JP2019002888A - Electrode extraction device, method for extracting electrode, and electrode extraction program - Google Patents

Abstract

To provide an electrode extraction device for extracting a plurality of electrodes from an image.SOLUTION: An electrode extraction device includes: an image acquisition unit 64 for acquiring an image including a plurality of electrodes; a pixel extraction unit for individually extracting pixels in different lightness ranges from the image; and an electrode extraction unit for extracting the electrodes from the individual extracted pixels.SELECTED DRAWING: Figure 2

Description

  The present invention relates to an electrode extraction apparatus, an electrode extraction method, and an electrode extraction program for extracting an electrode image from an image including electrodes.

  Conventionally, a technique for determining an indentation based on the brightness of an image obtained by imaging an electrode is known. For example, Patent Document 1 discloses a configuration in which electrodes are detected by pattern matching. Japanese Patent Application Laid-Open No. H10-228561 discloses that a portion where the lead terminal is located is displayed brightly or darkly with respect to the surroundings by binarization processing.

JP 2008-76184 A Japanese Unexamined Patent Publication No. 2014-21087

  In the prior art, it has been difficult to extract images of electrodes having different brightness. That is, in an electronic component or the like, a plurality of types of electrodes (for example, electrodes having different materials) may be used. It is complicated to prepare a pattern for each electrode and perform a pattern matching process for each pattern on an image obtained by imaging such an electrode. In addition, when the shapes of a plurality of types of electrodes are similar, it is difficult to distinguish the two by pattern matching processing. Furthermore, when the binarization process is performed, only one of a plurality of types of electrodes is extracted or both are extracted, and it is difficult to distinguish the two.

  The present invention solves the above-described problems, and an object thereof is to extract a plurality of types of electrodes from an image.

  In order to achieve the above object, an electrode extraction device includes an image acquisition unit that acquires an image including an electrode, a pixel extraction unit that extracts pixels of different brightness ranges from the image, and an electrode from each of the extracted pixels. An electrode extraction unit for extraction. That is, since the electrode extraction device extracts electrodes from pixels in different brightness ranges, it can distinguish a plurality of types of electrodes included in an image as images of different brightness. Therefore, a plurality of types of electrodes can be extracted from the image.

  Here, the image acquisition unit only needs to be able to acquire an image including a plurality of types of electrodes, and examples thereof include a configuration for acquiring an image captured by a sensor including a plurality of types of electrodes in the field of view. The image may be, for example, a visible light image or an X-ray image. The image may be captured by various optical systems. For example, a differential interference microscope or the like may be used, or another optical system may be used.

  The plurality of types of electrodes only need to have at least different brightness, and other elements, such as shapes, may differ. Note that the cause of the difference in brightness may be various causes. For example, the brightness of each electrode type reflects that the material, the position on the substrate, and the like differ for each electrode type. And the like are different. Of course, the hue may be different depending on the type of electrode, and the electrode may be extracted based on the difference in hue as well as the brightness. There may be at least two kinds of electrodes, and there may be three or more kinds.

  The pixel extraction unit only needs to be able to extract pixels in different brightness ranges from the image. That is, the pixel extraction unit only needs to be able to set a plurality of brightness ranges and extract pixels included in each brightness range from the image. The brightness range is a range in which the brightness of each image of the plurality of electrodes can be included, and may be the same number as the number of types of electrodes. For example, when the image includes two types of electrodes, two brightness ranges are set. The brightness range is preferably a range that does not overlap.

  The electrode extraction part should just be able to extract an electrode from each extracted pixel. In other words, the electrode extraction unit only needs to be able to extract pixels belonging to each lightness range as electrodes, or to extract a portion occupied by an electrode image from the pixels and extract them as electrodes. Of course, in this case, various types of noise removal processing, electrode shape correction, and the like may be performed. Once the electrode is extracted, it can be used for various inspections based on the image of the electrode. For example, the present invention can be used for inspection based on indentations of a conductive material formed when an electrode formed on two substrates (such as a glass substrate and a flexible substrate) is bonded with an anisotropic conductive film.

  In the pixel extraction unit, the brightness range may be determined in advance, may be determined automatically, or may be determined by a user instruction, and various configurations may be employed. When the brightness range is automatically determined, a configuration determined by analysis of an image including electrodes can be employed. For example, the pixel extraction unit may be configured to smooth the brightness frequency distribution of the pixels included in the image and acquire the boundary of the brightness range based on the change in the smoothed frequency distribution.

  That is, if the frequency distribution of brightness is smoothed, the influence of noise, the shape of the local distribution, etc. on the analysis results can be suppressed, and it is easy to analyze the global features. If the influence of noise or the like is small, smoothing may be omitted and the boundary of the brightness range may be acquired based on the change in the frequency distribution.

  Furthermore, since images of the same type of electrodes have similar brightness, and images of different types of electrodes have different brightness, a group is formed for each type of electrode image when the brightness frequency distribution is acquired. (For example, a distribution in which peaks occur at different positions). Therefore, by analyzing the lightness frequency distribution, it is possible to specify a lightness range in which the lightness of each type of electrode can exist.

  Various methods can be assumed as a method for analyzing the frequency distribution of lightness. However, if attention is paid to the change in the frequency distribution, it is possible to estimate whether the frequency distribution is changed due to different electrodes. it can. That is, it is estimated that different electrodes have different brightness, and it is considered that different groups are formed for each electrode in the frequency distribution. Therefore, if the boundary of the brightness range is acquired based on the change in the frequency distribution so that different electrodes are divided into different groups based on changes in the peak of the frequency distribution, the width of the distribution, the shape, etc., it is possible to acquire the brightness range. is there. According to this configuration, it is possible to determine a brightness range in which a plurality of types of electrodes can be extracted from an image.

