the angle difference between the imaging surface and the camera, which is peculiar to the coating of the inspection surface, and the light source position and viewpoint position change. However, it is possible to correctly determine whether the color unevenness is good or bad.

  Embodiments according to the present invention will be described below with reference to the drawings.

<First Embodiment>
FIG. 1 shows a vehicle color unevenness inspection apparatus according to an embodiment of the present invention, which is electrically connected to a three-dimensional measuring instrument 1 as an imaging means for imaging an inspection object and the three-dimensional measuring instrument 1. The color unevenness determination device 2 is configured.

  The three-dimensional measuring instrument 1 is provided with a three-dimensional camera 6 at the tip of an arm 5 of a robot 4 provided so as to be positioned in the vicinity of a vehicle 3 that is an object to be inspected. By measurement, the shape of the vehicle 3 and the color of the paint can be measured simultaneously. The three-dimensional measuring instrument 1 also has a collation function with CAD data.

  Three-dimensional information measurement by the three-dimensional camera 6 includes an active method for irradiating light for three-dimensional information measurement and a passive method for detecting natural illumination, but any known method including the phase shift method described above. Can be adopted. For example, a method using triangulation that projects light rays or patterns, a light cutting method by projecting sheet light or a lattice dot pattern, a moire topography by projecting a sheet light group, or the like can be used.

  The color unevenness determination apparatus 2 includes a computer 7 that functions as an image processing unit that processes an image captured by the three-dimensional camera 6, and a monitor 8 that displays an inspection result processed by the computer 7.

  FIG. 1 shows a state in which the three-dimensional camera 6 captures an image of an inspection surface region that straddles a joint portion of different parts such as a trunk and a rear fender of the vehicle 3 as an inspection surface 9 of an inspection object. A curve A in FIG. 2 shows a change in the cross-sectional shape (contour line shape) on the inspection surface 9, that is, a change in the shape of the ridge line 10 of the cross-section on the inspection surface 9. A curve B in FIG. 2 shows changes in hue, saturation, and brightness as color changes in the ridgeline 10 (details will be described later). In addition, since the drawings shown here are for explaining the present embodiment, the inspection surface 9 of the inspection object is naturally performed on a surface other than the inspection surface over which different parts are combined. is there.

  The computer 7 as the image processing means includes shape acquisition means for acquiring the shape of the ridge line 10 of the cross section on the inspection surface 9 from the image data captured by the three-dimensional camera 6, and ridge lines from the image data captured by the three-dimensional camera 6. The color acquisition means for acquiring 10 colors, the similarity of the degree of inflection (inflection rate) of the obtained change in the shape of the ridge line 10 and the change in the color, and the presence or absence of color unevenness according to the evaluation result It functions as a judging means for judging The determination means in this embodiment evaluates the degree of coincidence of inflection points and determines the presence or absence of color unevenness based on the evaluation result.

  FIG. 3 is a flowchart showing the flow of the hue inspection process in the color unevenness inspection apparatus according to the first embodiment.

  First, the inspection object is imaged by the three-dimensional camera 6, and three-dimensional data of the inspection object is generated from the captured image (S101), and an arbitrary cross-section is detected by detecting arbitrary continuous pixels on the inspection surface 9. Data (the ridgeline 10 of the cross section of the inspection surface 9) is acquired (S102). Next, this cross-sectional data is converted into a cross-sectional shape vector (S103), and an inflection point of the cross-sectional shape vector is extracted (S104). Thereby, the change in the shape of the ridgeline 10 (two-dimensional ridgeline) of the cross section on the inspection surface 9 is acquired. By accumulating the cross sections (tomographic images), the entire surface can be calculated and the relative shape change in three dimensions can be known.

  On the other hand, the hue, saturation, and brightness of the continuous pixels are extracted from the three-dimensional data of the inspection object (S105), and these are converted into hue, saturation, and brightness vectors to obtain color changes (S106). In addition, inflection points of the hue, saturation, and brightness vector are extracted (S107).

  Next, by comparing the inflection points of the three-dimensional shape change indicated by the cross-sectional shape vector with the inflection points of the color change indicated by the hue, saturation, and brightness vectors, the continuity is compared ( S108), the similarity of the degree of inflection is evaluated (S109).

