JP7207948B2 - Appearance inspection method and program - Google Patents

Description

The present invention relates to an appearance inspection method and program for inspecting the appearance of an object to be measured based on an image obtained by imaging the object to be measured.

An image measuring machine captures an image of an object to be measured (hereinafter referred to as "workpiece"), analyzes the image, extracts edge point groups contained in the image, and extracts edge points from the extracted edge point group. It is a device that evaluates the distance, inclination, diameter, width, etc. of approximated geometric shapes such as straight lines, circles, and polygons. When obtaining a geometric shape from the extracted edge point group by approximation, if the edge point group that is the basis of the approximation contains points that are greatly separated from other points, the actual workpiece shape may To obtain a fitted geometry, a process of excluding outliers to find the geometry is often performed. The geometrical shape obtained by approximation is evaluated by comparing it with an ideal geometrical shape determined by design data or the like.

JP 2011-185888 A

In this way, the conventional image measuring machine evaluates the geometric shape obtained by performing approximation while ignoring abnormal points. In addition, it could not be applied to the so-called appearance inspection focusing on the defective portion deviating from the approximated geometric shape.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to provide a visual inspection method that can be performed by a vision measuring machine, and a program for realizing such a visual inspection method.

(1) In order to solve the above problems, a visual inspection method according to the present invention includes a target area specifying step of specifying a target area in an inspection target image, a defect detection step of detecting a defect contained in the specified target area, and a detection a defect counting step for counting the number of defects detected; and a judgment step for judging acceptance/rejection of the object appearing in the image to be inspected based on the counted number of defects.

(2) In the present invention, the defect detection step preferably includes a step of extracting edges included in the target region, and a step of comparing the extracted edges with edges in a master image to detect defects.

(3) When a visual inspection method is applied to a plurality of images to be inspected, the master image may be the first image to which the visual inspection method is applied among the plurality of images to be inspected. (4) Further, the master image may be an image to which the visual inspection method has been applied immediately before among the plurality of images to be inspected. (5) Alternatively, the master image may be an image, among the plurality of images to be inspected, that has been determined to pass as a result of applying the appearance inspection method immediately before.

(6) A program according to the present invention for solving the above problems causes a computer to execute any one of the appearance inspection methods described above.

1 is a perspective view showing the configuration of an image measuring machine 100. FIG. 2 is a schematic diagram showing the configuration of an imaging unit 120 together with a stage 100; FIG. 2 is a block diagram showing the configuration of position acquisition means 110. FIG. 2 is a block diagram showing the configuration of a computer main body 141; FIG. FIG. 10 is a diagram showing an example of screen display; It is a figure which shows an example of a visual inspection screen. 4 is a flow chart showing the procedure of visual inspection. It is a figure which shows an example of a visual inspection screen. It is a figure which shows an example of a visual inspection screen. It is a figure which shows an example of a visual inspection screen.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a perspective view showing the configuration of the image measuring machine 1. As shown in FIG. The image measuring machine 1 includes a stage 100 , a position acquisition means 110 , an imaging unit 120 , a remote box 130 and a computer system 140 .

The stage 100 is arranged so that its upper surface is a horizontal surface, and a work (an object to be measured or inspected) W is placed on the upper surface. At least a portion of the upper surface of the stage 100 on which the workpiece W is placed is made of a material such as glass that transmits light. The stage 100 is driven by an X-axis drive motor and a Y-axis drive motor (not shown) and is movable in the X-axis direction and the Y-axis direction parallel to the horizontal plane. A drive control signal for the drive motor of each axis is given to the drive motor of each axis from a remote box 130 or computer system 140, which will be described later.

