JP2007101239A - Visual inspection device and visual inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a visual inspection device capable of efficiently detecting a surface flaw difficult to detect by a conventional visual inspection device to judge the kind of the detected flaw. <P>SOLUTION: An illumination/imaging system 2 is constituted so as to obliquely irradiate a copper clad laminated sheet being an inspection target with light to capture an image. An image processor 3 processes the image sent from the illumination/imaging system 2. A composite processor 4 is constituted so as to sense the surface flaw by applying composite processing to the image processed by the image processor 3 to judge the kind of the surface flaw. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an appearance inspection apparatus and an appearance inspection method for automatically detecting defects generated on the surface of a copper clad laminate, for example.

  With recent increases in the density, high precision, and multi-layered printed wiring boards, copper-clad laminates such as glass cloth epoxy copper-clad laminates (MCL) for multilayers are strongly demanded for products with a high appearance level. Became.

  Conventionally, the visual inspection of a copper clad laminate has been performed by a human visual method. However, in human visual inspection, there are individual differences, and even the same person cannot effectively inspect the entire surface to be inspected at the same level, resulting in uneven detection of defects. In addition, the defect may be overlooked.

Thus, as disclosed in Patent Document 1 and Patent Document 2 below, many methods using an appearance inspection apparatus have been proposed. For example, there is an appearance inspection method using an appearance inspection apparatus that detects a surface defect by detecting a change in luminance of light reflected from the surface of an inspection object using a monochrome line sensor camera and a single illumination.
JP-A-10-9838 JP-A-11-183395

  By the way, in the inspection method combining the monochrome line sensor camera and the single illumination, it cannot be detected unless the defect has a clear contrast, and it is difficult to detect a low-contrast defect such as discoloration depending on the imaging angle. In addition, it is impossible to discriminate the types of detected defects because the contrast difference between the defects is small.

  Further, in the conventional appearance inspection apparatus, since the defect is determined based on whether or not the luminance change exceeds a predetermined level (threshold), if the luminance change occurs even if it is non-defective, the defect cannot be distinguished from the non-defect (light spot). There is a problem. The light spot here is a specular point generated during the production of a copper clad laminate.

  In addition, there are various surface defects of the object to be inspected, but in the conventional appearance inspection apparatus, various different defects are often captured as the same change, and it is difficult to accurately determine the type of the defect.

  The present invention has been made in view of the above circumstances, and an appearance inspection apparatus and an appearance that can efficiently detect surface defects that have been difficult with a conventional appearance inspection apparatus and determine the type of detection defect. The purpose is to provide inspection methods.

  In order to solve the above problems, an appearance inspection apparatus according to the present invention is an appearance inspection apparatus that automatically inspects defects generated on the surface of an object to be inspected. ) Two illumination sources that irradiate the surface of the object to be inspected with parallel rays from the direction for specularly reflected light and irregularly reflected light, respectively, and the surface of the object to be inspected from the two illumination sources. In addition, two light receptions having a light receiving operation direction in a direction orthogonal to the moving direction (or the opposite direction of the moving direction) of the inspection object for detecting the regular reflection light and the irregular reflection light from the surface of the inspection object of each light. Means, image processing means for subjecting the detection outputs detected by the two light receiving means to image processing to generate binarized data, respectively, and composite processing of each binarized data result generated by the image processing means And And a composite processing means for detecting defects on the surface of the inspection object by the difference in Shako.

  Further, in this appearance inspection apparatus, the composite processing means distinguishes the defect and the light spot from the barycentric coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the binary value. Centroid coordinates (X1, Y1) of the bright detection result from the normalized specular image are detected, and centroid coordinate differences | X1-X2 |, | Y1-Y2 | between the centroid coordinates are calculated. X1-X2 | <= arbitrary, | Y1-Y2 | <= arbitrary are determined, and if the conditions match, it is determined as a light spot.

