JP2009008502A - Data registering method for surface state inspection and surface state inspection device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately and efficiently set an imaging target region while preventing a necessary quantity or above of imaging from being performed. <P>SOLUTION: The position and imaging direction of a camera 1 necessary for imaging the imaging target region is set while determining the imaging target region at every time prior to inspecting the surface uneven flaw of a workpiece W. In this setting processing, the normal direction histogram of the whole of a surface to be inspected is formed using CAD data which indicates that the surface shape of the workpiece is expressed as the aggregate of a triangular plane and a plane becoming a predetermined tolerance value or below in the angle difference with respect to the normal line direction of an imaging center point is extracted from the histogram to recognize an inspectable range. Further, an angle for showing the irregularity in the advance direction of reflected light with intensity required in inspection is specified and the value based on the angle is set as a tolerance value to ensure the recognizing precision of the inspectable range. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a method for registering information relating to setting of an imaging target region for a surface to be inspected in order to inspect the surface state of an object having a predetermined shape by an image processing technique, and a surface state inspection using this method Relates to the device.

  Many inspection apparatuses using an image processing technique have been proposed for inspecting the surface state of a molded body made of resin or metal. In general, this type of inspection apparatus illuminates a molded object to be inspected (hereinafter referred to as “work”) from a predetermined direction, images reflected light from the surface of the work, and differs in the generated image from the surroundings. An area where brightness or color appears is detected as a defect (for example, see Patent Document 1).

JP 2004-108833 A

In this type of inspection apparatus, in order to detect defects with high accuracy, it is necessary to adjust the direction of illumination and the imaging direction in accordance with the direction of the surface of the workpiece to be inspected (hereinafter referred to as “work surface”). In particular, when an inspection using specular reflection light such as an uneven defect is performed, it is necessary to accurately align the optical axis of the camera in the direction in which the specular reflection light travels.
Therefore, prior to the inspection, the user must align the camera and the workpiece using a non-defective workpiece so that imaging that satisfies the above conditions is performed on any part of the workpiece surface to be inspected. And an operation for determining the imaging direction (hereinafter referred to as “setting operation”).

When the workpiece surface to be inspected becomes a certain size, in order to ensure the inspection accuracy, it is necessary to divide the workpiece surface into a plurality of areas and take images.
Further, when inspecting each surface of a polyhedral work, it is necessary to perform the above setting operation for each surface. In the above-mentioned patent document 1, since a minute object (a fastener of a slide fastener) is an inspection target, an image necessary for the inspection is acquired by one imaging. It is thought that accuracy cannot be secured. Also, the configuration of the apparatus becomes large.

  Furthermore, in recent years, devices that place importance on fashion, such as mobile phones and portable music players, use case bodies with a combination of free-form surfaces including portions with various curvatures. When a workpiece having such a form is an inspection target, the range in which the specularly reflected light can be incident on the camera varies depending on the curvature of the surface in the imaging target region.

  In addition to the light reflected in the specular specular reflection direction from the surface of the workpiece, there is also specular reflection light that scatters to a specified range due to minute irregularities on the surface. Detection of defects by these scattered light The necessary brightness may be ensured. Therefore, in order to specify a range (hereinafter referred to as “inspectable range”) in which an image having a brightness necessary for detecting an uneven defect is obtained in the imaging target region, the theoretical specular reflection direction and the imaging target region In addition to the relationship with the curvature of the surface, it is necessary to consider the degree of variation in scattered light.

  However, it is very difficult to specify the curvature of the surface included in the imaging target region every hour and the degree of variation in scattered light as numerical values. For this reason, in the conventional inspection, the actual object is marked, for example, the imaging target area assigned to the workpiece is imaged in order, and the part where sufficient brightness is obtained by the hourly imaging is confirmed. Setting work is performed until confirmation of the entire surface is obtained. However, with such a method, the burden on the user is enormous, and there is a possibility that confirmation may occur. Further, an imaging target area more than necessary is allocated to prevent an inspection omission, and as a result, the inspection efficiency may be reduced.

  The present invention has been made paying attention to the above-mentioned problems, and accurately and efficiently executes processing for setting an imaging target region so that images necessary for inspection can be obtained for all portions of the surface to be inspected. In addition, it is an object to prevent unnecessary imaging from being performed and to improve inspection efficiency by correctly recognizing an inspectable range in the imaging target region every hour.

  In the present invention, an object whose surface is partly or entirely the surface to be inspected is to be inspected, and the surface to be inspected is divided into a plurality of times while changing the position and direction of the imaging device and the illuminating device with respect to the inspection object In order to register information related to the setting of the imaging target area in the hourly imaging in the inspection apparatus that inspects the state of the surface to be inspected using the specularly reflected light image in the generated image, the following steps A, B, and C are executed for a plurality of cycles.

In step A, an imaging target region is set at an arbitrary position on the surface to be inspected.
In step B, when the positions and directions of the imaging device and the illuminating device are determined so that the specularly reflected light from the center position of the imaging target region set in step A is incident on the imaging device, it is reflected with an intensity of a predetermined value or more. An angle representing the degree of variation in the traveling direction is determined for the light to be transmitted.

  In step C, using the three-dimensional design data representing the surface of the object as an aggregate of a plurality of planes, the plane corresponding to the center position of the area is selected from the planes corresponding to the imaging target area set in step A. A plane in which the difference in inclination corresponds to the range of angles obtained in step B is specified.

  Further, in the present invention, the above steps A, B, and C are executed in a plurality of cycles to identify all planes corresponding to the surface to be inspected, and information regarding the setting of the imaging target area set in step A every hour (set) Information necessary for imaging the imaging target area, for example, the position where the imaging device is arranged and its optical axis direction) are registered.

  In the above method, by executing steps B and C for the imaging target area set in step A, a range in which an image having a brightness necessary for detecting a defect is obtained by imaging the imaging target area is 3 Recognize by replacing with the approximate plane represented by the dimensional design data. Therefore, it is necessary to inspect any part of the surface to be inspected by any imaging by repeating the processes in steps A, B, and C until it is recognized that inspection of all the parts on the surface to be inspected is possible. Can be obtained.