It is a block diagram of an electrode extraction device. 2A is a perspective view of the imaging unit, and FIG. 2B is a schematic cross-sectional view of the substrate. FIG. 3A is a captured image, and FIG. 3B is a first electrode image. FIG. 4A is a second electrode image, and FIG. 4B is a frequency distribution of brightness. It is a flowchart of a joining state inspection process. 6A is an excerpt of the bonding state inspection process, and FIG. 6B is a frequency distribution of brightness. FIG. 7A is an excerpt of the bonding state inspection process, and FIG. 7B is a frequency distribution of brightness. FIG. 8A is an excerpt of the bonding state inspection process, and FIG. 8B is an example of selecting the brightness range. 9A is a display example of the selected brightness range, and FIG. 9B is a selection example of the brightness range.

Here, embodiments of the present invention will be described in the following order.
(1) Configuration of the electrode extraction device:
(2) Bonding state inspection process:
(3) Other embodiments:

(1) Configuration of the electrode extraction device:
FIG. 1 is a block diagram illustrating a schematic configuration of a bonding state inspection apparatus 10 that functions as an electrode extraction apparatus according to the present embodiment. In FIG. 1, the bonding state inspection apparatus 10 includes a control unit 20, a recording medium 30, a display device 40, an input device 50, and an imaging unit 60.

  The imaging unit 60 includes an illumination 61, a differential interference microscope 62, an XY stage 63, and a CCD sensor 64. FIG. 2A is a perspective view of the imaging unit 60. The XY stage 63 is a stage on which the substrate B is placed, and moves the substrate B in a direction parallel to the surface of the substrate B under the control of the control unit 20. The X axis and the Y axis are axes in the direction in which the XY stage 63 moves, and are coordinate axes orthogonal to each other on a plane parallel to the substrate B.

  The board | substrate B is a board | substrate which joined the semiconductor component to the glass substrate of the liquid crystal display. The semiconductor component is, for example, a liquid crystal driver circuit that generates a drive signal for driving a liquid crystal element of a liquid crystal display.

  The illumination 61 is a ring light. The CCD sensor 64 is an image sensor that captures a differential interference image (in the visual field V) formed by the differential interference microscope 62. The differential interference microscope 62 has an optical system that causes the polarized lights reflected at positions in the vicinity of the visual field V to interfere with each other. The differential interference microscope 62 causes the polarized light reflected at nearby positions to interfere with each other, thereby forming a differential interference image, which is an image in which the brightness changes according to the phase difference of the polarized light, on the CCD sensor 64. Since the phase difference of the polarized light reflected at a position in the vicinity of the field of view V corresponds to the difference in the optical path length of the polarized light, the differential interference image becomes an image that clearly shows the height difference at the position where the polarized light is reflected. . That is, the differential interference image is an image that clearly shows the level difference and inclination of the reflector of the light existing on the substrate B.

  Further, the differential interference image not only changes in brightness according to the height difference at the position where the polarized light is reflected, but the differential interference image has a higher brightness as the reflectance at the position where the polarized light is reflected. In other words, the change in brightness according to the height difference in the part with high reflectivity is seen in the bright brightness range, and the change in brightness according to the height difference in the part with low reflectivity is seen in the dark brightness range. It will be. In the present embodiment, differential interference images are captured in a state where two types of electrodes are included in the same field of view, but these two types of electrodes have different reflectivities due to different materials. Accordingly, in the differential interference image, different types of electrodes are captured as images of different brightness.

  The control unit 20 shown in FIG. 1 is a computer composed of a CPU, ROM, and RAM (not shown), and executes processing necessary for inspection of the bonding state using programs and various information recorded on the recording medium 30. As shown in FIG. 1, the control unit 20 executes a bonding state inspection program 21.

  The recording medium 30 records the captured image 30a, the first pixel information 30b, the second pixel information 30c, the first electrode image 30d, and the second electrode image 30e. The captured image 30a is an image captured by the CCD sensor 64, and is a differential interference image including images of two types of electrodes. The first pixel information 30b is information indicating pixels in the first brightness range included in the captured image 30a. In the present embodiment, the first pixel information 30b is image information binarized with pixels in the first brightness range and other pixels. is there. Specifically, the first pixel information 30b is image information obtained by binarizing the pixels in the first brightness range included in the captured image 30a to 1 and the pixels outside the first brightness range to 0.

  The second pixel information 30c is information indicating pixels in the second brightness range included in the captured image 30a. In the present embodiment, the second pixel information 30c is image information binarized by pixels in the second brightness range and other pixels. is there. Specifically, the second pixel information 30c is image information obtained by binarizing the pixels in the second brightness range included in the captured image 30a to 1 and the pixels outside the second brightness range to 0. In the present specification, one of the two types of electrodes is referred to as a first electrode and the other is referred to as a second electrode. The first brightness range is a brightness range in which the brightness of the image of the first electrode can be included, and the second brightness range is a brightness range in which the brightness of the image of the second electrode can be included.

  The first electrode image 30d is information indicating the pixel of the first electrode included in the captured image 30a, and is image information binarized with the pixel of the first electrode and other pixels in the present embodiment. Specifically, the first electrode image 30d is generated based on the first pixel information 30b, and images (for example, wiring and noise) other than the first electrode are excluded from the first pixel information 30b. The pixel information is expressed as 1 and the pixels other than the first electrode are expressed as 0.