  Further, the composite processing means distinguishes the concave defect from the scratches and the barycentric coordinates (X1, Y1) of the dark detection result from the binarized diffuse reflection image and the barycenter of the bright detection result from the same diffuse reflection image. Whether coordinates (X2, Y2) are detected and the center of gravity coordinates of the dark detection result is at an arbitrary distance of the center of gravity coordinates of the light detection result, or is the center of gravity coordinate of the light detection result at an arbitrary distance of the center of gravity coordinates of the dark detection result If there is a barycentric coordinate of the dark (bright) detection result at an arbitrary distance of the barycentric coordinate of the bright (dark) detection result, it is determined as a concave defect, and if not, it is determined as a scratch flaw defect.

  Further, the composite processing means detects the barycentric coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the dark detection result from the binarized regular reflection image in order to detect the discoloration defect. Centroid coordinates (X1, Y1) are detected, and it is determined whether the centroid coordinates of the bright detection result are at an arbitrary distance of the centroid coordinates of the dark detection result of the diffuse reflection image, and then dark detection from the diffuse reflection image is performed. The barycentric coordinate difference | X1-X2 |, | Y1-Y2 | between the barycentric coordinates (X2, Y2) of the result and the barycentric coordinates (X1, Y1) of the dark detection result from the specular reflection image is calculated, and the barycentric coordinates The difference | X1-X2 | <= arbitrary, | Y1-Y2 | <= arbitrary is determined, and if the condition matches, it is determined that the color change defect.

  The two illumination sources are two line illumination sources that can reveal all defects of the inspection object, and perform oblique illumination on the inspection object.

  The two light receiving means use two line sensor type cameras as an image capturing camera using specularly reflected light and an image capturing camera using irregularly reflected light.

  In order to solve the above problems, an appearance inspection method according to the present invention is an appearance inspection method for automatically inspecting a defect generated on a surface of an inspection object. Or an illumination step of irradiating light from the two illumination sources for specular reflection light and irregular reflection light with parallel rays from the direction of movement), and from the two illumination sources by the illumination step. Receiving operation of regular reflection light and irregular reflection light from the surface of the inspection object of each light irradiated on the surface of the inspection object in a direction orthogonal to the movement direction (or the opposite direction of the movement direction) of the inspection object A light receiving step for receiving light by two light receiving means having a direction, and image processing for generating binarized data by performing image processing on the detection outputs detected by the two light receiving means in the light receiving step And step, combined processes each binary data results produced by said image processing step, and a composite processing step of detecting a defect of a surface of an object to be inspected by the difference of the reflected light.

  Further, in the composite processing step, in order to distinguish the defect and the light spot, the barycentric coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the binarized regular reflection image The barycentric coordinates (X1, Y1) of the bright detection result are detected, barycentric coordinate differences | X1-X2 |, | Y1-Y2 | between the barycentric coordinates are calculated, and this barycentric coordinate difference | X1-X2 | <= arbitrary, | Y1-Y2 | <= Arbitrary is determined, and if the condition matches, it is determined as a light spot.

  Further, in the composite processing step, the barycentric coordinates (X1, Y1) of the dark detection result from the binarized irregular reflection image and the bright detection result from the same irregular reflection image are used to distinguish the concave defect and the scratch. The barycentric coordinates (X2, Y2) are detected, and the barycentric coordinates of the dark detection result are at an arbitrary distance of the barycentric coordinates of the bright detection result, or the barycentric coordinates of the bright detection result are at an arbitrary distance of the barycentric coordinates of the dark detection result If there is a barycentric coordinate of the dark (bright) detection result at an arbitrary distance of the barycentric coordinate of the bright (dark) detection result, it is determined as a concave defect, and if not, it is determined as a scratch flaw defect.

  Further, in the composite processing step, the barycentric coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the dark detection result from the binarized regular reflection image are used to detect the discoloration defect. Centroid coordinates (X1, Y1) are detected, and it is further determined whether or not the centroid coordinates of the bright detection result are at an arbitrary distance of the centroid coordinates of the dark detection result of the diffuse reflection image, and then dark detection from the diffuse reflection image is performed. The barycentric coordinate difference | X1-X2 |, | Y1-Y2 | between the barycentric coordinates (X2, Y2) of the result and the barycentric coordinates (X1, Y1) of the dark detection result from the specular reflection image is calculated, and the barycentric coordinates The difference | X1-X2 | <= arbitrary, | Y1-Y2 | <= arbitrary is determined, and if the condition matches, it is determined that the color change defect.