  In addition, since the inspectable range can be specifically specified in the set imaging target area, the setting position of the imaging target area is set so that the already specified parts are not included in the subsequent setting of the imaging target area. Can be adjusted. Thereby, it is possible to prevent the imaging target region from being set more than necessary.

  In a preferred mode, in step B, the non-defective model of the inspection object is targeted, and the field of view of the imaging device is aligned with the imaging target area set in step A, and the specularly reflected light from the center position of the imaging target area The first model image is generated by determining the positions and directions of the imaging device and the illuminating device so as to be incident on the imaging device, and the degree of variation in the traveling direction of reflected light from the imaging target region is determined. Assuming the value of the angle to be represented, a second model image of the imaging target region set in step A is generated using the assumed angle and the three-dimensional design data.

  Further, in the above aspect, the first and second model images are displayed in a collatable manner, and the second model image is updated in accordance with an operation for changing the assumed angle, and the display of each model image is confirmed. When an operation is performed, an arithmetic operation is performed using a theoretical formula with the assumed angle at that time as the range of scattering, and an angle representing the degree of variation in the traveling direction of light reflected at an intensity of a predetermined value or more is specified.

  The above aspect displays an actual image generated by imaging a non-defective model and a second model image created by simulation based on a temporary angle representing the degree of variation in the traveling direction of reflected light, and By collating, an angle representing the degree of variation in the traveling direction of light reflected at the intensity required for inspection is automatically calculated. The user determines whether or not the second model image is close to the first model image, and performs an operation of changing the temporary angle when the two images are different. When it is determined that the second model image is finally close to the first model image and a determination operation is performed, a theoretical formula (for example, a Gaussian scattering model of the Gaussian scattering model) with a temporary angle at that time as a range of scattering is performed. The angle representing the degree of variation in the light reflected with the intensity necessary for detecting the defect is specified by the calculation using the theoretical formula.

  In another preferred embodiment, in step B, a non-defective model of the inspection object is targeted, and the field of view of the imaging device is aligned with the imaging target region set in step A, and the direction of illumination by the lighting device is switched multiple times. The imaging is executed. Then, an angle representing the degree of variation in the traveling direction of light reflected at an intensity of a predetermined value or higher is specified based on the illumination direction in imaging in which an image having a brightness of a predetermined value or higher is obtained.

  According to the above aspect, the width of the change in the illumination direction in which an image having the brightness necessary for detecting the defect is obtained by actually performing imaging for each illumination while switching the illumination direction with respect to the non-defective model. The angle representing the variation in the traveling direction of the light reflected with the intensity required for the inspection is specified.

  In yet another preferred aspect, in step C, a histogram showing the distribution state in the normal direction of each plane corresponding to the imaging target region is created using the three-dimensional design data, and the plane at the center of the imaging target region is created from this histogram. The normal direction included in the allowable range based on the angle obtained in step B with the angle difference with respect to the normal direction corresponding to is extracted, and the plane corresponding to each extracted normal direction is specified. According to such processing, it is possible to easily and accurately recognize a part where an image necessary for an inspection can be obtained in the imaging target region.

  A surface state inspection apparatus to which the above-described inspection target area registration method is applied includes an imaging apparatus for imaging an inspection target, an illumination apparatus for illuminating an imaging target area of the imaging apparatus, an imaging apparatus, and an illumination apparatus A position of the imaging device and the illumination device with respect to the inspection object by moving at least one of the first support mechanism for supporting the inspection object, the second support mechanism for supporting the inspection object, and the first and second support mechanisms; A process of changing the direction and causing the imaging device to image a predetermined imaging target area on the surface to be inspected is executed a plurality of times, and the surface state of the surface to be inspected is inspected using an image generated by the hourly imaging And a control processing device that executes processing for the above.

The inspection apparatus further includes a display unit for displaying a model image of each imaging target region and an input unit for receiving a user operation.
Further, the control processing device targets storage means for registering information relating to setting of the imaging target area, area setting means for setting the imaging target area at an arbitrary position on the surface to be inspected, and a non-defective model of the inspection target. The positions and directions of the imaging device and the illumination device so that the field of view of the imaging device is aligned with the imaging target region set by the region setting means, and the specularly reflected light from the center position of the imaging target region is incident on the imaging device. And a degree of variation in reflected light from the imaging target region set by the region setting unit, the first image generating unit generating the first model image of the imaging target region by causing the imaging device to perform imaging Assuming that a predetermined value is used as the angle representing the first angle, and using the assumed angle and the three-dimensional design data representing the surface of the inspection object as an aggregate of a plurality of planes, The second image generating means for generating the model image, and the first and second model images are displayed on the display means so that they can be collated, and the input means receives the operation for changing the assumed angle. Display control means for causing the second image generation means to update the second model image on the basis of the angle changed by the step of updating the display of the display means with the updated image, and input means under the display state of each model image. When a confirmation operation is accepted, a calculation is performed according to a theoretical formula with the assumed angle at the time of the operation as the range of scattering, and the degree of variation in the traveling direction of reflected light incident on the imaging device with an intensity of a predetermined value or more is calculated. The angle with respect to the plane of the center position of the area from the plane corresponding to the imaging target area set by the area setting means using the angle specifying means for specifying the angle to be expressed and the three-dimensional design data Plane specifying means for specifying a plane whose difference is included in the allowable range based on the angle specified by the angle specifying means, region setting means, first image generating means, second image generating means, display control means, angle Registration means for registering information related to the setting of each of a plurality of imaging target areas set while the processing by each means of the specifying means and the plane specifying means is executed in a plurality of cycles;

  In a preferred aspect of the inspection apparatus described above, the plane specifying unit creates a histogram indicating the distribution state in the normal direction of each plane corresponding to the imaging target area using the three-dimensional design data, and the imaging target area is determined from the histogram. A normal direction whose angle difference with respect to the central plane is included in the range of the allowable value is extracted, and a plane corresponding to each extracted normal direction is specified.