  The second electrode image 30e is information indicating the pixel of the second electrode included in the captured image 30a. In the present embodiment, the second electrode image 30e is image information binarized with the pixel of the second electrode and other pixels. Specifically, the second electrode image 30e is generated based on the second pixel information 30c, and images (for example, wiring and noise) other than the second electrode are excluded from the second pixel information 30c. The pixel information is binarized to 1 and the pixels other than the second electrode to 0.

  In the recording medium 30, various other information not shown, for example, control information when the control unit 20 controls the illumination 61, the differential interference microscope 62, the XY stage 63, the CCD sensor 64, etc. The design information of the substrate B to be recorded is recorded.

  FIG. 2B is a schematic diagram illustrating the configuration of the substrate B. 2B shows the structure at different positions in the Y-axis direction on the right side and the left side of the drawing (for example, the left side of FIG. 2B shows the structure at the position Y1 in FIG. 3A. On the right side, the structure at position Y2 in FIG. 3A is extracted and shown). The substrate B includes a glass substrate B1 of a liquid crystal display and an electronic component U (which may be a flexible substrate or the like). The glass substrate B1 includes a plurality of first electrodes P1 and second electrodes P2. Since the first electrode P1 and the second electrode P2 are made of different materials, the glass substrate B1 includes two types of electrodes. In the present embodiment, the reflectance of the first electrode P1 is greater than that of the second electrode P2. The reflectance varies depending on the material used for the electrode, for example, the type and concentration of Ta (tantalum), Mo (molybdenum), Al (aluminum), and alloys thereof. For example, the first electrode P1 is made of molybdenum, and the second electrode P2 is made of tantalum having a lower reflectance than molybdenum.

  The glass substrate B1 is positioned closer to the positive side in the Z-axis direction than the electronic component U, and the first electrode P1 and the second electrode P2 are formed on the electronic component U side on the glass substrate B1. The electronic component U is located closer to the negative side in the Z-axis direction than the glass substrate B1, and the electrodes I1 and I2 face the glass substrate B1 side on the electronic component U. In the Z-axis direction, an anisotropic conductive film F is sandwiched between the glass substrate B1 and the electronic component U. In the anisotropic conductive film F, the conductive particles E are dispersed almost uniformly in the thermosetting binder.

  When pressure in the Z-axis direction is applied between the glass substrate B1 and the electronic component U while heating, the first electrode P1, the second electrode P2, and the electrodes I1, I2 are electrically connected via the conductive particles E. In the state, the anisotropic conductive film F is cured. The conductive particles E are conductive particles whose surfaces are coated with an insulating film. When pressure is applied to the conductive particles E sandwiched between the glass substrate B1 and the electronic component U, the insulating film is broken, and the first electrode P1 and the electrode I1 are electrically connected via the conductive particles E, and the first The two electrodes P2 and the electrode I2 are electrically connected.

  FIG. 2B shows a state where the first electrode P1, the second electrode P2, and the electrodes I1 and I2 are pressure-bonded by the anisotropic conductive film F. The conductive particles E are harder than the first electrode P1 and the second electrode P2, and when the glass substrate B1 and the electronic component U are pressure-bonded, the indentation Q of the conductive particles E is formed on the first electrode P1 and the second electrode P2. The

  The display device 40 is a display that displays various images under the control of the control unit 20. The input device 50 is a device that receives user operations, and may be a mouse, a keyboard, a touch panel, or the like.

  Next, the software configuration of the bonding state inspection program 21 will be described. The bonding state inspection program 21 includes an image acquisition unit 21a, a pixel extraction unit 21b, an electrode extraction unit 21c, and an inspection unit 21d. The image acquisition unit 21a, the pixel extraction unit 21b, and the electrode extraction unit 21c constitute an electrode extraction program that extracts electrodes from an image.

  The image acquisition unit 21a is a program module that causes the control unit 20 to realize a function of acquiring an image including a plurality of types of electrodes. That is, the control unit 20 captures the substrate B from the glass substrate B1 side by the function of the image acquisition unit 21a, and acquires the captured image 30a in which the first electrode P1 and the second electrode P2 are captured in the field of view.

  FIG. 3A is a schematic diagram of the captured image 30a. In FIG. 3A, for the sake of explanation, the distribution of each region existing on the captured image 30a is schematically shown instead of the differential interference image itself. The first electrode P1 is shown in white, the second electrode P2 is shown in gray, and the region where the first electrode P1 and the second electrode P2 do not exist is shown in black. Note that the characters P1 and P2 that are superimposed on the electrodes in FIG. 3A are symbols and are not included in the actual captured image 30a.

  Since the glass substrate B1 is almost transparent, the illumination light reaches the first electrode P1 and the second electrode P2. In addition, the 1st electrode P1 and the 2nd electrode P2 become a reflector which reflects the light of the illumination 61. FIG. Since the first electrode P1 has a higher reflectance than the second electrode P2, the first electrode P1 is brighter than the second electrode P2 in the captured image 30a. In the region where the first electrode P1 and the second electrode P2 are not present, the illumination light is transmitted or absorbed by the electronic component U. In FIG. 3A, the indentation of the conductive particles E is omitted.

  The pixel extraction unit 21b is a program module that causes the control unit 20 to realize a function of extracting pixels in different brightness ranges from the captured image 30a. That is, the control unit 20 uses the function of the pixel extraction unit 21b to include the first brightness range in which the brightness of the pixels of the first electrode can be included based on the captured image 30a and the brightness of the pixels of the second electrode. Set the brightness range.