  As described above, the appearance inspection apparatus and method according to the present invention irradiate specular reflection and irregular reflection with parallel rays from the opposite (or movement) direction of the inspection object, for example, the copper clad laminate, for example. Then, the line sensor having the light receiving operation direction in the direction orthogonal to the direction of movement of the copper clad laminate (or opposite to the moving direction) detects and scans the inspection surface reflected light of the copper clad laminate, and according to the amount of light received Each detection output from the line sensor is binarized, the result is combined, and a defect on the surface of the inspection object is detected by the difference in reflected light.

  According to the appearance inspection apparatus and the appearance inspection method according to the present invention, it is possible to efficiently detect and determine the type of surface defects that have been difficult with the conventional appearance inspection apparatus. In addition, it is possible to provide countermeasures against defects in the previous step by quality assurance and type determination of the copper clad laminate.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

  FIG. 1 is an overall configuration diagram of an appearance inspection apparatus 1 according to an embodiment. This appearance inspection apparatus 1 is an apparatus for inspecting the appearance of a copper clad laminate such as a multilayer glass cloth epoxy copper clad laminate (MCL) as an object to be inspected, and executes the appearance inspection method of the present invention.

  The types of defects in the copper clad laminate are roughly classified into a concave defect, a scratch defect, and a smooth defect. These defects may become fatal defects that cause disconnection or the like, causing trouble in product functions.

  Concave defects include dents, copper wrinkles, mold flaws, cracks, and wrinkles. The dent is a dent of φ0.5 mm or more, and is generated by a foreign matter between the copper foil and the end plate. A copper wrinkle is a wrinkle of a copper foil, and is caused by a thin copper foil. A mold flaw is a strip-shaped depression having a depth of 0.7 μm or more, and is caused by a flaw in a copper plate. The crack is a crack in the copper foil at the resin protrusion (tread) portion of the prepreg, and is caused by the heavy deviation of the copper foil due to the crack. Also, tame wrinkles are copper wrinkles in the thigh part, and are caused by a shift in the longitude of the copper foil by the tuna.

  A flaw defect is a flaw having a length of 30 mm or more or a depth of 7 μm or more, and is caused by a machine trouble. In addition, smooth defects include discoloration and deposit discoloration. Discoloration is rust and dirt on the copper clad laminate copper foil, and is caused by water residue on the end plate. Adherent discoloration occurs when foreign matter adheres to the surface of the copper-clad laminate. This is caused by adhesion of machine oil or the like.

  Note that the copper clad laminate has a light spot generated during production. The light spot is a flat specular spot generated on the surface of the copper clad laminate. The light spot is not a fatal defect that causes disconnection or the like, and does not hinder the product function. This light spot causes a false alarm. Therefore, it is necessary to distinguish from the above defects.

  The appearance inspection apparatus 1 irradiates light on a copper-clad laminate, which is an object to be inspected, based on an irradiation method to be described later, and illuminates and images the imaging system 2 and images sent from the illumination and imaging system 2. An image processing device 3 for processing and binarizing, and a composite processing for detecting a surface defect by performing a composite processing on the binarized image processed by the image processing device 3 and further determining the type of the surface defect The apparatus 4 is provided.

  Here, as will be described later, the illumination / imaging system 2 is used for specularly reflected light and irregularly reflected light on the surface of the inspected object with parallel rays from the opposite (or moving) direction of the inspected object. Two illumination sources 22 and 23 that respectively irradiate light, and the specularly reflected light from the surface of the object to be inspected and the respective lights irradiated on the surface of the object 21 from the two illumination sources 22 and 23 and It comprises two light receiving means (line sensor type image capturing cameras 24 and 25) having a light receiving operation direction in a direction orthogonal to the moving direction (or the opposite direction of the moving direction) of the inspection object for detecting diffusely reflected light.

  As shown in FIG. 1, the illumination / imaging system 2 removes all defects from the copper-clad laminate 21 placed on the table and conveyed (moved) along the arrow 20 direction. Oblique illumination is performed from two line illumination sources 22 and 23 that can be realized. The line illumination source 22 performs oblique illumination for regular reflection, and uses metal halide illumination with high illuminance output. The line illumination source 23 performs oblique illumination for irregular reflection, and similarly uses metal halide illumination with high illuminance output. These illumination lights are parallel light.