  According to the inspection apparatus having the above configuration, the user can obtain a first model image generated by actual imaging of a non-defective work, a second model image created from the temporarily set angle and the three-dimensional design data, By comparing and confirming the temporary angle, the angle representing the variation in the traveling direction of the reflected light from the imaging target region can be easily and accurately specified. Therefore, it is possible to accurately specify a portion (plane) that can obtain an image having brightness suitable for the inspection using the specified angle, and to easily set an appropriate number of imaging target regions. Can do.

  According to the present invention, even when a plurality of surfaces are included in the surface to be inspected or portions with various curvatures are included, it is possible to reliably obtain an image having a brightness necessary for detecting a defect for each portion. An imaging target area can be assigned. In addition, since it is possible to specifically identify a part where an image having the brightness necessary for defect detection can be obtained in each imaging target area, the imaging target is configured so that as few parts as possible are included in multiple imaging target areas. It is possible to adjust the setting position of the area, and it is possible to prevent an inspection from being missed and an excessive number of imaging target areas from being set. Therefore, both inspection efficiency and accuracy can be improved.

FIG. 1 shows a configuration example of an inspection apparatus to which the present invention is applied.
The inspection apparatus of this embodiment is for inspecting whether or not the surface of the workpiece W has a defect with the workpiece W having a plurality of surfaces as an inspection target, and is referred to as a camera 1 (hereinafter simply referred to as “camera 1”). .), An illumination device 2, a multi-axis robot 3, a robot controller 4, and a control processing device 5 using a personal computer.

  The multi-axis robot 3 (hereinafter simply referred to as “robot 3”) has a structure in which an arm unit 30 is connected to a base 36 having a predetermined size via an arm support unit 35. The arm portion 30 has a configuration in which three intermediate arms 31, 32, and 33 and a tip arm 34 are sequentially connected, and each connecting portion includes a rotation shaft (not shown). A support plate 37 is connected to the tip surface of the tip arm 34, and the camera 1 and the lighting device 2 are attached to the plate surface of the support plate 37.

  The camera 1 is a digital still camera for generating a color image. The illumination device 2 is for illuminating the workpiece W to be inspected, and includes two types of illumination units 2A and 2B described later. The robot controller 4 is a controller dedicated to robot control, and controls the rotation angle of each axis of the robot 3 in accordance with a command from the control unit 5, thereby supporting the azimuth, height, and tip of the tip arm 34. The inclination of the plate surface of the plate 37 is changed. The operation of the robot 3 makes it possible to change the position and direction of the camera 1 and the illumination device 2 as viewed from the workpiece W while maintaining the positional relationship between them.

  The alignment control of the camera 1 and the illumination device 2 and the workpiece W is not limited to this example. For example, the camera 1 and the illumination device 2 may be fixed and the position and posture of the workpiece W may be changed by a robot. Moreover, you may make it change the position and attitude | position of the camera 1 and the illuminating device 2, mounting and rotating the workpiece | work W on a rotary table.

  In addition to the CPU 50, ROM 51, and RAM 52, the control unit 5 includes a memory 53, an illumination control unit 501, an imaging control unit 502, an image input unit 503, an image processing unit 504, an inspection unit 505, a robot control unit 506, and the like. . Further, peripheral devices such as an input unit 54 and a display 55 are connected to the CPU 50. The memory 53 is a hard disk, a program for setting an inspection function in the CPU 50, setting information for imaging described later, a reference image of a non-defective workpiece, inspection parameters (binarization threshold, determination For example). The input unit 54 includes a keyboard, a mouse, and the like.

  The illumination control unit 501 selects an illumination unit to be lit for the illumination device 2, and adjusts the illumination light amount and switches illumination colors. The imaging control unit 502 causes the camera 1 to perform imaging by outputting a drive signal. The illumination control unit 501, the imaging control unit 502, and the robot control unit 506 may be independent of the control processing device 5 by incorporating the functions of these control units into a PLC (programmable logic controller).

  The image input unit 503 includes an interface circuit that inputs an image signal (generated for each color of R, G, and B) from the camera 1. The image input by the image input unit 503 is stored in the RAM 52 or the memory 53, and thereafter processed by the image processing unit 504 and the inspection unit 505. Note that the image processing unit 504 generates a difference grayscale image and binarization processing described later. The inspection unit 505 determines the presence / absence of a defect from the binarization result.

The CPU 50 outputs the control signal to the robot controller 4 via the robot control unit 506 based on the setting information for imaging registered in the memory 53. The robot controller 4 controls the operation of the robot 3 according to this control signal.
Further, the CPU 50 causes the imaging control unit 502 to output a drive signal to the camera 1 at the timing when the operation of the robot 3 is stopped. By repeating the control of the robot 3 and the driving of the camera 1, the positions and postures of the camera 1 and the illumination device 2 in the hourly imaging are changed, and the surface of the workpiece W is imaged in a plurality of regions. The distance between the camera 1 and the workpiece W is controlled so as to be always constant.

  FIG. 2 shows a detailed configuration of the optical system (camera 1 and illumination device 2) of the inspection apparatus. For convenience of drawing, the size of the illumination device 2 is greatly exaggerated with respect to the camera 1 and the optical axis of the camera 1 is shown in the vertical direction of the drawing. In addition, the surface of the workpiece W is expressed as a flat surface, the range of the imaging target region R of the camera 1 (the region where the field of view of the camera 1 is matched), and the center point C (hereinafter referred to as “imaging center point” Or simply “center point”).

  The illumination device 2 incorporates a half mirror 20, a first illumination unit 2A for coaxial incident light, and a second illumination unit 2B for oblique incidence illumination. The half mirror 20 is provided on the optical axis of the camera 1, the first illumination unit 2 </ b> A is disposed on the side of the half mirror 20, and the second illumination unit 2 </ b> B is disposed below the half mirror 20.