  In this embodiment, two lightness ranges are set, but the lightness range may be the same as the number of electrode types. If there are three electrode types, three lightness ranges are set. The brightness range may be set by various methods. In the present embodiment, the control unit 20 acquires the brightness frequency distribution of the captured image 30a, and acquires the local minimum value of the frequency distribution as the boundary of the brightness range. To set the first brightness range and the second brightness range. Details of setting the brightness range will be described later.

  When the first brightness range and the second brightness range are set, the control unit 20 specifies the position of the pixel in the first brightness range from the captured image 30a by the function of the pixel extraction unit 21b. Then, the control unit 20 generates the first pixel information 30b in which the value of the pixel is 1 and the values of the other pixels are 0, and records the first pixel information 30b on the recording medium 30. For example, in the example shown in FIG. 3A, the first electrode P1 shown in white, the wiring W1 having the same lightness as the first electrode P1, and the noise pixel having the same lightness as the first electrode P1 are 1 pixel. First pixel information 30b in which the other pixels are set to 0 is generated.

  Further, the control unit 20 specifies the position of the pixel in the second brightness range from the captured image 30a by the function of the pixel extraction unit 21b. Then, the control unit 20 generates the second pixel information 30c in which the value of the pixel is 1 and the values of the other pixels are 0, and records the second pixel information 30c on the recording medium 30. For example, in the example shown in FIG. 3A, the second electrode P2 shown in gray, the wiring W2 having the same brightness as the second electrode P2, and the noise pixel having the same brightness as the second electrode P2 are 1 pixel. Second pixel information 30c in which the other pixels are 0 is generated.

  The electrode extraction unit 21c is a program module that causes the control unit 20 to realize a function of extracting an electrode from each pixel extracted by the pixel extraction unit 21b. In the present embodiment, in view of the fact that the wiring made of the same material as the electrode is connected to the electrode, the control unit 20 uses the function of the electrode extraction unit 21c to control the first pixel information 30b and the second pixel information. The first electrode image 30d and the second electrode image 30e are generated by excluding portions other than these electrodes from 30c and leaving electrode portions.

  The control unit 20 realizes a process of leaving the electrode portion by performing image processing based on the first pixel information 30b and the second pixel information 30c, and details thereof will be described later. 3B and 4A show examples of the first electrode image 30d and the second electrode image 30e. That is, the first electrode P1 is extracted based on the captured image 30a shown in FIG. 3A, and the first electrode P1 is 1 and the other part is 0. For the first electrode image 30d, 1 pixel is white, 0 An example in which pixels are shown in black is FIG. 3B. Based on the captured image 30a shown in FIG. 3A, the second electrode P2 is extracted, the second electrode P2 is 1, and the other part is 0. For the second electrode image 30e, 1 pixel is white and 0 pixel is An example shown in black is FIG. 4A.

  The inspection unit 21d is a program module that causes the control unit 20 to realize a function of inspecting the bonding state of the first electrode P1 and the second electrode P2. In this embodiment, based on the indentations formed on the first electrode P1 and the second electrode P2, the quality of the bonded state between the first electrode P1 and the anisotropic conductive film F, the second electrode P2 and the anisotropy The quality of the joining state with the conductive film F is inspected. That is, the control unit 20 specifies the position of the first electrode P1 based on the first electrode image 30d, and extracts the image of the portion that is the position of the first electrode P1 from the captured image 30a. Then, an indentation image is extracted from the image.

  The indentation image can be extracted by analyzing, for example, a change in lightness or a lightness pattern. When the indentation image is detected, the control unit 20 inspects the bonding state of the first electrode based on the indentation image. For the inspection of the bonding state, for example, a configuration that determines that the bonding state of the first electrode P1 is normal when the number of indentations detected at the individual first electrode P1 is equal to or greater than a threshold value can be employed. .

  The same applies to the inspection of the second electrode P2, and the control unit 20 extracts an image of the indentation from the captured image 30a based on the second electrode image 30e, and extracts an image of the indentation. And the quality of the joining state of the 2nd electrode P2 is determined based on criteria, such as the number of indentations. However, since the brightness of the first electrode P1 and the brightness of the second electrode P2 are different from each other, the parameters for extracting the impression from each electrode (such as a parameter indicating the change in the brightness to be the impression) may be different. In the present embodiment, as described above, in order to extract electrodes from pixels having different brightness ranges, it is possible to distinguish the first electrode P1 and the second electrode P2 included in the image as images having different brightness. it can. Therefore, the analysis can be performed by applying different parameters to the first electrode P1 and the second electrode P2, and it is possible to accurately specify the quality of the bonding state of each electrode.

(2) Bonding state inspection process:
FIG. 5 is a flowchart of the bonding state inspection process. First, the control unit 20 moves the XY stage 63 to a position where the inspection target of the substrate B can be imaged by the function of the inspection unit 21d (step S100). That is, the control unit 20 specifies the positions of the first electrode P1 and the second electrode P2 on the substrate B based on the design information of the substrate B (not shown) and the like so that these electrodes fall within the field of view V. -The Y stage 63 is moved.

  Next, the control unit 20 acquires an inspection region by the function of the inspection unit 21d (Step S105). The inspection area is an area including a portion to be imaged as the captured image 30a, and may be acquired by various methods. For example, an area including the first electrode P1 and the second electrode P2 may be acquired as an inspection area based on the design information described above, or an image captured by the CCD sensor 64 is displayed on the display device 40. In this state, the user may designate an examination area from the image by the input device 50, and various configurations can be adopted.