  Further, the illumination / imaging system 2 uses two line sensor type cameras as the image capturing light receiving unit (camera) 24 using specularly reflected light and the image capturing light receiving unit (camera) 25 using irregularly reflected light. The line sensor type camera has a feature that the number of pixels is 7500 in the horizontal direction and a high resolution image can be captured, and the scanning speed is twice that of the area sensor type camera compared with the selection. Therefore, in this appearance inspection apparatus 1, a monochrome line sensor camera is used because of the features of high resolution and high speed processing. The line sensor type image capturing cameras 24 and 25 have a light receiving operation direction in a direction orthogonal to the moving direction of the copper clad laminate (or the direction opposite to the moving direction).

  The two lines of oblique illumination for regular reflection and irregular reflection by the two line illumination sources 22 and 23 are irradiated on the surface of the copper-clad laminate 21, and the inspection surface reflected light of the copper-clad laminate 21 is reflected by the line sensor method. Light is received by the image capturing cameras 24 and 25.

  FIG. 2 is a diagram showing a difference in received light amount of each defect by the two line illumination sources 22 and 23. Concave defects (such as dents and wrinkles) are detected in the dark by the amount of light received by specular reflection, and are detected both brightly and darkly by irregular reflection. Indentations and wrinkles can be generally classified by the ratio of horizontal: vertical = X: Y. Further, the scratch is detected in the dark, and the bright is detected in the irregular reflection. In addition, the color change is darkly detected by the amount of light received by regular reflection and irregular reflection. The light spot is different from bright detection in regular reflection and dark reflection in irregular reflection. Here, out of 255 gradations for monochrome, bright detection was set to 131 or more, dark detection was set to 129 or less, and none was set to 130. From the results shown in FIG. 2, it can be seen that significant defect classification is possible due to the difference in the amount of light received by the defect detection unit due to regular reflection and irregular reflection. That is, it is possible to distinguish a concave defect, a smooth defect, and a light spot by the two-optical system combined processing of regular reflection and irregular reflection, and the defect can be detected under each determination condition.

  The amount of the reflected light received by the image capturing cameras 24 and 25 is supplied to the image processing device 3. The image processing apparatus 3 can determine a concave defect, a scratch defect, and a smooth defect on the surface of the copper-clad laminate 21 with respect to the detection output from the line sensor (image capturing cameras 24 and 25) according to the amount of received light. Then, each detection output is binarized.

  The binarization result by the image processing device 3 is passed to the composite processing device 4. The composite processing apparatus 4 performs composite processing on the binarization result to determine a defect type, and a determination unit 41 that determines a defect based on a parameter set for each defect type based on the composite processing result. And a result output unit 42 for outputting the determination result of the determination unit 41.

  FIG. 3 is a flowchart showing a processing procedure of an appearance inspection method executed by the appearance inspection apparatus 1. First, in step S1, the illumination / imaging system 2 images the copper-clad laminate 21, which is an inspection object. Specifically, the surface of the copper clad laminate 21 is irradiated with two systems of oblique illumination for regular reflection and irregular reflection by the two line illumination sources 22 and 23, and the inspection surface reflected light of the copper clad laminate 21 is irradiated. Is received by the line sensor type image capturing cameras 24 and 25. These image capturing cameras 24 and 25 convert each received light amount into an image signal and supply it to the image processing apparatus 3.

  Next, in step S <b> 2, the image processing device 3 performs shading correction or luminance correction on each of the imaging signals. The imaging signal shaded or brightness-corrected by shading correction is smoothed in step S3, and then converted into binary data in the image processing apparatus 3 in step S4, and supplied to the composite processing apparatus 4. .

  In the composite processing apparatus 4, in step S5, the composite processing unit 40 is used to calculate the difference in reflected light and determine the type of defect. In step S <b> 6, the determination unit 41 determines the quality of the defect based on the composite processing result, and outputs the determination result from the result output unit 42.