  If the color change of hue, saturation, and brightness is based on the change of the cross-sectional shape of the inspection surface 9, the color change of hue, saturation, and lightness and the change of the cross-sectional shape are changed as shown in FIG. The same phenomenon occurs at the same position (pixel) on the inspection surface 9, and the similarity of the degree of inflection (inflection rate) appears strongly. 2, the cross-sectional shape of the inspection surface 9 has irregularities 11 such as ridges, gaps, and surface differences of the mating portion as viewed in the shape pattern (curve A) of the ridge line 10 of the cross section. Depending on the unevenness 11, the color change pattern (curve B) of hue, saturation, and brightness also changes greatly, and the inflection points a1, a2, a3 and the inflection points b1, b2, b3 on both patterns are The positions on the inspection surface 9 match. Therefore, processing without color unevenness such as alarm non-display is performed (S110).

  On the other hand, if the color change of hue, saturation, and lightness is not based on the change of the cross-sectional shape of the inspection surface 9, the color change of hue, saturation, and lightness is as shown in FIG. The similarity of the degree of inflection (inflection rate) does not appear between changes in the cross-sectional shape. In FIG. 2B, the cross-sectional shape of the inspection surface 9 is almost flat as seen from the shape pattern (curve A) of the cross-sectional ridge line 10, whereas the hue, saturation, and lightness color change patterns ( Since the curve B) is prominently raised regardless of this, it is understood that so-called “color unevenness” occurs on the inspection surface 9 at this portion. Therefore, processing with uneven color such as alarm display is performed (S111).

  Then, the evaluation of the color unevenness of each tomographic image is spread over the entire inspection surface, and is accumulated to determine the presence or absence of color unevenness for the entire inspection surface 9.

  As described above, the first embodiment determines whether the inflection point existing on the continuous data depends on the shape on the inspection surface 9, depends on the light source, or is uneven in color or hue. Therefore, the color unevenness in each part is detected without detecting the color change in the part where the shape of the inspection surface is changed like the joint part of two parts like the joint part of two adjacent parts. In addition, it is possible to determine whether or not there is uneven color between the two parts.

  Here, the painted surface between the two parts is, for example, a painted surface between various parts such as the aforementioned trunk and rear fender, bonnet and front fender, and body and bumper in the case of an automobile body. In particular, the body and bumper are made of steel (body) on the one hand and resin material (bumper) on the other, so the difference in color tends to be noticeable and the consistency of the hue is high. According to this embodiment, it is possible to reliably inspect the hue between these two parts.

  In this embodiment, since the three-dimensional camera is used to obtain the three-dimensional shape of the inspection surface and the color is acquired from the image data captured by the same camera, the shape acquisition position and the color acquisition position are shifted. Therefore, it is possible to reliably evaluate the color unevenness in the present embodiment. In particular, by doing so, it is not necessary to perform alignment between the shape acquisition position and the color acquisition position as compared with the case of using another three-dimensional shape measurement apparatus, so it is easy to determine the color village. It can be performed. Of course, in addition to the method using a 3D camera, by accurately matching the shape acquisition position and color acquisition of the inspection surface, it is possible to acquire the account using other 3D shape measurement devices, etc. May be obtained.

<Second Embodiment>
FIG. 4 is a flowchart showing the flow of the hue inspection process in the color unevenness inspection apparatus according to the second embodiment. The configuration of the color unevenness inspection apparatus according to the second embodiment is the same as that of the first embodiment described above, and only the processing is different as will be described later. Therefore, a description of the device configuration is omitted.

  First, the inspection object is imaged by the three-dimensional camera 6, and three-dimensional data of the inspection object is generated from the captured image (S201), and any continuous pixel (cross-sectional ridgeline) that straddles the matching portion of the inspection surface 9 10) is extracted (S202), converted into a cross-sectional shape vector (S203), and an inflection point of the cross-sectional shape vector is extracted (S204). Thereby, a change in the shape of the ridgeline 10 of the cross section on the inspection surface 9 is acquired.

  On the other hand, the color composed of R, G, B of the continuous pixels is detected from the three-dimensional data of the inspection object (S205), and these are converted into change amount vectors to extract the color change (S206). Further, inflection points of the R, G, and B vectors are extracted (S207).

  Next, the inflection points of the cross-sectional shape vector and the inflection points of the R, G, and B vectors are collated. That is, by comparing the inflection point of the shape change of the ridge line 10 with the inflection points of the color change of R, G, and B, the continuity is compared (S208), and the similarity of the inflection degree is obtained. Evaluate (S209).