FIG. 2 is a schematic diagram showing the configuration of the imaging unit 120 together with the stage 100. As shown in FIG. Imaging unit 120 includes optical system 122 , imaging means 124 , and light source 126 . The optical system 122 comprises, for example, a telecentric optical system by combining a plurality of lenses and diaphragms. In a telecentric optical system, the principal ray can be regarded as parallel light, so the dimension in the captured image does not depend on the position in the Z-axis direction (height direction). Therefore, it is suitable for measuring a work W having undulations (for example, steps, holes, etc.). The light source 126 irradiates at least the imaged portion of the work W with light under the control of the computer system 140 when the image of the work W is captured. In this embodiment, a light source 126a for epi-illumination that irradiates the work W with light from above (that is, the imaging means 124 side) through the optical system 122, and light from below the work W (that is, the back side of the stage 100). A light source 126b for transmitted illumination is provided. The imaging means 124 is, for example, a two-dimensional image sensor such as CCD or CMOS. An image of the workpiece W is formed on the light receiving surface of the imaging means 124 by the optical system 122 . The imaging means 124 captures the formed image and outputs image data in a predetermined format. This image data includes at least an index indicating the order in which the images are captured, in addition to information on the pixels forming the image. The imaging unit 120 transmits image signals output by the imaging means 124 to the computer system 140 . The computer system 140 and the imaging unit 120 are connected by a general-purpose communication standard such as USB (Universal Serial Bus). Further, the imaging unit 120 outputs a trigger signal to the latch means 118 at the timing when imaging of one image (one frame) is completed.

The imaging unit 120 is driven by a Z-axis driving motor (not shown) and is movable in the Z-axis direction (that is, the direction perpendicular to the upper surface of the stage 100). Focus adjustment is performed by adjusting the position of the imaging unit 120 in the Z-axis direction. A drive control signal for the Z-axis drive motor is given from a remote box 130 or a computer system 140, which will be described later.

FIG. 3 is a block diagram showing the configuration of the position acquisition means 110. As shown in FIG. The position acquisition means 110 comprises an X-axis encoder 112 , a Y-axis encoder 114 , a Z-axis encoder 116 and a latch means 118 .

The X-axis encoder 112 measures and outputs the position coordinates of the stage 100 in the X-axis direction. A Y-axis encoder 114 measures and outputs the position coordinates of the stage 100 in the Y-axis direction. The Z-axis encoder 116 measures and outputs the position coordinates of the imaging unit 120 in the Z-axis direction. Each encoder has a graduated scale and a scale reading unit that reads the graduations of the scale. The scales are attached to movable parts of the stage 100 and imaging unit 120 along each axis. On the other hand, the scale reader is arranged on the non-movable part.

The latch means 118 comprises a counter 118a and a buffer 118b. The counter 118a increases the count value by one when a trigger signal (for example, a pulse signal) is supplied from the outside. Note that the value of the counter 118a is appropriately reset based on instructions from the computer system 140. FIG. The buffer 118b has storage areas of a plurality of addresses, and latches and stores the output values of the encoders of each axis in the storage areas of addresses corresponding to the count values of the counter 118a at the timing when the trigger signal is supplied. The trigger signal may be supplied from the imaging means 124, for example, at the timing when imaging of one image is completed. The position coordinates of each axis held by the latch means 118 are associated with address values (that is, count values) and are taken into the computer system 140 as appropriate. The computer system 140 and the latch means 118 are connected by a general-purpose communication standard such as USB (Universal Serial Bus). The image data and the positional coordinates are separately fetched into the computer system 140, but the image data is assigned an index indicating the order of imaging, and the positional coordinates are assigned a count value indicating the order of imaging. 140, it can be associated after import.

Returning to FIG. 1, the remote box 130 is an operation means for setting the positions of the stage 100 and the imaging unit 120, and according to the operation by the operator, the X-axis drive motor, Y-axis It transmits drive control signals for the drive motor and the Z-axis drive motor. The remote box 130 has a joystick 132 and a jog shuttle 134 . The joystick 132 is an operation input means for setting the position of the stage 100, and the remote box 130 moves the stage 100 in the X-axis direction and the Y-axis direction according to the tilt direction of the joystick 132. Send a drive control signal for The jog shuttle 134 is operation input means for setting the Z-axis direction position of the imaging unit 120, and the remote box 130 controls the imaging unit 120 in accordance with the rotation direction, rotation amount, rotation speed, and the like of the jog shuttle 134. in the Z-axis direction.