  Next, the control unit 20 captures the captured image 30a of the inspection region with the differential interference microscope 62 and the CCD sensor 64 by the function of the image acquisition unit 21a (step S110). Next, by the function of the image acquisition unit 21a, the control unit 20 acquires a captured image 30a of the inspection area captured by the CCD sensor 64 (step S115). The captured image 30 a is recorded on the recording medium 30.

  Next, the control unit 20 acquires the frequency distribution of the inspection region by the function of the pixel extraction unit 21b (step S120). That is, the control unit 20 obtains a frequency distribution of brightness by converting the gradation value of each pixel of the captured image 30a into a brightness value and counting the number of pixels having the same brightness value. Next, the control unit 20 smoothes the frequency distribution until there are three distributions by the function of the pixel extraction unit 21b (step S125). That is, the control unit 20 performs smoothing with a predetermined intensity on the frequency distribution obtained in step S120, and determines whether or not there are three upwardly convex peaks. If there are more than three peaks that are convex upward, the process of performing smoothing with a predetermined intensity is repeated.

  FIG. 4B shows an example of the frequency distribution after the frequency distribution obtained in step S120 is repeatedly smoothed until there are three upwardly convex peaks. In the example shown in FIG. 4B, there are a total of three upwardly convex peaks Pk0 to Pk2, and there are no upward inflection points in the other parts. In the inspection area in the present embodiment, there are mainly the first electrode P1 and the second electrode P2, and the part other than each electrode is imaged as a dark background. That is, as shown in FIG. 3A, the captured image 30a includes a dark background, a bright first electrode P1, and a dark second electrode P2.

  Therefore, if the general characteristics of the frequency distribution are captured, it is estimated that the frequency distribution reflects the brightness of these three types of images. Therefore, in the present embodiment, it is considered that the three distributions in which the brightness of the background, the first electrode P1 and the second electrode P2 has a peak are formed in the frequency distribution, and three by smoothing. It is configured to capture peaks. Therefore, for example, in the case of an image in which whiteout pixels can exist to some extent, a configuration in which smoothing is performed until there are four upwardly protruding peaks may be employed.

  Next, the control unit 20 sets a brightness range with the minimum value of the frequency distribution as a boundary by the function of the pixel extraction unit 21b (step S130). That is, the control unit 20 specifies the minimum value of the frequency distribution acquired in step S120 and acquires it as the boundary of the brightness range. As described above, in the present embodiment, three peaks are formed due to the brightness of the three images in the captured image 30a, so that there are at least two minimum values between the peaks. . The minimum minimum value M1 is estimated as the boundary between the background distribution and the distribution of the second electrode P2, and the second maximum minimum value M2 is the boundary between the distribution of the second electrode P2 and the distribution of the first electrode P1. Presumed.

  Therefore, in the present embodiment, the local minimum value M3 is set as the boundary of the lightness range. When there is a local minimum value M3 larger than the second largest local minimum value, the local minimum value M3 is also the boundary of the lightness range. Specifically, the control unit 20 sets a range from the minimum minimum value M1 to the second largest minimum value M2 as the second brightness range R2. Further, the control unit 20 sets a range from the second largest minimum value M2 to the third largest minimum value M3 (the maximum value or default value of lightness when M3 does not exist) as the first lightness range R1.

  Next, the control part 20 acquires the 1st pixel information 30b by the function of the pixel extraction part 21b (step S135). That is, the control unit 20 compares the lightness value of each pixel of the captured image 30a with the first lightness range, determines 1 as a pixel having a lightness within the first lightness range, and a pixel having a lightness outside the first lightness range. First pixel information 30 b is generated as 0 and recorded on the recording medium 30.

  Next, the control unit 20 performs a contraction process on the first pixel information 30b by the function of the electrode extraction unit 21c (step S140). That is, there is a possibility that the first pixel information 30b includes not only the first electrode P1 but also a wiring image close to the brightness of the first electrode P1, noise, and the like as the value 1. Therefore, the control unit 20 scans the image for each pixel, and performs a contraction process for replacing the pixel of interest with 0 when at least one 0 is present around the pixel of interest, and performs noise and the first electrode P1. Remove the wiring connected to.

  Further, the control unit 20 performs an expansion process on the information after the contraction process in step S140 by the function of the electrode extraction unit 21c (step S145). That is, since the size of the image of the first electrode P1 has been reduced by the contraction process in step S140, the control unit 20 executes the expansion process to return to the original size. Specifically, the control unit 20 scans the image for each pixel, and performs an expansion process for replacing the pixel of interest by 1 when at least one 1 exists around the pixel of interest. When the expansion process ends, the control unit 20 acquires the image after the expansion process as the first electrode image 30d by the function of the electrode extraction unit 21c and records it on the recording medium 30 (step S150). As a result, for example, the first electrode image 30d as shown in FIG. 3B is recorded on the recording medium 30.

  Next, the control unit 20 acquires the second pixel information 30c by the function of the pixel extraction unit 21b (step S155). That is, the control unit 20 compares the lightness value of each pixel of the captured image 30a with the second lightness range, determines 1 as a pixel having a lightness within the second lightness range, and a pixel having a lightness outside the second lightness range. Second pixel information 30 c is generated as 0 and recorded on the recording medium 30.

  Next, the control unit 20 performs a contraction process on the second pixel information 30c by the function of the electrode extraction unit 21c (step S160). That is, there is a possibility that the second pixel information 30c includes, as the value 1, an image of wiring, noise, or the like close to the brightness of the second electrode P2, but not the second electrode P2. Therefore, the control unit 20 scans the image for each pixel, and executes a contraction process for replacing the pixel of interest with 0 when there is even one around the pixel of interest, and performs noise and the second electrode P2 Remove the wiring connected to.