  In order to distinguish the defect and the light spot, the composite processing device 4 centroid coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the light from the binarized regular reflection image The barycentric coordinates (X1, Y1) of the detection result are detected, barycentric coordinate differences | X1-X2 |, | Y1-Y2 | between the barycentric coordinates are calculated, and this barycentric coordinate difference | X1-X2 | <= arbitrary, | Y1-Y2 | <= arbitrary is determined, and if the condition matches, it is determined as a light spot.

  Further, in order to distinguish the concave defect and the scratch flaw, the composite processing device 4 uses the barycentric coordinates (X1, Y1) of the dark detection result from the binarized irregular reflection image and the bright detection result from the same irregular reflection image. The barycentric coordinates (X2, Y2) are detected, and the barycentric coordinates of the dark detection result are at an arbitrary distance of the barycentric coordinates of the bright detection result, or the barycentric coordinates of the bright detection result are at an arbitrary distance of the barycentric coordinates of the dark detection result If there is a barycentric coordinate of the dark (bright) detection result at an arbitrary distance of the barycentric coordinate of the bright (dark) detection result, it is determined as a concave defect, and if not, it is determined as a scratch flaw defect.

  For this reason, the composite processing unit 40 and the determination unit 41 of the composite processing apparatus 4 classify defects according to each detection algorithm described below based on the difference in reflected light in FIG.

  FIG. 4 is a diagram for explaining the light spot detection algorithm. FIG. 5 is a flowchart of light spot detection. First, in step S11, the barycentric coordinates (X2, Y2) are detected from the dark detection result from the irregular reflection image. Further, the barycentric coordinates (X1, Y1) are detected from the bright detection result from the regular reflection image. In step S12, the barycentric coordinate difference between these barycentric coordinates is calculated as | X1-X2 |, | Y1-Y2 |. In step S13, this barycentric coordinate difference | X1-X2 | <= arbitrary, | Y1-Y2 | <= arbitrary is determined, and if the conditions match, it is determined as a light spot (step S14).

  FIG. 6 is a diagram for explaining a method of classifying a concave defect and a scratch. FIG. 7 is a flowchart illustrating the classification of the concave defect and the scratch. FIG. 8 is a diagram showing processing for determining the type of defect according to the flowchart from specific examples of a concave defect and a scratch.

  As shown in FIG. 6, the diffuse reflection image with respect to the concave defect and the rubbing scratch detects both bright and dark in the concave defect and bright detection in the rubbing scratch. This is as shown in FIG.

  In step S21 of FIG. 7, a binarization result is obtained from the irregular reflection image of the concave defect and the scratch. In step S22, the composite processing unit 40 calculates the barycentric coordinates of the dark detection result and the light detection result. In step S23, the composite processing unit 40 determines whether the barycentric coordinates (X2, Y2) of the light detection result are present at an arbitrary distance of the barycentric coordinates (X1, Y1) of the dark detection result. It may be determined whether or not the barycentric coordinates (X1, Y1) of the dark detection result are at an arbitrary distance of the barycentric coordinates (X2, Y2) of the bright detection result. For example, if the barycentric coordinates (X2, Y2) of the bright detection result are present at an arbitrary distance of the barycentric coordinates (X1, Y1) of the dark detection result (YES), the process proceeds to step S24 and the composite processing unit 40 is a concave defect. Judge that. If there is no center of gravity coordinate (X2, Y2) of the bright detection result at an arbitrary distance of the center of gravity coordinate (X1, Y1) of the dark detection result (NO), the process proceeds to step S25 and the composite processing unit 40 has a scratch flaw defect. Judge that. As shown in FIG. 8, in the case of a scratch, only the light detection gravity center coordinates (X2, Y2) can be calculated in the gravity center coordinate calculation. Therefore, in the gravity center position search, the light detection gravity center coordinates (X2,. This is because there is no other barycentric coordinate at an arbitrary distance from Y2). As described above, the concave defect and the scratch can be divided.

  Then, the determination unit 41 determines whether or not a defect is possible based on the determination condition for each type of defect based on the determination results in step S24 and step S25 by the composite processing unit 40. The determination condition by the determination unit 41 is mainly based on the density of 255 gradations and the defect area (number of pixels).