The computer system 140 comprises a computer main body 141 , a keyboard 142 , a mouse 143 and a display 144 . FIG. 4 is a block diagram showing the configuration of the computer main body 141. As shown in FIG. The computer main body 141 includes a CPU 40 that forms the center of control, a storage unit 41, a work memory 42, interfaces (shown as "IF" in FIG. 4) 43 and 44, and display control for controlling display on the display 144. a portion 45;

The operator's instruction information input from the keyboard 142 or the mouse 143 is input to the CPU 40 via the interface 43 . The interface 44 is connected to the imaging unit 120 and the stage 100, supplies various control signals from the CPU 40 to the imaging unit 120 and the stage 100, and receives various status information and measurement results from the imaging unit 120 and the stage 100. Input to the CPU 40 .

The display control unit 45 displays the image captured by the imaging unit 120 on the display 144 . In addition to the image captured by the imaging unit 120, the display control unit 45 displays on the display 144 an interface for inputting a control instruction to the image measuring machine 1, an interface for a tool for analyzing the captured image, and the like. do.

A work memory 42 provides a work area for various processes of the CPU 40 . The storage unit 41 is configured by, for example, a hard disk drive, RAM, etc., and stores programs executed by the CPU 40, image data obtained by imaging with the imaging unit 120, and the like.

The CPU 40 controls the imaging unit 120, the X-axis drive motor, the Y-axis drive motor, the Z By controlling the axis drive motor, etc., setting the moving path of the imaging unit 120, adjusting the moving speed and exposure time, adjusting the light amount of the light source 126, capturing a two-dimensional image by the imaging unit 120, and pasting together a plurality of partial images. Various kinds of processing such as image stitch processing and analysis of the whole image obtained by imaging are executed.

Measurement performed using the image measuring machine 1 described above will be described below.

[Basic image measurement]
First, the operator operates the joystick 132 or the computer system 140 controls the stage 100 so that the work W is within the field of view. Then, the position of the imaging unit 120 in the Z-axis direction is adjusted so that the workpiece W is in focus. After focusing on the work W, the imaging means 124 captures an image for measurement. At this time, the coordinates of the stage 100 output by the X-axis encoder 112 and the Y-axis encoder 114 are taken into the computer system 140 and stored in the storage unit 41 together with the captured image. Specifically, a pulse serving as a trigger signal to the latch means 118 is output at the timing when the imaging means 124 completes imaging of one image. The latch means 118 latches and holds the position coordinates of each axis at the timing of the rise transition of the pulse (that is, substantially at the same time as the imaging of the image is completed). The computer system 140 fetches the image signal from the imaging means 124, fetches the position coordinates when the image is taken from the latch means 118, and stores them in association with each other.

Computer system 140 displays the resulting measurement image on display 144 along with a measurement tool interface for analyzing the image. FIG. 5 is a diagram showing an example of screen display. This screen display is displayed on the display 144 by a program (measurement application software) executed by the CPU 40 of the computer system 140 .

As shown in FIG. 5, the main window MW is displayed on the display 144 by executing the program. A plurality of windows (first window W1 to eighth window W8) are displayed in the main window MW. Menus and icons for various operations and settings are also displayed on the upper side of the main window MW. In this embodiment, eight windows are displayed as an example, but windows other than eight may be displayed as necessary, and windows may be divided, integrated, or omitted depending on the application. good too. Also, the layout of each window can be freely changed by the operator's operation.

An image WG of the workpiece W captured by the imaging unit 120 is displayed in the first window W1. The operator can adjust the position of the image WG of the work W displayed in the first window W1 by operating the mouse 143 or the joystick 132 of the remote box 130, for example. The operator can also enlarge/reduce the image WG of the work W by selecting an icon with the mouse 143, for example.

The second window W2 displays icons of measurement tools that can be selected by the operator. The icon of the measurement tool is provided corresponding to the specification method for specifying the measurement point from the image WG of the workpiece W. Examples of measurement tools include straight edge detection tools, circular edge detection tools, and the like.