  If the color change of R, G, B is based on the change of the cross-sectional shape of the inspection surface 9, the color change of R, G, B and the change of the cross-sectional shape occur in the same way, and the degree of inflection (inflection rate) ) Strongly appears, and at least the position of the inflection point on the inspection surface 9 matches. Therefore, processing without color unevenness such as alarm non-display is performed (S210).

  On the other hand, if the color change of R, G, B is not based on the change of the cross-sectional shape of the inspection surface 9, the degree of inflection is between the color change of R, G, B and the change of the cross-sectional shape. Since the similarity of (inflection rate) does not appear and at least the position of the inflection point on the inspection surface 9 is inconsistent, it can be seen that so-called “color unevenness” occurs on the inspection surface 9 at this site. Therefore, processing with uneven color, such as displaying an alarm, is performed (S211).

  In this way, by dividing the acquired colors into the three primary colors R, G, and B and evaluating each of them, not only the color unevenness of the entire paint color but also what color elements are missing Further, it is possible to perform detailed evaluation of color unevenness such as whether or not it is excessive.

<Third Embodiment>
FIG. 5 is a flowchart showing the flow of the hue inspection process in the color unevenness inspection apparatus according to the third embodiment of the present invention. The configuration of the color unevenness inspection apparatus according to the third embodiment is the same as that of the first embodiment described above, and only the processing is different as will be described later. Therefore, a description of the device configuration is omitted.

  First, the inspection object is imaged by the three-dimensional camera 6 using a light cutting method by projection of a lattice dot pattern, and three-dimensional data of the inspection object is generated from the captured image (S301, S302).

  Next, the light source → the inspection object (subject) → the reflection angle of the camera is calculated for each pixel of arbitrary cross-sectional data straddling the matching portion of the inspection surface 9 (S303). Here, the reflection angle refers to a reflection angle at which the reflected light from the inspection object onto which the lattice dot pattern is projected from the light source travels toward the three-dimensional camera 6 that is the observation direction. In the case of the light cutting method based on the projection of the lattice dot pattern, the positions of the light source, the inspection object, and the three-dimensional camera 6 are known, so that the reflection angle from the inspection object toward the three-dimensional camera 6 can be calculated. The cross sectional shape can be known. Next, the inflection point of the reflection angle is extracted (S304). Thereby, the change in the shape of the ridge line 10 of the cross section on the inspection surface 9 is indirectly acquired.

  On the other hand, the color consisting of R, G, B of the continuous pixels is detected from the three-dimensional data of the inspection object (S305), and these are converted into R, G, B vectors to extract color changes ( In step S306, the inflection points of the R, G, and B vectors are extracted (S307).

  Next, the inflection point of the reflection angle for each pixel of the cross-sectional data is compared with the inflection points of the R, G, and B color changes to compare their continuity (S308). The degree of similarity is evaluated (S309).

  If the color change of R, G, B is based on the change of the cross-sectional shape of the inspection surface 9 indicated by the reflection angle, the color change of R, G, B and the change of the cross-sectional shape occur in the same way. The similarity of the degree of inflection (inflection rate) appears strongly, and at least the position of the inflection point on the inspection surface 9 matches. Therefore, processing without color unevenness such as alarm non-display is performed (S310).

  On the other hand, if the color change of R, G, B is not based on the change of the cross-sectional shape of the inspection surface 9 indicated by the reflection angle, the color change of R, G, B and the change of the cross-sectional shape In this case, the similarity of the degree of inflection (inflection rate) does not appear, and at least the position of the inflection point on the inspection surface 9 is inconsistent. You can see that Therefore, processing with uneven color such as alarm display is performed (S311).

  In the above-described embodiment, the case where the change in the shape of the cross-sectional ridge line 10 on the inspection surface 9 is acquired by image processing from image data obtained by imaging the inspection object, but the information on the change in the cross-sectional shape on the inspection surface 9 is as follows. You may acquire by diverting CAD data.