  FIG. 9 is a diagram for explaining a color change defect detection algorithm. FIG. 10 is a flowchart for detecting a discoloration defect. First, in step S31 and step S32, the barycentric coordinates (X2, Y2) are detected from the dark detection result from the irregular reflection image. Further, the barycentric coordinates (X1, Y1) are detected from the dark detection result from the regular reflection image.

  Next, in step S33, the composite processing unit 40 determines whether or not the barycentric coordinates of the bright detection result are present at an arbitrary distance of the barycentric coordinates (X2, Y2) of the dark detection result of the irregular reflection image. This determination confirms that the defect is not a concave defect. The processing in step S33 is the same as part of the algorithm (see FIGS. 7 and 8) that divides the concave defect and the scratch defect. In this step S33, the irregular reflection dark detection portion which is a concave defect is removed. The specular reflection dark detection part which is a concave defect is also automatically removed. If it is determined in step S33 that the center of gravity coordinates of the light detection unit do not exist at an arbitrary distance of the center of gravity coordinates of the dark detection unit (NO), it is confirmed that there is no concave defect, and the process proceeds to step S34.

  In step S34, the barycentric coordinate difference between the two barycentric coordinates of the dark detection part of the irregular reflection image and the dark detection part of the regular reflection image calculated in step S32 is expressed as | X1-X2 |, | Y1-Y2 |. To calculate. Further, this center-of-gravity coordinate difference | X1-X2 | <= arbitrary, | Y1-Y2 | <= arbitrary is determined, and if the condition is met (YES), in step S35, the composite processing unit 40 determines that the color change defect is present. judge. That is, the process in step S34 compares the gravity center positions of the irregular reflection and regular reflection dark detection units, as shown in FIG.

  As described above, in the flowchart shown in FIG. 10, as a processing order, it is detected whether or not it is a concave defect in step S33, and whether or not it is a discoloration defect is detected in step S34. That is, in the flow of the color change defect detection algorithm, it can be determined that the defect is a color change defect while confirming that it is not a concave defect, so that efficient processing can be performed.

  As described above, the appearance inspection apparatus 1 that executes the appearance inspection method of the present invention can efficiently detect surface defects that have been difficult with conventional appearance inspection apparatuses, and determine the type of defect. it can.

  Next, an inspection system 50 for the copper clad laminate 21 including the appearance inspection apparatus 1 will be described with reference to FIGS. 11 to 13. FIG. 11 is a side view (A) and a top view (B) of the inspection system 50. FIG. 12 is a perspective view of a main part of the appearance inspection apparatus 1. FIG. 13 is a top view of the appearance inspection apparatus 1.

  As shown in FIG. 11, the inspection system 50 includes an input device 51 that inputs a copper-clad laminate as an inspection object into the system, a C / V 52 that performs two-sheet detection, plate thickness detection, and copper foil thickness inspection, The inspection was completed by the clean roller 53, the alignment unit 54 for aligning the copper clad laminate before moving to the first inspection machine, the first inspection machine 1 as the appearance inspection apparatus 1, and the first inspection machine 1. A reversing machine 55 for reversing the copper-clad laminate and reversing from one surface to the other, for example, from the back surface to the front surface, and a marking alignment unit 56 for aligning the markings on the copper-clad laminate reversed by the reversing machine 55; Place the second inspection machine 58 and the copper-clad laminate, which has been inspected by the second inspection machine 58, on the basis of the inspection result, and classify them into non-defective products, inspected items that require re-inspection, and defective products. Part 59.

  The copper clad laminate that has been input to the inspection system 50 by the input device 51 is subjected to two-sheet detection, plate thickness detection, and copper foil thickness inspection by the C / V 52. The aligning unit 54 aligns the copper clad laminate 21 moved from the C / V 52 and puts it on the table 26 shown in FIGS. 12 and 13 of the appearance inspection apparatus 1 used as a back surface inspection machine, for example. The copper clad laminate 21 placed on the table 26 reciprocates between the fixed illumination sources 22 and 23 and the lower portions of the cameras 24 and 25 as the table 26 moves in the direction of the arrow, that is, reciprocates.