Icons of functions that can be selected by the operator are displayed in the third window W3. A function icon is provided for each measurement method. For example, how to measure the coordinates of a point, how to measure the length of a straight line, how to measure a circle, how to measure an ellipse, how to measure a square hole, how to measure an oblong hole, how to measure a pitch. method, method of measuring the tolerance of two lines, and so on. The computer system 140, according to the operator's selection, measures dimensions such as the length of straight lines, distances between straight lines, diameters of circles, and deviations from ideal geometric shapes such as straightness, roundness, parallelism, etc. Insanity) evaluation.

A fourth window W4 displays a guidance showing an operation procedure for measurement.

In the fifth window W5, various sliders for controlling the lighting applied from the imaging unit 120 to the workpiece W are displayed. The operator can apply desired illumination to the work W by operating this slider.

The XY coordinate values of the stage 100 are displayed in the sixth window W6. The XY coordinate values displayed in the sixth window W6 are the coordinates in the X-axis direction and the Y-axis direction of the stage 100 with respect to a predetermined origin.

The seventh window W7 displays the result of tolerance determination. That is, in the seventh window W7, the result is displayed when a measurement method capable of judging the tolerance is selected.

A measurement result is displayed in the eighth window W8. That is, in the eighth window W8, the measurement result is displayed when a measurement method that obtains the measurement result by a predetermined calculation is selected. Details of the display of the tolerance determination result in the seventh window W7 and the measurement result in the eighth window W8 are omitted from the illustration.

[Appearance inspection screen]
In the image measuring machine 1 of the present embodiment, the program (measurement application software) executed by the CPU 40 of the computer system 140 performs the basic image measurement as described above, and also measures fine chipping, deformation, and burrs of the work W. , and provides a function of performing a so-called appearance inspection focusing on defects such as stains.

FIG. 6 shows an example of a screen for visual inspection (hereinafter referred to as a visual inspection screen). The visual inspection screen consists of an image pane P1, a filmstrip pane P2, a measurement result display pane P3, and a control pane P4. The appearance inspection screen may be provided as different tabs for each algorithm of image processing and defect determination, which will be described later. It consists of a control pane P4.

The image pane P1 is an area for displaying an image to be inspected for appearance. Image processing and defect determination are performed on the image displayed in the image pane P1 under the conditions set in the control pane P4.

When a mouse is dragged on the image pane P1, a rectangular area whose diagonal line is a line segment connecting the drag start point and the drag end point can be specified as an area for performing defect determination on the displayed image. . When this drag operation is performed, an area tool showing the outline of the specified rectangular area is displayed in the image pane P1 superimposed on the image. The region tool can be selected by clicking and resized by dragging the handles that appear upon selection. The area tool can also be deleted by pressing the DEL key on the keyboard in the selected state.

The filmstrip pane P2 is an area for displaying images read for visual inspection in a thumbnail format. When one of the images displayed in the filmstrip pane P2 is double-clicked, the image is displayed in the image pane P1 for image processing and defect determination. In the initial state immediately after the images are read, a predetermined image (for example, the first image when sorted by name, date of saving, etc.) is selected in the filmstrip pane P2 and displayed in the image pane P1.

If the image processed and judged in the image pane P1 has a defect, the image in which the defect is found in the filmstrip pane P2 is hatched H so that it can be easily distinguished from the image in which no defect has been found. At this time, it is preferable to change the hatching mode according to the number of defects. For example, if the processed image has defects and the number of defects is greater than or equal to a predetermined number, the image in the filmstrip pane P2 is hatched in red, and the processed image has defects but the number of defects is less than a certain number. In this case, the image in question in the filmstrip pane P2 should be hatched in orange. Further, it is preferable that the predetermined number, which is the boundary of the hatching pattern, can be set in the control pane P4, which will be described later.

The measurement result display pane P3 is an area for displaying defect information when a defect is found in the image displayed in the image pane P1. If there are multiple defects, information on all defects is displayed in a list format. When the defect information displayed in the list format in the measurement result display pane P3 is clicked, the defect information is displayed in the image pane P1 in a form associated with the defect in the image. Specifically, a figure surrounding the defective area D (for example, a contour line of the defect, a rectangle containing the defect, etc.) is superimposed on the image displayed in the image pane P1, and a balloon is displayed for the figure. to display defect information.