  The first illumination unit 2A for coaxial epi-illumination has R, G, B in a housing 23 provided with a circular opening (opening toward the left in the figure) from which light is emitted. The light sources 21R, 21G, and 21B (specifically, LEDs) that emit the respective color lights are accommodated. Each of the light sources 21R, 21G, and 21B is provided with its optical axis directed toward the half mirror 20. A diffusion plate 22 is provided in the opening.

  The second illuminating unit 2B for oblique incidence illumination has a main body unit that is a casing 25 in which a viewing hole 26 of the camera 1 is formed on the upper surface. The inside of the housing 25 is divided into three regions s, t, and u arranged concentrically, and each of the regions s, t, and u has a plurality of light sources 21R, 21G, and 21B that are the same as the first illumination unit. It is arranged concentrically over the entire area. Both light sources are wired to a substrate 27 in the housing 25 so that the optical axis is parallel to the optical axis of the camera 1.

A lower portion of the housing 25 is opened, and a light diffusion member 23 is provided in the opening. Further, the center of the peep hole 26 is aligned with the optical axis of the camera 1.
The light diffusion member 23 has a configuration in which three inclined surfaces 23s, 23t, and 23u having sizes corresponding to the regions s, t, and u are continuous. Each inclined surface 23s, 23t, 23u descends from the inside toward the outside, and the inclination becomes steeper as the inclined surface is farther from the peephole.

  In the first illumination unit 2A, the light emitted from the light sources 21R, 21G, and 21B is mixed by the diffusion plate 22 to generate light having a circular cross section. This light travels along the optical axis of the camera 1 after reaching the half mirror 20. Therefore, when the first illumination unit 2A is turned on, so-called coaxial incident illumination light is irradiated to the imaging target region R.

  On the other hand, in the second illumination unit 2B, the light emitted from the light sources 21R, 21G, and 21B is mixed in the regions s, t, and u of the housing 25, and further below the peephole 26 through the light diffusion member 23. Is emitted. Therefore, when the 2nd illumination part 2B lights, the illumination light which injects from the diagonal direction is irradiated with respect to an imaging object area | region.

  Note that the illumination units 2A and 2B can change the color of the illumination light by selecting the color of the light source to be lit. Further, in the second illumination unit 2B, any one of the three regions s, t, and u or two adjacent regions are selected, and only the light source in the selected region is turned on, thereby changing the illumination direction. It can be limited.

  In this embodiment, when inspecting for the presence or absence of irregularities on the surface of the workpiece W, imaging is performed under coaxial epi-illumination by the first illumination unit 2A. On the other hand, when inspecting the presence or absence of a color defect, imaging is performed under oblique incidence illumination by the second illumination unit 2B.

  Of the above two types of inspection, the inspection of uneven defects uses a phenomenon in which the traveling direction of specular reflection light changes due to the presence of defects. It is necessary to adjust the imaging direction in accordance with the direction of the workpiece surface so that it is incident on 1. In this embodiment, the optical axis of the camera 1 in hourly imaging corresponds to the normal direction of the imaging center point C (specifically, the direction in which the vector indicating the normal direction of the center point C is inverted and the camera 1 The position and orientation of the camera 1 are adjusted so that the vector indicating the imaging direction of the camera 1 coincides with the vector.

  FIG. 3 shows the coaxial epi-illumination light at the imaging center point C and the two points F1 and F2 that are separated from the imaging center point C by a predetermined distance among the points on the workpiece surface included in the imaging target region R. 2 shows a state in which specularly reflected light is incident on the lens 11 of the camera 1 or is directed in a direction not incident on the lens 11. In addition, the dotted line in a figure shows the range of the illumination light which goes to the points C, F1, and F2. The halftone dot pattern is a range in which specular reflection light with respect to the illumination light travels (range of light reflected in the specular specular direction with respect to each illumination light).

  In the example of FIG. 3, all the specular reflection light from the imaging center point C is incident on the lens 11, but only a part of the specular reflection light from the point F <b> 1 slightly away from the center point C is incident. . Further, when the point F2 is further away from the center point C than the point F1, the specularly reflected light is not incident on the lens 11 at all.

  Thus, even if it is a site | part included in the visual field of the camera 1, depending on the inclination angle, a part where regular reflection light cannot be incident on the camera 1 like the point F2, or a part like the point F1 There is a part where only specularly reflected light can enter. However, the distribution range of specularly reflected light from one point of the workpiece W is not limited to the range shown in FIG. 3 and is scattered around the periphery with a certain degree of variation. May be secured.

  Therefore, the range (inspectable range) for obtaining an image with the brightness necessary for detecting irregularities is not only the relationship between the curvature of the surface to be inspected and the theoretical regular reflection direction, but also the variation in the traveling direction of scattered light. It is thought that it varies depending on the degree.

  By the way, when inspecting uneven defects by multiple imaging for polyhedral workpieces containing multiple parts with different curvatures, it is possible to inspect any part of the imaging target area without omission. The imaging target area must be allocated so as to be included in the range. If the imaging target area can be determined by correctly recognizing the inspectable range during this allocation process, it is considered that the imaging target area is not allocated more than necessary, and the inspection efficiency can be improved.

  However, it is not easy to recognize the degree of curvature in the imaging target region set at an arbitrary position. In addition, scattering of specularly reflected light on the work W is considered to be caused by minute irregularities on the surface of the work or materials contained in the paint (for example, metal pieces contained in the paint of the metallic work). It is very difficult to determine if there is scattering. Therefore, it is not easy for the user at the site who sets the imaging target area to determine which range of the imaging target area is the inspection possible range.

  Therefore, in this embodiment, a plurality of non-defective models of the workpiece W and CAD data in the STL format created for the workpiece W design are used to check a plurality of inspection surfaces of the workpiece W while checking the inspectable range. The imaging target areas are allocated, and setting information necessary for imaging these imaging target areas is registered in the memory 53.

  The CAD data in the STL format represents the surface of an object by approximating it to an aggregate of a plurality of triangular planes. Each triangular plane is a vector (hereinafter referred to as “normal line” representing the coordinates of three vertices and the direction of the normal. It is expressed as “vector”.)