  By this reciprocation, the front surface (the back surface in this system) of the copper clad laminate 21 is inspected to detect the defect and determine the defect type. When the inspection by the first inspection machine is completed, the copper clad laminate is discharged.

  The reversing machine reverses the copper-clad laminate discharged at the end of the inspection by the first inspection machine 1 and reverses it from one surface to the other surface, for example, from the back surface to the front surface. The copper-clad laminate that has been stamped and rearranged by the stamp alignment unit 56 is put into the second inspection machine 58.

  For example, the copper clad laminate that has been surface-inspected by the second inspection machine 58 is placed by the placement unit 59 into a non-defective product, an inspected object that needs re-inspection, and a defective product based on both inspection results. Placed.

It is a block diagram of an external appearance inspection apparatus. It is a figure which shows the difference in the light reception amount of each defect by the two line illumination sources 23 and 22. FIG. It is a flowchart which shows the process sequence of the external appearance inspection method performed with an external appearance inspection apparatus. It is a figure for demonstrating a light spot detection algorithm. It is a flowchart of a light spot detection. It is a figure for demonstrating the classification method of a concave defect and a abrasion crack. It is a division | segmentation flowchart of a concave defect and a abrasion crack. It is a figure which shows the process which determines the classification of a defect with the flowchart of FIG. 7 from the specific example of a concave defect and a flaw. It is a figure for demonstrating a discoloration defect detection algorithm. It is a flowchart of a discoloration defect detection. It is an external appearance side view of the inspection system containing an external appearance inspection apparatus. It is a perspective view of the principal part of an appearance inspection apparatus. It is a top view of an appearance inspection apparatus.

Explanation of symbols

1 visual inspection device, 2 illumination / imaging system, 3 image processing device, 4 composite processing device, 21 copper-clad laminate, 22 line illumination source (for irregular reflection), 23 line illumination source (for regular reflection), 24 by irregular reflection light Image capturing camera, 25 Image capturing camera using specularly reflected light, 40 Composite processing unit, 41 Determination unit, 42 Result output unit

Claims (10)