The control pane P4 is an area where a user interface is displayed for setting conditions for image processing and defect determination to be performed on the image displayed in the image pane P1. In addition to the above-described user interface, the control pane P4 also displays a comprehensive judgment indicating the overall quality of the work W corresponding to the image displayed in the image pane P1.

Using the user interface provided in the control pane P4, image processing parameters (e.g., threshold value for binarization, presence or absence of luminance value inversion, etc.), edge detection parameters (e.g., luminance change amount regarded as an edge) threshold, etc.), parameters for defect determination (for example, the upper limit of the allowable defect size (width, height, area, etc.), and the upper limit of the number of defects allowed in one image, etc. can be set. be done. The user interface may be provided as a GUI (Graphical User Interface) control such as a slider bar, a switch, etc., as well as a format for directly inputting numerical values.

Comprehensive judgment indicates a comprehensive judgment result of the work W corresponding to the image displayed in the image pane P1. For example, when the judgment result is pass (OK), the judgment result display area R1 is displayed in yellow green and "OK" is displayed, and when the judgment result is fail (NG), the judgment result display area is changed to It is preferable to make the determination result visually recognizable by displaying "NG" together with the color red. Although the criteria for determination are arbitrary, for example, when a plurality of images are obtained for one workpiece W, even one of these plurality of images detects a defect of an unacceptable size exceeding the allowable number. If the work W is found to be defective, the inspection result of the work W itself should be rejected (NG). Furthermore, even if the number of unacceptable size defects detected in each image does not exceed the allowable number, the total number of defects detected in a plurality of images corresponding to one workpiece W exceeds the allowable number. If it exceeds, the inspection result of the workpiece W may be rejected (NG).

[Flow of appearance inspection]
Next, the procedure of appearance inspection will be described with reference to the flowchart shown in FIG. In the following description, if there is no particular mention of the subject of processing, it should be understood that the program executed by CPU 40 of computer system 140 is the subject.

Prior to starting visual inspection, the user selects a menu for visual inspection in the program (measurement application software). In response to this, the display 144 displays an appearance inspection screen. The program then prompts the user to designate one or more images for visual inspection. When the user designates an image file in response to this, the image file is read, all the read images are displayed in the filmstrip pane P2, and the first image is displayed in the image pane P1. Here, it is assumed that information for identifying the corresponding work W is recorded as additional information in all the read images. Alternatively, the information for identifying the work W may not be recorded as additional information, and a plurality of images designated to be read collectively may be treated as images of the same work W. FIG.

Although the image displayed in the image pane P1 is subject to visual inspection, if the user wishes to have an image other than the top image inspected, the user can double-click the desired image in the filmstrip pane P2. , the desired image can be displayed in the image pane P1. With the image to be inspected displayed in the image pane P1 in this way, the visual inspection is started.

When the appearance inspection is started, the program first accepts the setting of various conditions by the user (step S100). Specifically, selection of tabs corresponding to image processing and defect judgment algorithms, setting of image processing parameters (binarization threshold, etc.), defect judgment parameters (permissible defect width, height, area etc.) and allowable number of defects. Note that these settings can be changed at any timing.

Subsequently, a target area for defect detection is identified (step S110). Any method can be used to specify the target area. For example, the target region may be specified using a region tool by user operation, or a region partitioned by edges extracted by edge extraction may be automatically identified as the target region. Alternatively, a region of a specific shape or position contained within the image may be specified as the target region.

Subsequently, image processing is performed based on the image processing parameters set in step S100 (step S120). Image processing may be applied to the entire image or only to the region of interest. The image processing performed here includes, for example, inversion of light and dark, noise removal, binarization, etc. In order to make it easier to identify the shape and size of the defect, the image is finally binarized. and

Next, in the image processed and binarized in step S120, a black region (that is, the pixel value is 0) in the target region is detected as a defect (step S130). If there are multiple black areas, each is detected as a separate defect. At this time, the program performs screen display to facilitate the user's recognition of the detected defects. That is, a figure surrounding the defective area D (for example, a contour line of the defect, a rectangle containing the defect, etc.) is displayed superimposed on the image displayed in the image pane P1. Further, when the defect information displayed in the list form in the measurement result display pane P3 is clicked, the defect information is displayed by providing a balloon for the graphic.