The normal vector is represented by two angles α and β as shown in FIG.
The angle α indicates the azimuth of the projection image VL ′ when the normal vector VL is projected onto the XY plane of the spatial coordinate system based on CAD data. From the positive direction of the Y axis to VL ′ in the clockwise direction. It is represented by the angle. The angle β is an angle from the XY plane to the normal vector VL. Hereinafter, the angle α is referred to as “azimuth angle α”, and the angle β is referred to as “elevation angle β”.

  The setting information for imaging is configured by the position of the camera 1 and the imaging direction. The position of the camera 1 (hereinafter referred to as “camera position”) is specifically represented by the X, Y, Z coordinates of the center point of the camera lens 11. The imaging direction is originally expressed as a direction from the camera 1 along the optical axis toward the workpiece W. In this embodiment, for convenience, the normal vector VL (specifically, the azimuth angle α) of the imaging center point C is used. And elevation angle β).

  When registering the setting information, in this embodiment, an image representing a three-dimensional model of the work W based on CAD data is created and displayed on the display 55, and the user is allowed to designate a predetermined position on the display image. The designation of the imaging center point C is accepted. The CPU 50 sets the normal vector VL of the triangular plane including the point designated here as information indicating the imaging direction. In addition, the coordinates of a point separated from the designated point by a predetermined distance along the direction indicated by the normal vector VL are set as the camera position.

  Further, in this embodiment, a triangular plane corresponding to the imaging target area when the camera 1 is aligned with the workpiece W based on the camera position and the imaging direction is extracted. Further, an inspectable range in the imaging target region is recognized by a method of identifying a plane in which the inclination angle with respect to the imaging center point C is within a predetermined allowable value from among these triangular planes. In addition, by setting a value in consideration of the above-described variation degree of scattered light as the allowable value, the recognition accuracy of the inspectable range can be improved and the number of imaging can be reduced.

  FIG. 5 shows a series of processes for setting the camera position and the imaging direction. Hereinafter, the contents of the setting process will be described along the flow of FIG. 5 with reference to FIGS.

In the first ST1 (ST is an abbreviation of “step”, the same applies hereinafter), the acceptance of a non-defective product model of the workpiece W (hereinafter referred to as “non-defective workpiece”) is accepted.
In the next ST2, the size of the field of view of the camera 1 is determined by receiving the input of the resolution necessary for the inspection. For example, if the number of pixels of the image sensor is 1000 pixels × 1000 pixels and the input resolution is 0.01 mm / pixel, the field of view of the camera 1 is 10 mm square. The size of the imaging target area is also the same as this visual field size.

  In the next ST3, an initial value 1 is set to a counter n indicating the order of registration. In the next ST4, the image of the above-described three-dimensional model is displayed, and designation of the imaging center point C by the user is accepted. In ST5, the camera position and the imaging direction are set based on the imaging center point C designated in ST4.

  In the next ST6, the position and direction of the camera 1 are actually adjusted based on the camera position and imaging direction set in ST5, and a non-defective workpiece is imaged. In ST7, the image generated by this imaging is temporarily stored in the RAM 52 as the first model image.

Next, in ST8, a triangular plane corresponding to the imaging target area based on the above setting is extracted using CAD data.
FIG. 6 shows an example of this extraction. In this figure, among the triangular plane group representing the workpiece W, the plane that can be recognized from the currently set imaging direction is indicated by a solid line, and the set imaging target region R is indicated by a dotted frame. Further, a halftone dot pattern is added to the triangular plane extracted in ST8. As indicated by the range of the halftone dot pattern, in ST8, only the triangular plane that is completely included in the imaging target region R is extracted.

  Next, in ST9, a normal direction histogram is created using the extracted normal vector of the plane (hereinafter, this histogram is referred to as "normal direction histogram"). Specifically, as shown in FIG. 7, a histogram in which the frequency is associated with the combination of the azimuth angle α and the elevation angle β indicating the normal vector VL of each triangular plane is created. In this example, the angles α and β are expressed in units of 5 °. However, the present invention is not limited to this, and may be expressed in units of 1 °, for example.

When the normal direction histogram is created as described above, in the next ST10, a process of setting an allowable value for determination is executed (details of this process will be described later).
When the allowable value is set, in next ST11, the normal direction in which the angle difference with respect to the normal direction of the imaging center point C is within the allowable value set in ST10 is extracted from the normal direction histogram. Furthermore, in ST12, the sum total of the frequencies in the normal direction extracted is obtained, and the sum is compared with a predetermined threshold value. If the sum of the frequencies is equal to or greater than the threshold value, ST13 becomes “YES” and the process proceeds to ST14.

In ST14, a detected flag is set on each triangular plane corresponding to the extracted normal direction.
This detected flag indicates that the current imaging target area is a plane corresponding to a part capable of generating an image having the brightness necessary for detecting the irregular defect, that is, a part included in the inspectable range. Is. That is, portions where the inclination angle with respect to the imaging center point C is within an allowable value are extracted by ST11 and ST12, and these portions are recognized as an inspectable range by ST14.

  In ST15, the camera position and imaging direction set in ST5 are registered in the inspection data file in the memory 53 as the nth imaging setting information. In ST16, the first model image stored in ST7 is registered in the inspection data file as the nth reference image.

  In next ST17, it is determined whether or not the detected flag is set in all the triangular planes corresponding to the surface to be inspected. If there is a triangular plane for which the detected flag is not set, the determination in ST17 is “NO” and the process proceeds to ST18, where the value of n is incremented. Further, in ST19, a screen showing a part where the detected flag is not set is displayed (for example, the same three-dimensional model image displayed when the imaging center point C is specified is displayed on the triangular plane where the detected flag is not set. The corresponding part is marked with a predetermined color and displayed). Thereafter, the process returns to ST4, and designation of a new imaging center point C is accepted using the currently displayed screen. At this stage, only the designation for a plane for which the detected flag is not set may be valid.