  Further, the control unit 20 performs an expansion process on the information after the contraction process in step S160 by the function of the electrode extraction unit 21c (step S165). That is, since the size of the image of the second electrode P2 has been reduced by the contraction processing in step S160, the control unit 20 executes the expansion processing to return to the original size. Specifically, the control unit 20 scans the image for each pixel, and performs an expansion process for replacing the pixel of interest by 1 when at least one 1 exists around the pixel of interest. When the expansion process ends, the control unit 20 acquires the image after the expansion process as the second electrode image 30e by the function of the electrode extraction unit 21c, and records it on the recording medium 30 (step S170).

  As a result, for example, the second electrode image 30e as shown in FIG. 4A is recorded on the recording medium 30. In addition, although the wiring and noise removal can be performed only by the contraction process and the expansion process in the present embodiment, other methods, for example, a method of excluding a part that does not have a certain area or a certain width may be combined. .

  Next, the control unit 20 detects an indentation by the function of the inspection unit 21d (step S175). That is, the control unit 20 extracts the image of the first electrode P1 from the captured image 30a based on the first electrode image 30d, and detects the indentation based on the parameter for detecting the indentation of the first electrode P1. Moreover, the control part 20 extracts the image of the 2nd electrode P2 from the captured image 30a based on the 2nd electrode image 30e, and detects an indentation based on the parameter for detecting the indentation of the 2nd electrode P2.

  Next, the control unit 20 performs pass / fail determination based on the detection result of the indentation by the function of the inspection unit 21d (step S180). That is, the control unit 20 determines the quality of the bonding state of each first electrode P1 based on the indentation on the first electrode P1 detected in step S175. Moreover, the control part 20 determines the quality of the joining state of each 2nd electrode P2 based on the impression on the 2nd electrode P2 detected by step S175.

(3) Other embodiments:
The above-described embodiment is an example, and various other embodiments may be adopted. For example, the captured image 30a is desirably a differential interference image, but may be an image captured by a normal microscope or an image obtained by performing a differential operation on the image. The electrodes to be inspected may be electrodes having different brightness. The factor that the brightness of the electrode is different does not necessarily have to be the difference in the material of the electrode. For example, the brightness of the electrode may differ depending on the surface treatment of the electrode.

  In the above-described embodiment, the expansion process is performed after the contraction process to remove the wiring and noise images, but these processes are also called opening processes, such as removing images other than the image of interest. Used for purposes. On the other hand, a series of processes for performing a contraction process after an expansion process is called a closing process, and is used for the purpose of correcting a missing part when a part that should be an image is missing. Such a closing process may be performed.

  Furthermore, the method for setting the brightness range may be another method. For example, in the configuration similar to FIG. 1 described above, the control unit 20 may acquire a minimum value having a relatively large change rate of the frequency distribution as a boundary of the lightness range by the function of the pixel extraction unit 21b. . This configuration can be realized by, for example, a configuration in which steps S125 and S130 in the flowchart shown in FIG. 5 are replaced with steps S125-1 and S130-1 shown in FIG. 6A.

  In step S125-1, the control unit 20 smoothes the frequency distribution by the function of the pixel extraction unit 21b. At this time, it is sufficient that the smoothing is performed to such an extent that a sudden change in the frequency distribution due to noise or the like can be eliminated. In general, step S125-1 may have a lower level of smoothing than step S125. . FIG. 6B shows the frequency distribution after smoothing in step S125-1. In FIG. 6B and FIG. 4B, it is assumed that the frequency distribution of the brightness of the same captured image 30a is smoothed. However, the intensity of smoothing is smaller in FIG. 6B, and as a result, the number of peaks is larger. It has become.

  Next, in step S <b> 130-1, the control unit 20 sets a lightness range with the upper two minimum values having a large change as a boundary by the function of the pixel extraction unit 21 b. That is, the control unit 20 specifies the minimum value of the frequency distribution acquired in step S125-1, and sets the change rate of the minimum value, for example, a default value including the absolute value or the minimum value of the gradient of the change rate near the minimum value. Acquired based on the frequency range of change in the range.

  When a plurality of minimum values are obtained by the above processing, the control unit 20 acquires two minimum values in descending order of the rate of change. The minimum values M1 and M2 shown in FIG. 6B are the top two minimum values having a large change rate. Also in this example, since the captured image 30a is assumed to mainly include three types of images of the background, the first electrode P1 and the second electrode P2, the control unit 20 has two minimum values. A range from a minimum value M1 having a relatively small brightness value to a minimum value M2 having a relatively large brightness value is set as the second brightness range R2. Moreover, the control part 20 sets the range more than the relatively large minimum value M2 as 1st brightness range R1. In the example shown in FIG. 6B, since the maximum value of the brightness value is M3, the first brightness range may be a range from the minimum value M2 to the maximum value M3.

  When the brightness range is set by the above processing, images of the first electrode P1 and the second electrode P2 can be extracted by the same processing as in the above-described embodiment, and the bonding state of each electrode is also inspected. be able to. In addition, since the first electrode P1 and the second electrode P2 can be distinguished and extracted, analysis can be performed by applying different parameters to the first electrode P1 and the second electrode P2. It becomes possible to accurately specify the quality of the bonding state. In this example, a minimum value with a relatively large change rate was appropriate for the separation of the brightness range, so a minimum value with a relatively large change rate was used as the boundary of the brightness range, but the frequency distribution was relatively In the case where the images are neatly separated, a configuration in which a minimum value having a relatively small change rate is used as the boundary of the brightness range may be employed.