Subsequently, for each of the defects detected in step S130, it is checked against the defect determination parameters set in step S100 to determine whether or not the size exceeds the allowable size, and the number of defects exceeding the allowable size is counted (step S140). Then, the number of defects exceeding the allowable size is compared with the allowable number of defects set in step S100, and pass/fail determination is made for the image according to the comparison result (step S150).

In addition, the measurement application software displays a screen to facilitate the user's recognition of the defect detection status according to the number of detected defects and the result of comparison with the allowable number of defects (step S160). That is, if the number of detected defects exceeds the allowable number of defects, the image in filmstrip pane P2 is hatched in a prominent color such as red. Also, if the number of detected defects is not zero but does not exceed the allowable number of defects, this image in the filmstrip pane P2 is hatched in a less conspicuous color (e.g., orange) than the hatching applied when the number of allowable defects is exceeded. attached. Also, if no defect is detected (0 detections), this image in the filmstrip pane P2 is not hatched. The hatching added in this way is maintained even after the image has been visually inspected. Therefore, the user can recognize at a glance the detection status of defects in a plurality of images that have undergone visual inspection.

Subsequently, in addition to the pass/fail determination for the image performed in step S150, comprehensive determination is performed for the work W corresponding to the image (step S160). This comprehensive judgment is performed by aggregating the number of defects detected in other images of the same workpiece W as the current image and the results of pass/fail judgments. For example, if even one image fails the pass/fail judgment, or if the number of unacceptable size defects detected in an individual image does not exceed the allowable number, the defect detected in each image If the total number exceeds the allowable number, the inspection result of the workpiece W corresponding to the image may be rejected.

The above completes the appearance inspection for one image. Appearance inspection is then performed on all the read images while appropriately selecting targets.

[Specific example of algorithm for image processing and defect determination]
Next, a specific example of an algorithm for image processing and defect determination performed on the image displayed in the image pane P1 will be described.

[Example 1: Binarization]
"Binarization" is an algorithm that detects defects contained within a given region in an image. As shown in FIG. 6 above, when the "binarization" algorithm tab is selected, the setting parameters are the threshold for binarization (Brightness Threshold), the allowable defect size (width, height size, area) and the number of allowable defects (Allowable defect number) can be set in the control pane P4. By appropriately setting the parameters and selecting an image to be inspected in the filmstrip pane P2, the inside of the chamfered rectangle in the center of the image is identified as the defect target area.

Then, the inside of the target area is binarized according to the threshold value, and the area that becomes black as a result of the binarization is detected as a defect. Concerning the detected defect, a contour line surrounding the defect area D and a rectangle containing the defect are displayed superimposed on the image displayed in the image pane P1. Also, the information of the detected defects is displayed in the measurement result display pane P3 in a list format. The number of detected defects that exceed the allowed size is counted. The corresponding image in the filmstrip pane P2 is hatched according to this count value. Also, if the count exceeds the allowable number of defects, the workpiece corresponding to the image is rejected.

[Example 2: Area]
"Area" is an algorithm that detects defects contained within regions bounded by edges extracted in the image. As shown in Fig. 8, when you select the "Area" algorithm tab, the setting parameters are the threshold for binarization (Brightness Threshold), the inversion of the brightness value (Defective in the dark), and the threshold for the amount of change that is regarded as an edge. (Amount of change considered as an edge), allowable defect size (width, height, area), and allowable defect number (Allowable defect number) can be set in the control pane P4. When the parameters are appropriately set and an image to be inspected is selected in the filmstrip pane P2, the selected image is displayed in the image pane P1.

A region for edge detection is set with the region tool for the image displayed in the image pane P1. Edges existing within the set area are then extracted. The image is partitioned into two regions by a straight line fitting this edge. The brighter area of these two areas is specified as the target area. If the dark side area is desired to be the target area, the brightness value is set to be inverted to reverse the brightness.