  If the sum of the frequencies in the normal direction extracted from the normal direction histogram is below the threshold value for the imaging target region set based on the imaging center point C specified in ST4, the determination in ST13 is “NO”. Then, the process proceeds to ST20, and it is notified by displaying a message or the like that imaging according to the current setting is not suitable for inspection. Thereafter, the process proceeds to ST19 without performing the processing of ST14 to 18, so that information relating to imaging that does not conform to this inspection is not registered. Therefore, it is possible to prevent a wasteful process from being performed by setting the imaging target region in a range where the inspectable range is extremely small.

  Hereinafter, by repeating the above processing until the detected flag is set on all the triangular planes corresponding to the surface to be inspected, a plurality of image capturing target areas are allocated to the surface to be inspected, and these image capturing target areas are imaged. A camera position and an imaging direction necessary for the image acquisition, and a reference image of the imaging target area are registered. When the detected flag is finally set for all the triangular planes, ST17 becomes “YES”, the process proceeds to ST21, and the current value of n is stored as the registered total number N. Furthermore, the non-defective workpiece is unloaded at ST22 and the process is terminated.

  According to the above processing, since settings relating to imaging every hour are performed using CAD data while specifying a part included in the inspectable range, it is easy to set so that all of the inspected surface can be inspected without omission. It becomes possible to do.

Next, the allowable value setting process in ST10 will be described in detail.
In this embodiment, as described above, the angle φ representing the spread range in the direction in which the reflected light having the intensity necessary for detecting the concavo-convex defect proceeds in consideration of the magnitude of the scattered light variation due to the reflection characteristics of the workpiece W. And an allowable value is set based on the angle φ.

FIG. 8 shows a scattering state of reflected light based on a Gaussian scattering model. A vector K in the figure indicates a theoretical regular reflection direction with respect to illumination light. The intensity of the specularly reflected light traveling in the direction of the vector K is P ₀ , the intensity of the reflected light reflected in the direction shifted by an angle θ from the vector K is P (θ), and the standard deviation of the traveling direction of each reflected light is σ. Then, P (θ) is expressed by the following equation (1).

FIG. 9 shows the deviation angle θ with respect to the theoretical specular reflection direction (the direction of the vector K) and the angle θ when the standard deviation σ is 5 degrees and 30 degrees, respectively, based on the above equation (1). Is a graph showing the relationship with the intensity of specularly reflected light in the direction of. The value of the specular reflection light intensity is normalized with the theoretical intensity in the specular reflection direction being “1”.
As shown in the graph, when the variation in the traveling direction of the scattered light increases, the range in which relatively strong reflected light can be obtained also increases.

  For example, in FIG. 9, when the range in which the relative intensity of the regular reflection light is 0.5 or more is confirmed, when the standard deviation σ is 5 degrees, the direction is shifted by about 5 to 6 degrees from the theoretical regular reflection direction. If it is not within the range, regular reflection light satisfying this condition cannot be obtained. On the other hand, when the standard deviation σ is 30 degrees, regular reflected light satisfying the above conditions can be obtained in a range up to about 35 degrees from the theoretical regular reflection direction.

Accordingly, assuming that the spread angle of the range in which the reflected light having the intensity necessary for detecting the irregular defect travels is φ, in theory, if the tilt angle with respect to the imaging center point C is within φ, it is within the above range. It is considered that light can be incident on the camera, and it is possible to detect uneven defects.
Therefore, in this embodiment, it is assumed that the distribution state of scattered light in this area approximates a Gaussian distribution for each set imaging target area, and an image of a non-defective work generated by the above-described imaging in ST6 ( A value of the standard deviation σ is obtained using a first model image) and a simulation image (second model image) created based on CAD data. Further, the value of σ is applied to the equation (1) to calculate the intensity of specularly reflected light in each direction, and the angle representing the size of the range in which the calculated intensity is a value necessary for detecting uneven defects is described above. Obtained as an angle φ, and based on this angle φ, the allowable value used in ST11 is determined.

FIG. 10 shows the detailed procedure of the allowable value setting process corresponding to ST10 of FIG.
In this process, first, in ST101, a predetermined initial value (for example, 5 degrees) is set as the standard deviation σ in the above equation (1).
Next, in ST102, with respect to the imaging target area according to the current setting (the setting in ST5 in FIG. 4), the above-described standard deviation σ is set on each triangular plane (extracted in ST8) corresponding to this area. As a result, the intensity of the light incident on the image sensor of the camera 1 is calculated for each pixel. Then, a second model image is generated using each calculated intensity.

  In the next ST103, a screen in which the first model image generated by imaging the non-defective workpiece W (temporarily stored in the RAM 52 in ST7) and the second model image is displayed in parallel is displayed on the display 55. increase. FIG. 11 shows an example of this display screen. In the figure, the “real image” on the left is the first model image, and the “simulation image” on the right is the second model image. Further, on this screen, a graph representing the luminance distribution along the horizontal direction of the image is displayed below each image, change buttons 31 and 32 for raising and lowering the set value of the standard deviation σ, and this setting A confirmation button 33 for confirming the value is provided.

In this embodiment, when the brightness distribution in the second model image obtained by the simulation is close to the first model image obtained by imaging the actual non-defective workpiece, the actual standard deviation of the scattered light is also set to the set value of σ. Let the user specify the value of σ, assuming that it is close.
Specifically, when the user changes the set value of σ by operating the change button 31 or 32, ST104 becomes “YES” and the process proceeds to ST105, and a new second model image is displayed based on the changed σ value. Create and update the second model image in the display screen with the new image. If the confirm button 33 is operated at a predetermined time, ST106 becomes “YES” and the process proceeds to ST107. Based on the formula (1) based on the value of σ at the confirming operation, regular reflection of intensity required for the inspection is performed. A spread angle φ in the direction in which light travels is obtained.

  Further, in ST108, an allowable value is determined based on the specified value of the angle φ. In this process, the angle φ is not set as the allowable value as it is, but a value obtained by subtracting a predetermined adjustment value from φ is set as the allowable value in consideration of the sensitivity of the camera 1 and the like. The magnitude of the adjustment value is desirably changed according to the curvature of the workpiece W in the imaging target area.