  Further, the pixel extraction unit analyzes the characteristics of the frequency distribution of the brightness, and specifies the brightness range in which the brightness characteristics of the first electrode P1 and the second electrode P2 appear, whereby the first brightness range and the second brightness range. You may get As such a configuration, for example, in the same configuration as in FIG. 1 described above, the control unit 20 uses a function of the pixel extraction unit 21b to change the brightness distribution of the brightness of the pixels included in the image to a plurality of separation degrees. The light intensity range may be separated. This configuration can be realized by, for example, a configuration in which steps S125 and S130 in the flowchart illustrated in FIG. 5 are replaced with steps S125-2 and S130-2 illustrated in FIG. 7A.

  In step S125-2, the control unit 20 receives an input of the background portion of the captured image 30a by the function of the pixel extraction unit 21b. That is, it is considered that the background other than the first electrode P1 and the second electrode P2 is included in the captured image 30a and may become noise in the analysis of the frequency distribution. Therefore, the background portion to be excluded in step S125-2 is determined. Accept information for identification.

  Various input methods can be employed as the background portion input method. For example, the input device 50 is configured to accept a certain upper ratio and a lower certain ratio of brightness, and the user can input using the input apparatus 50. The structure etc. which specify a background part by the ratio which performed are mentioned. FIG. 7B is a diagram showing an example of the frequency distribution of brightness. In this example, assuming that the upper 5% and the lower 10% of the brightness are input as the background, the boundary of the upper 5% is set as the threshold T1, The lower 10% boundary is shown as the threshold value T2.

  Of course, the background portion input method shown here is an example, and the parameter for specifying the background portion may be a predetermined value or may be determined automatically. Further, only the upper level or the lower level may be input as the background portion according to the brightness of the background in the image. Note that the brightness frequency distribution shown in FIG. 7B is not smoothed, but may be smoothed.

  Next, the control unit 20 excludes the brightness range of the background portion of the captured image 30a, and acquires the brightness range of the electrode image from the remaining range. Specifically, in step S130-2, the control unit 20 sets a lightness range with a threshold at which the degree of separation of the frequency distribution excluding the background portion is maximized by the function of the pixel extraction unit 21b. That is, when the captured image 30a includes the background, the first electrode P1, and the second electrode P2, it is considered that a distribution reflecting the brightness of the two types of electrodes remains in the image from which the background portion is excluded. Therefore, if the frequency distribution of the image excluding the background portion is separated into two distributions, one is considered to be the lightness frequency distribution of the first electrode P1, and the other is the lightness frequency distribution of the second electrode P2.

  The degree of separation is an index that increases as the value increases, and includes, for example, the degree of separation defined by discriminant analysis (binarization of Otsu). For this degree of separation, a temporary threshold value T3 is set for the frequency distribution from which the background portion is excluded, and the number of pixels, average, and variance are obtained for each of the lightness distribution below the threshold T3 and the lightness distribution higher than the threshold T3. The intra-class variance and inter-class variance are obtained from these values, and can be defined by calculating the inter-class variance / intra-class variance. Therefore, the brightness range can be defined so that the two distributions are most appropriately separated by assuming the threshold value T3 and acquiring the threshold value T3 that maximizes the degree of separation.

  When a threshold value that maximizes the degree of separation is acquired, the control unit 20 sets a brightness range with the threshold value as a boundary. If the threshold value T3 shown in FIG. 7B is a threshold value that maximizes the degree of separation, the control unit 20 sets the range R1 as the first lightness range and the range R2 as the second lightness range. When the brightness range is set by the above processing, images of the first electrode P1 and the second electrode P2 can be extracted by the same processing as in the above-described embodiment, and the bonding state of each electrode is inspected. be able to. In addition, since the first electrode P1 and the second electrode P2 can be distinguished and extracted, analysis can be performed by applying different parameters to the first electrode P1 and the second electrode P2. It becomes possible to accurately specify the quality of the bonding state. Of course, the exclusion of the background portion may be applied to any embodiment.

  Furthermore, the structure which the control part 20 acquires the brightness range based on a user's input information may be sufficient. As such a configuration, for example, in the same configuration as in FIG. 1 described above, the control unit 20 receives input information input by the user through the input device 50 by the function of the pixel extraction unit 21b, and based on the input information. Thus, a configuration for acquiring the brightness range is given.

  The input information may be information for specifying the brightness range, and various information can be adopted. For example, the control unit 20 may acquire the first lightness range and the second lightness range by inputting the image ranges of the first electrode P1 and the second electrode P2 based on the captured image 30a. This configuration can be realized by, for example, a configuration in which steps S125 and S130 in the flowchart illustrated in FIG. 5 are replaced with steps S125-3 and S130-3 illustrated in FIG. 8A.

  In step S125-3, the control unit 20 receives an input of the range of the first electrode P1 and the range of the second electrode P2 by the function of the pixel extraction unit 21b. That is, in this example, the control unit 20 displays the captured image 30a on the display device 40. The user can designate the range of the electrodes on the image by operating the input device 50.

  FIG. 8B is a diagram illustrating a state in which the user designates the range Z1 of the first electrode P1 and designates the range Z2 of the second electrode P2 on the captured image 30a using the input device 50 such as a mouse. In this example, the user can designate a rectangular range with the input device 50 such as a mouse, and designates a rectangular range included in each of the first electrode P1 and the second electrode P2. The control part 20 acquires the range in the image of the 1st electrode P1 and the 2nd electrode P2 by receiving the said input.