Then, the inside of the target area is binarized according to the binarization threshold, and the area that becomes black as a result of the binarization is detected as a defect. Concerning the detected defect, a contour line surrounding the defect area D and a rectangle containing the defect are displayed superimposed on the image displayed in the image pane P1. Also, the information of the detected defects is displayed in the measurement result display pane P3 in a list format. The number of detected defects that exceed the allowed size is counted. The corresponding image in the filmstrip pane P2 is hatched according to this count value. Also, if the count exceeds the allowable number of defects, the workpiece corresponding to the image is rejected.

In addition, even if the count value in the image does not exceed the allowable number of defects, if the total of the count value for the image showing another part of the same work exceeds the allowable number of defects for the entire work , it is better to make a general judgment that the work is rejected. For example, in the example of FIG. 8, the edge of a circular workpiece is captured by dividing it into a plurality of images, and these images are displayed in the filmstrip pane P2. The count values for these images are aggregated to obtain the count value for the number of defects for the entire work, and the total number of defects for the entire work is compared with the allowable number of defects for comprehensive judgment.

[Example 3: Comparison]
"Comparison" is an algorithm that extracts edges in an image and compares the extracted edge positions with the edge positions in the master image to detect defects. burrs, etc.). As shown in FIG. 9, when the "Comparison" algorithm tab is selected, as setting parameters, the threshold of the error from the edge of the master image as a whole detected edge (Master Threshold), the threshold of the amount of change to be regarded as an edge (Amount of change considered as an edge), Allowable defect size (length (number of consecutive pixels), depth (error from edge in master image), Allowable defect number) can be set in the control pane P4.When appropriate parameters are set and an image to be inspected is selected in the filmstrip pane P2, the selected image is displayed in the image pane P1.

A region for edge detection is set with the region tool for the image displayed in the image pane P1. In the "comparison" algorithm, the area set at this time is the target area. After setting the region of interest, the program extracts the edges that lie within the region of interest. Then, the extracted edge is compared with the edge extracted from the master image, and the misaligned portion is detected as a defect. Then, the information of the detected defects is displayed in the measurement result display pane P3 in a list format. The number of detected defects exceeding the allowable size (length and depth) is counted. The corresponding image in the filmstrip pane P2 is hatched according to this count value. Also, if the count exceeds the allowable number of defects, the workpiece corresponding to the image is rejected.

In addition, even if the count value in the image does not exceed the allowable number of defects, if the total of the count value for the image showing another part of the same work exceeds the allowable number of defects for the entire work , it is better to make a general judgment that the work is rejected. Further, if the edges detected in the image as a whole deviate from the edges detected in the master image by exceeding a threshold (Master Threshold), the overall judgment may be rejected.

With the "comparison" algorithm described above, defects can be detected using any of a large number of workpieces as a reference (master) without preparing an ideal reference such as design data in advance. The method of determining the master image is arbitrary, but, for example, among a plurality of images read to be subjected to visual inspection, the first image subjected to visual inspection may be set as the master image. In this way, defects in a large number of images can be detected and visually inspected against a common reference.

Alternatively, in a case where a plurality of images read to be subjected to visual inspection are sequentially subjected to visual inspection, the image that was visually inspected immediately before (or the image that was judged to be acceptable immediately before) may be used as the master image. In this case, multiple images would not be checked against a common reference. However, if it is assumed that the position, shape, etc. of the edge to be inspected will change gradually due to changes over time in the manufacturing and processing processes of the workpiece, the appearance inspection will be performed according to the order in which manufacturing and processing were performed. Under the rule, by using the previous image instead of the first image, it is possible to eliminate the effects of changes over time while appropriately detecting defects larger than expected due to changes over time and performing visual inspections. becomes possible.