  Specifically, for the inspection surface 9 of the inspection object, a color change due to the three-dimensional shape of the inspection object using computer graphics is simulated, and the color change of the simulation result is stored as reference information, The actual color change on the inspection surface 9 is acquired from the image data obtained by imaging the inspection object with the three-dimensional camera 6, and the acquired actual color change and the inflection point of the color change of the reference information are displayed. In the color unevenness inspection method, the similarity is evaluated based on the degree of coincidence of the inflection points, and the presence or absence of color unevenness is determined based on the evaluation result. According to this, it is possible to evaluate the hue of the inspection object over a wide range.

  In this way, the color change for the three-dimensional object is obtained in advance (or after the fact) by simulation, and the result and the actual color change are compared and evaluated, for example, from various directions. Illuminate the inspection surface, simulate the color change from the difference in the position of the light source of the illumination, and use this as a reference to match the actual captured image of the inspection object with the color inflection point on the surface, Uneven defects can also be detected. By using the simulation in this way, it becomes possible to discriminate between the difference in the appearance of the color simply due to the change in the lighting condition and the case where the color unevenness actually occurs.

  INDUSTRIAL APPLICABILITY The present invention can be used in various fields as a technique for inspecting color unevenness of a painted surface, and can be applied to, for example, an automobile body or a home appliance.

It is a perspective view which shows the outline | summary of the color nonuniformity inspection apparatus which concerns on embodiment of this invention. It is the figure which showed the change of the shape of the ridgeline of the cross section in the test | inspection surface which concerns on the 1st Embodiment of this invention, and the change of the color in this ridgeline. It is a flowchart which shows the color nonuniformity inspection process which concerns on the 1st Embodiment of this invention. It is a flowchart which shows the color nonuniformity inspection process which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows the color nonuniformity inspection process which concerns on the 3rd Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Three-dimensional measuring instrument 2 ... Color unevenness determination apparatus 3 ... Vehicle 4 ... Robot 5 ... Arm 6 ... Three-dimensional camera 7 ... Computer 8 ... Monitor 9 ... Inspection surface 10 ... Ridge line 11 ... Unevenness

Claims (9)

Shape acquisition means for acquiring the shape of the inspection object on the inspection surface;
Color acquisition means for acquiring the color of the same part as the part from which the shape has been acquired;
About the change in the shape and the change in the color, the similarity of the degree of inflection is evaluated, and determination means for determining the presence or absence of color unevenness from the evaluation result;
An apparatus for inspecting color unevenness, comprising: The shape acquisition means
Imaging means for imaging the inspection surface;
Image processing means for processing an image picked up by the image pickup means,
The image processing means obtains the shape of the ridge line of the cross section on the inspection surface from the captured image data,
The color acquisition unit acquires the color of the surface of the ridge line portion from the image data captured by the imaging unit,
The color unevenness inspection apparatus according to claim 1, wherein the determination unit evaluates the similarity of the degree of inflection of the acquired shape change of the ridge line and the color change of the surface of the ridge line part.   The color unevenness inspection apparatus according to claim 2, wherein the imaging unit is a three-dimensional camera.   The color unevenness inspection apparatus according to claim 2, wherein the inspection surface is a region straddling a matching part of different parts. The color acquisition means acquires a change in hue, saturation, and brightness from the image data of the inspection surface,
The color unevenness inspection apparatus according to claim 1, wherein the determination unit performs the evaluation for each of the acquired hue, saturation, and brightness. The color acquisition means acquires the three primary colors separately from the image data of the inspection surface,
The color unevenness inspection apparatus according to claim 1, wherein the determination unit performs the evaluation for each of the acquired three primary colors. Obtaining the shape of the inspection surface of the inspection object and obtaining the color of the inspection surface;
A color unevenness inspection method, wherein similarity of the degree of inflection is evaluated for the change in shape and the change in color, and the presence or absence of color unevenness is determined based on the evaluation result. The shape on the inspection surface and the color of the inspection surface are:
The color unevenness inspection method according to claim 7, wherein the inspection surface is acquired from image data captured by an imaging unit.   With respect to the inspection surface of the inspection object, computer graphics are used to simulate the color change due to the three-dimensional shape of the inspection object, and the color change of the simulation result is stored as reference information, while the inspection object is a 3D camera. The actual color change on the inspection surface is acquired from the image data captured in step, the acquired actual color change is compared with the inflection point of the reference information color change, and the inflection point A color unevenness inspection method characterized by evaluating similarity based on the degree of coincidence and determining the presence or absence of color unevenness based on the evaluation result.