  According to the methods shown in FIGS. 10 and 11, assuming that the degree of variation (standard deviation σ) in the traveling direction of scattered light becomes a predetermined value, a simulation image based on this assumption and an actual image are displayed to the user. By collating, it is possible to easily specify the degree of variation. Further, if the value of σ can be changed in fine units, a numerical value that can be specified can be made highly reliable.

  Therefore, by using this specified value of σ, an angle φ indicating the degree of variation in the direction in which the reflected light with the intensity necessary for detecting the irregularity defect travels is specified, and an allowable value is set with the value of φ. Thus, it is possible to accurately recognize the inspectable range in the imaging target region.

  However, the process for obtaining the angle φ is not limited to the above example. For example, the areas s, t, and u of the light source 2B for oblique incidence illumination shown in FIG. The value of φ may be derived from the range of the illumination direction when an image having the brightness necessary for the above is obtained.

If the variation in curvature of the surface of the workpiece W is small, measurement outside the inspection apparatus may be performed using a goniophotometer as shown in FIG.
In the example of FIG. 12, the intensity of the reflected light in each direction is measured while changing the direction of the light receiver 102 with respect to the work W while irradiating the work W with the projector 101 arranged in a certain direction. . In this case, the width of the azimuth change in which reflected light having an intensity of a predetermined value or more can be received is specified as the angle φ.

  Next, FIG. 13 shows a processing procedure (related to the inspection of one workpiece W) in the case of inspecting the concavo-convex defect based on the set camera position and imaging direction after completing the setting process shown in FIG. Show.

  First, in ST201, a transport mechanism (not shown) is driven to transport the workpiece W to a predetermined inspection position. Note that the posture of the workpiece W at the time of loading is adjusted so as to be always in a predetermined state.

  Next, in ST202, the counter n is set to an initial value of 1. In ST203, the nth registration data (the camera position and imaging direction in the nth imaging, and the nth reference image) are read out.

  In ST204, the operation of the robot 3 is controlled based on the read camera position and imaging direction, and the position and imaging direction of the camera 1 are adjusted to match the read data. In ST205, the illumination device 2 is set in the coaxial epi-illumination state, and the camera 1 performs imaging.

In ST206, a difference calculation is performed on the inspection image generated by the above imaging with the reference image to generate a difference grayscale image indicating a difference in brightness with respect to the reference image. In this embodiment, since a color image is generated, the difference calculation process is specifically performed for each color data of R, G, and B, and a grayscale image is created by the average value of the differences for each color data. The
Furthermore, in ST207, a significant difference is detected by binarizing the difference grayscale image. In ST208, the presence or absence of a defect is determined using the binarized image.

  When the above determination is completed, the determination result is temporarily stored in the RAM 52, and then the process proceeds from ST209 to ST210. After updating the counter n, the process returns to ST203. Thereafter, by executing the same processing until the counter n reaches the total number N of registrations, the imaging target region is sequentially imaged, and processing for determining the presence / absence of an uneven defect is performed.

  When the determination for all the imaging target regions is completed, the process proceeds to ST211 and inspection result data in which the determination processing results are integrated is created and output to the display 55 or an external device (not shown). Further, in ST212, the work W that has been inspected is carried out, and the process is terminated.

In the example of FIG. 13 described above, a region different from the reference image is detected in the entire generated image without considering the inspectable range in each imaging target region. If it can be reliably detected, there will be no particular problem in this way. This is because even if there is a defect outside the inspectable range and the defect cannot be detected, the defect can be detected by processing the imaging target region including the portion corresponding to the defect in the inspectable range.
In this embodiment, in the setting process, the imaging target area is allocated so that each part of the surface to be inspected is always included in the inspectable range of any imaging target area. Therefore, by performing the process of FIG. Thus, it is possible to detect irregularities on the entire surface to be inspected without omission.

  In the above embodiment, the user designates the imaging center point C and sets the camera position and the imaging direction based on the designation. However, the present invention is not limited to this, and the imaging center point C is automatically set using CAD data. Alternatively, it can be set semi-automatically. For example, create a normal direction histogram for the entire surface to be inspected of the workpiece, focus on the peaks in this histogram in descending order of frequency, and set the imaging center point on the plane in the plane group corresponding to the peak of interest. A method of setting C is conceivable.

  In the concavo-convex defect inspection of the above embodiment, the camera position and the imaging direction are set so that strong reflected light spreading around the theoretical specular reflected light direction is incident on the camera 1. If the degree of light scattering increases, even if the direction of regular reflection changes due to a defect, a considerable amount of reflected light may enter the camera 1 and the sensitivity to the defect may deteriorate. In such a case, even if the method of illumination or imaging is changed so that strong reflected light including light traveling in the theoretical regular reflection direction escapes from the camera 1 in the range in which the regular reflected light travels. Good. For example, oblique illumination is performed by the illumination unit 2B instead of coaxial incident illumination, and light that travels in a direction deviating from the theoretical regular reflection direction out of the regular reflection light with respect to the illumination light is incident on the camera 1. Can respond.

It is a figure which shows the external appearance and electrical structure of a one part mechanism of a test | inspection apparatus. It is explanatory drawing which shows the structure of the optical system of a test | inspection apparatus. It is explanatory drawing which shows a mode that the range of progress of the regular reflection light from a workpiece | work changes with the inclination angles of a reflective position. It is explanatory drawing which shows the information showing a normal vector. It is a flowchart which shows the flow of the setting process for imaging. It is explanatory drawing which shows the example of extraction of the triangular plane in an imaging object area | region. It is explanatory drawing which shows the example of a setting of a normal direction histogram. It is explanatory drawing which shows the distribution state of the reflected light by a Gaussian scattering model. It is the graph which showed the relationship between each direction and the intensity | strength of reflected light based on a Gaussian scattering model. It is a flowchart which shows the flow of an allowable value setting process. It is explanatory drawing which shows the example of the screen for collation. It is explanatory drawing which shows the other example of the identification method of angle (phi). It is a flowchart which shows the flow of the process in a test | inspection.