[Example 4: Outlier]
"Outlier" is an algorithm that extracts edges in an image and compares the extracted edges with ideal edges such as straight lines to detect defects. etc.). As shown in FIG. 10, when the "abnormal value" algorithm tab is selected, as setting parameters, the threshold of the error from the reference ideal edge (Error), the threshold of the amount of change to be regarded as an edge (Amount of change Considered as an edge), Allowable defect size (length (number of continuous pixels), depth (error from smooth edge)), Allowable defect number are displayed in the control pane P4. can be set by When the parameters are appropriately set and an image to be inspected is selected in the filmstrip pane P2, the selected image is displayed in the image pane P1.

A region for edge detection is set with the region tool for the image displayed in the image pane P1. In the "abnormal value" algorithm, the area set at this time is the target area. After setting the region of interest, the program extracts the edges that lie within the region of interest. Then, an ideal edge such as a straight line approximated to the extracted edge is obtained. Then, a portion that deviates from the ideal edge obtained by fitting is detected as a defect. Then, the information of the detected defects is displayed in the measurement result display pane P3 in a list format. The number of detected defects exceeding the allowable size (length and depth) is counted. The corresponding image in the filmstrip pane P2 is hatched according to this count value. Also, if the count exceeds the allowable number of defects, the workpiece corresponding to the image is rejected.

In addition, even if the count value in the image does not exceed the allowable number of defects, if the total of the count value for the image showing another part of the same work exceeds the allowable number of defects for the entire work , it is better to make a general judgment that the work is rejected. In addition, if the edges detected in the image as a whole deviate from ideal edges by exceeding a threshold value (Error), the overall judgment may be rejected.

According to the image measuring machine 1 of the present embodiment, using the algorithm described above, it is possible to perform a so-called appearance inspection focusing on defects such as minute chipping, deformation, burrs, dirt, etc. of the workpiece.

[Modification of Embodiment]
The present invention is not limited to the above-described embodiment examples, and includes modifications, improvements, etc. within the scope of achieving the object of the present invention. Appropriate additions, deletions, and design changes of components to the above-described embodiments or specific examples thereof by those skilled in the art are also included in the scope of the present invention as long as they have the gist of the present invention. .

1 Image measuring machine
40 CPU 41 storage unit 42 work memory 43, 44 interface 45 display control unit 100 stage
110 position acquisition means 112 X-axis encoder 114 Y-axis encoder 116 Z-axis encoder 118 latch means 118a counter 118b buffer 120 imaging unit
122 optical system
124 Imaging Means 126, 126a, 126b Light Source 130 Remote Box 132 Joystick 134 Jog Shuttle 140 Computer System 141 Computer Main Body 142 Keyboard 143 Mouse 144 Display P1 Image Pane P2 Film Strip Pane P3 Measurement Result Display Pane P4 Control Pane W Work

Claims (6)

a target area identifying step of identifying a target area in an image to be inspected;
a defect detection step of detecting defects included in the identified target area;
a defect counting step of counting the number of detected defects;
a judgment step for judging whether the object appearing in the image to be inspected is acceptable based on the counted number of defects ;
When a plurality of the inspection target images are obtained for one of the objects, in the determination step, even if the number of defects detected in each inspection target image does not exceed the allowable number, A visual inspection method , wherein the object is rejected if the total number of defects detected in the plurality of images to be inspected for one object exceeds an allowable number . The defect detection step includes:
extracting edges included in the region of interest;
2. The visual inspection method according to claim 1, further comprising the step of comparing the extracted edges with edges in a master image to detect defects. 3. The method according to claim 2, wherein when a visual inspection method is applied to a plurality of images to be inspected, the master image is an image to which the visual inspection method is first applied among the plurality of images to be inspected. Appearance inspection method. 3. The method according to claim 2, wherein when a visual inspection method is applied to a plurality of images to be inspected, the master image is an image to which the visual inspection method has been applied immediately before among the plurality of images to be inspected. Appearance inspection method. characterized in that, when applying a visual inspection method to a plurality of images to be inspected, the master image is an image, among the plurality of images to be inspected, that has been determined to pass as a result of applying the visual inspection method immediately before. The appearance inspection method according to claim 2. A program for causing a computer to execute the visual inspection method according to any one of claims 1 to 5.

Images