  Next, in step S130-3, the control unit 20 sets a lightness range from the lightness of the pixels belonging to the input range by the function of the pixel extraction unit 21b. That is, the control unit 20 acquires the brightness of the pixels in the range Z1 designated as the range of the first electrode P1, and acquires the minimum brightness value to the maximum brightness value as the first brightness range. Further, the control unit 20 acquires the brightness of the pixels in the range Z2 designated as the range of the second electrode P2, and acquires the minimum brightness value to the maximum brightness value as the second brightness range.

  When the brightness range is set by the above processing, images of the first electrode P1 and the second electrode P2 can be extracted by the same processing as in the above-described embodiment, and the bonding state of each electrode is also inspected. be able to. In addition, since the first electrode P1 and the second electrode P2 can be distinguished and extracted, analysis can be performed by applying different parameters to the first electrode P1 and the second electrode P2. It becomes possible to accurately specify the quality of the bonding state.

  Moreover, the structure which the user inputs the input information which shows the numerical value of the brightness range directly or indirectly may be sufficient. For example, the configuration may be such that the user can input the boundary of the lightness range while confirming the lightness frequency distribution and the captured image 30a. This configuration can be realized, for example, by omitting steps S125 and S130 in the flowchart shown in FIG. 5 and accepting the input of the brightness range by the user by the control unit 20 instead.

  9A and 9B are diagrams showing a user interface when accepting an input by a user. FIG. 9A shows a state in which pixels belonging to the brightness range designated by the user are emphasized and superimposed on the captured image 30a. The control unit 20 displays the captured image 30a on the display device 40. When the user designates a brightness range on the frequency distribution by operating the input device 50, the control unit 20 emphasizes pixels belonging to the brightness range. In FIG. 9A, the emphasized pixels are indicated by hatching.

  FIG. 9B is a graph showing the frequency distribution of brightness, and the control unit 20 displays the frequency distribution of the captured image 30 a as shown in FIG. 9B on the display device 40. The user can move the boundaries Bo1 and Bo2 of the brightness range on the frequency distribution by operating the input device 50. The control unit 20 acquires the boundaries Bo1 and Bo2 after movement as boundaries of the brightness range, and emphasizes pixels belonging to the brightness range as illustrated in FIG. 9A. According to this configuration, the user can input the brightness range while confirming on the captured image 30a whether or not the brightness range specified by the user on the frequency distribution is appropriate. If the above processing is executed for each of the first brightness range and the second brightness range, the control unit 20 can acquire the first brightness range and the second brightness range.

  When the brightness range is set by the above processing, images of the first electrode P1 and the second electrode P2 can be extracted by the same processing as in the above-described embodiment, and the bonding state of each electrode is also inspected. be able to. In addition, since the first electrode P1 and the second electrode P2 can be distinguished and extracted, analysis can be performed by applying different parameters to the first electrode P1 and the second electrode P2. It becomes possible to accurately specify the quality of the bonding state.

DESCRIPTION OF SYMBOLS 10 ... Joining state inspection apparatus, 20 ... Control part, 21 ... Joining state inspection program, 21a ... Image acquisition part, 21b ... Pixel extraction part, 21c ... Electrode extraction part, 21d ... Inspection part, 30 ... Recording medium, 30a ... Imaging Image, 30b ... 1st pixel information, 30c ... 2nd pixel information, 30d ... 1st electrode image, 30e ... 2nd electrode image, 40 ... Display apparatus, 50 ... Input device, 60 ... Imaging part, 61 ... Illumination, 62 ... Differential interference microscope, 63 ... XY stage, 64 ... CCD sensor, B ... Substrate, B1 ... Glass substrate, E ... Conductive particles, F ... Anisotropic conductive film

Claims (10)

An image acquisition unit for acquiring an image including a plurality of types of electrodes;
A pixel extraction unit for extracting each pixel of a different brightness range from the image;
An electrode extraction unit for extracting the electrode from each of the extracted pixels;
An electrode extraction apparatus comprising: The pixel extraction unit
Setting the same number of brightness ranges as the number of types of electrodes;
The electrode extraction apparatus according to claim 1. The pixel extraction unit
Smoothing the frequency distribution of the brightness of the pixels included in the image, and obtaining a boundary of the brightness range based on a change in the frequency distribution after smoothing;
The electrode extraction device according to claim 1. The pixel extraction unit
Obtaining a minimum value of the frequency distribution as a boundary of the brightness range;
The electrode extraction apparatus according to claim 3. The pixel extraction unit
Obtaining the minimum value having a relatively large change rate of the frequency distribution as a boundary of the brightness range;
The electrode extraction apparatus according to claim 4. The pixel extraction unit
Separating the brightness frequency distribution of the pixels included in the image into a plurality of brightness ranges in which the degree of separation is maximized;
The electrode extraction apparatus in any one of Claims 1-5. The pixel extraction unit
Excluding the brightness range of the background portion of the image based on the brightness distribution of the pixels included in the image, and obtaining the brightness range from the remaining range;
The electrode extraction apparatus in any one of Claims 1-6. The pixel extraction unit
Obtaining the brightness range based on user input information;
The electrode extraction apparatus in any one of Claims 1-7. An image acquisition step of acquiring an image including a plurality of types of electrodes;
A pixel extraction step of extracting pixels of different brightness ranges from the image,
An electrode extraction step of extracting the electrode from each of the extracted pixels;
An electrode extraction method comprising: An image acquisition function for acquiring an image including a plurality of types of electrodes;
A pixel extraction function for respectively extracting pixels of different brightness ranges from the image;
An electrode extraction function for extracting the electrode from each of the extracted pixels;
Electrode extraction program that makes a computer realize.