In visual inspection equipment that automatically inspects defects that occur on the surface of the inspection object,
Two illumination sources that respectively irradiate the surface of the inspection object with parallel rays from the opposite (or movement) direction of the movement direction of the inspection object for regular reflection light and irregular reflection light;
The movement of the inspection object (or the opposite of the movement direction) for detecting regular reflection light and irregular reflection light from the surface of the inspection object of each light irradiated on the surface of the inspection object from the two illumination sources. ) Two light receiving means having a light receiving operation direction in a direction orthogonal to the direction;
Image processing means for generating binarized data by performing image processing on the detection outputs detected by the two light receiving means;
A visual inspection apparatus comprising: composite processing means for performing composite processing on each binarized data result generated by the image processing means and detecting defects on the surface of the object to be inspected based on a difference in reflected light.   In order to distinguish the defect from the light spot, the composite processing means is configured to display the barycentric coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the light from the binarized regular reflection image. The barycentric coordinates (X1, Y1) of the detection result are detected, barycentric coordinate differences | X1-X2 |, | Y1-Y2 | between the barycentric coordinates are calculated, and barycentric coordinate differences | X1-X2 | <= arbitrary, | Y1 The appearance inspection apparatus according to claim 1, wherein -Y2 | <= arbitrary is determined, and if the condition matches, the light spot is determined.   In order to distinguish the concave defect and the scratch flaw, the composite processing means uses the barycentric coordinates (X1, Y1) of the dark detection result from the binarized irregular reflection image and the barycentric coordinates of the bright detection result from the same irregular reflection image. (X2, Y2) is detected, and whether there is a barycentric coordinate of the dark detection result at an arbitrary distance of the barycentric coordinate of the bright detection result, or whether there is a barycentric coordinate of the bright detection result at an arbitrary distance of the barycentric coordinate of the dark detection result And determining a concave defect if the barycentric coordinate of the dark (bright) detection result is present at an arbitrary distance of the barycentric coordinate of the bright (dark) detection result, and determining that it is a scratch flaw defect if not. 1. An appearance inspection apparatus according to 1.   The composite processing means detects the barycentric coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the dark detection result from the binarized regular reflection image in order to detect a discoloration defect. The barycentric coordinates (X1, Y1) are detected, and it is determined whether or not the barycentric coordinates of the bright detection result are at an arbitrary distance of the barycentric coordinates of the dark detection result of the diffuse reflection image, and then the dark detection result from the diffuse reflection image Centroid coordinate difference | X1-X2 |, | Y1-Y2 | between the centroid coordinates (X2, Y2) of the image and the centroid coordinates (X1, Y1) of the dark detection result from the specular reflection image is calculated. 2. The appearance inspection apparatus according to claim 1, wherein | X1-X2 | <= arbitrary, | Y1-Y2 | <= arbitrary are determined, and if the condition is met, it is determined as a discoloration defect.   2. The visual inspection apparatus according to claim 1, wherein the two illumination sources are two line illumination sources capable of revealing all defects of the inspection object, and perform oblique illumination on the inspection object. .   2. The visual inspection apparatus according to claim 1, wherein the two light receiving units use two line sensor type cameras as an image capturing camera using specular reflection light and an image capturing camera using diffuse reflection light. In the appearance inspection method for automatically inspecting defects that occur on the surface of the inspection object,
Illumination step of irradiating light from two illumination sources respectively for specular reflection light and irregular reflection light on the surface of the inspection object with parallel rays from the opposite (or movement) direction of the inspection object movement direction;
The specularly reflected light and the irregularly reflected light from the surface of the inspection object of the respective lights irradiated from the two illumination sources by the illumination step are moved (or moved in the moving direction) of the inspection object. A light receiving step for receiving light by two light receiving means having a light receiving operation direction in a direction orthogonal to the opposite direction;
An image processing step of performing image processing on the detection outputs detected by the two light receiving means in the light receiving step and generating binarized data, respectively;
A visual inspection method comprising: a composite processing step of performing composite processing on each binarized data result generated by the image processing step and detecting a defect on the surface of the inspection object based on a difference in reflected light.   In the composite processing step, in order to distinguish the defect and the light spot, the barycentric coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the light from the binarized regular reflection image are obtained. The barycentric coordinates (X1, Y1) of the detection result are detected, barycentric coordinate differences | X1-X2 |, | Y1-Y2 | between the barycentric coordinates are calculated, and barycentric coordinate differences | X1-X2 | <= arbitrary, | Y1 The appearance inspection method according to claim 7, wherein -Y2 | <= arbitrary is determined, and if the condition matches, the light spot is determined.   In the composite processing step, the barycentric coordinates (X1, Y1) of the dark detection result from the binarized irregular reflection image and the barycentric coordinates of the bright detection result from the same irregular reflection image are used to distinguish the concave defect and the scratch flaw. (X2, Y2) is detected, and whether there is a barycentric coordinate of the dark detection result at an arbitrary distance of the barycentric coordinate of the bright detection result, or whether there is a barycentric coordinate of the bright detection result at an arbitrary distance of the barycentric coordinate of the dark detection result And determining a concave defect if the barycentric coordinate of the dark (bright) detection result is present at an arbitrary distance of the barycentric coordinate of the bright (dark) detection result, and determining that it is a scratch flaw defect if not. 7. An appearance inspection method according to 7. In the composite processing step, the barycentric coordinates (X2, Y2) of the dark detection result from the binarized irregular reflection image and the dark detection result from the binarized regular reflection image are detected in order to detect a discoloration defect. The barycentric coordinates (X1, Y1) are detected, and it is determined whether or not the barycentric coordinates of the bright detection result are at an arbitrary distance of the barycentric coordinates of the dark detection result of the irregular reflection image, and then the dark detection result from the irregular reflection image Centroid coordinate difference | X1-X2 |, | Y1-Y2 | between the centroid coordinates (X2, Y2) of the image and the centroid coordinates (X1, Y1) of the dark detection result from the specular reflection image is calculated. 8. The appearance inspection method according to claim 7, wherein | X1-X2 | <= arbitrary, | Y1-Y2 | <= arbitrary are determined, and if the condition is met, it is determined as a discoloration defect.