Explanation of symbols

W Work C Imaging center point R Imaging target area 1 Camera 2 Illumination device 3 Robot 4 Robot controller 5 Control processing device 2A Illumination unit for coaxial epi-illumination 50 CPU

Claims (6)

Taking an object whose surface to be inspected as a part or all of the surface as an inspection target, imaging the inspection surface in multiple times while changing the position and direction of the imaging device and the illuminating device with respect to the inspection target, In a method of registering information related to setting of an imaging target area in imaging every hour in advance in an inspection apparatus that inspects the state of the surface to be inspected using a regular reflection light image in a generated image,
Setting an imaging target region at an arbitrary position on the surface to be inspected;
Regarding light reflected at an intensity of a predetermined value or more when the positions and directions of the imaging device and the illuminating device are determined so that specularly reflected light from the center position of the imaging target region set in step A is incident on the imaging device Step B for obtaining an angle representing the degree of variation in the traveling direction;
Using the three-dimensional design data representing the surface of the object as an aggregate of a plurality of planes, a difference in inclination with respect to the plane at the center position of the area from among the planes corresponding to the imaging target area set in Step A is obtained. Specifying a plane C corresponding to the range of angles determined in step B;
A plurality of cycles of the above steps are performed to identify all the planes corresponding to the surface to be inspected, and information relating to the setting of the imaging target area set in the hourly step A is registered.
An information registration method for surface condition inspection. In step B,
For the non-defective model of the inspection object, the field of view of the imaging apparatus is matched with the imaging target area set in step A, and specularly reflected light from the center position of the imaging target area is incident on the imaging apparatus. A first model image is generated by determining and capturing the positions and directions of the imaging device and the illumination device,
Assuming an angle value representing the degree of variation in the traveling direction of the reflected light from the imaging target region, using this assumed angle and the three-dimensional design data, the imaging target region set in step A is used. Generating a second model image;
The first and second model images are displayed in a collatable manner, and the second model image is updated in accordance with an operation for changing the assumed angle.
When the confirmation operation for the display of each model image is performed, the calculation is performed according to a theoretical formula with the assumed angle at that time as the range of scattering, and the traveling direction of the light reflected at the intensity of the predetermined value or more is calculated. The information registration method for surface condition inspection according to claim 1, wherein an angle representing a degree of variation is specified.   In the step B, for the non-defective model of the inspection object, the field of view of the imaging device is matched with the imaging target region set in the step A, and the direction of illumination by the lighting device is switched to perform multiple imaging. The angle representing the degree of variation in the traveling direction of the light reflected at the intensity equal to or greater than the predetermined value is specified based on the illumination direction in imaging that is executed and an image having a brightness equal to or greater than the predetermined value is obtained. Information registration method for inspection of surface condition.   In the step C, a histogram indicating a distribution state in a normal direction of each plane corresponding to the imaging target region is created using the three-dimensional design data, and a method corresponding to the central plane of the imaging target region is generated from the histogram. The normal direction in which the angle difference with respect to the line direction is included in the range of the allowable value based on the angle obtained in the step B is extracted, and a plane corresponding to each extracted normal direction is specified. Information registration method for surface condition inspection. An imaging device for imaging an object to be inspected using an object whose part or the entire surface is a surface to be inspected, an illuminating device for illuminating an imaging target area of the imaging device, and the imaging device And an imaging device for the inspection object by moving at least one of the first support mechanism for supporting the illumination apparatus, the second support mechanism for supporting the inspection object, and the first and second support mechanisms, and A process of changing the position and direction of the illuminating device and causing the imaging device to image a predetermined imaging target area of the surface to be inspected is executed a plurality of times, and the surface to be inspected is used using an image generated by imaging every hour And a control processing device that executes processing for inspecting the surface condition of
A display unit for displaying a model image of each imaging target region; and an input unit for receiving a user operation;
The control processing device includes:
Storage means for registering information regarding the setting of the imaging target area;
Area setting means for setting an imaging target area at an arbitrary position on the surface to be inspected;
For the non-defective model of the inspection object, the field of view of the imaging apparatus is matched with the imaging target area set by the area setting means, and specularly reflected light from the center position of the imaging target area is incident on the imaging apparatus. Determining a position and a direction of the imaging device and the illumination device, and causing the imaging device to perform imaging, thereby generating a first model image of the imaging target region;
Assuming a predetermined value as an angle representing the degree of variation in reflected light from the imaging target region set by the region setting means, the assumed angle and the surface of the inspection target are represented as an assembly of a plurality of planes. Second image generation means for generating a second model image of the imaging target area using the three-dimensional design data obtained;
The first and second model images are displayed on the display means so as to be collated, and when the input means accepts an operation for changing a hypothetical angle, the second model image is based on the angle changed by the operation. Display control means for causing the image generation means to update the second model image and updating the display of the display means with the updated image;
When the input means accepts a definite operation under the display state of each model image, the calculation is performed with a theoretical formula with the assumed angle at the time of the operation as the range of scattering, and imaged with an intensity of a predetermined value or more Angle specifying means for specifying an angle representing the degree of variation in the traveling direction of reflected light incident on the apparatus;
Using the three-dimensional design data, a difference in inclination with respect to the plane of the center position of the area from the plane corresponding to the imaging target area set by the area setting means is based on the angle specified by the angle specifying means. A plane identifying means for identifying a plane included within the allowable range;
Regarding a plurality of imaging target areas set during a plurality of cycles of processing by each of the area setting means, first image generation means, second image generation means, display control means, angle specification means, and plane specification means A surface condition inspection apparatus comprising registration means for registering information relating to the setting of each region in the storage means.   The plane specifying unit creates a histogram indicating a distribution state in a normal direction of each plane corresponding to the imaging target area using the three-dimensional design data, and corresponds to a central plane of the imaging target area from the histogram. The surface condition inspection apparatus according to claim 5, wherein a normal direction in which an angle difference with respect to a normal direction is included in the range of the allowable value is extracted, and a plane corresponding to each extracted normal direction is specified.

Images