JP4354174B2 - Manufacturing method for electronic circuit components - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present invention is an electronic circuit component.Made ofFor electronic manufacturing, more specifically, electronic circuit components suitable for high precision visual inspectionOutsideThe present invention relates to a method of manufacturing an electronic circuit component having a process in which an appearance inspection is performed by an inspection inspection method.
[0002]
[Prior art]
In recent years, not only dramatic progress in computers such as personal computers (PCs) and workstations (WS), but also in image input / output devices such as cameras and scanners, and image recording devices such as CDs and MOs, etc. Accordingly, processing techniques such as processing speed and processing accuracy in image processing have been remarkably developed.
Under such circumstances, the technical inspection of the appearance inspection of electronic circuit components such as package substrates and semiconductor components, which has conventionally relied on visual inspection, is going to be converted to automation using image processing. The electronic circuit component in this specification is based on a known package substrate such as a ceramic package substrate or a plastic package substrate, and a known electronic component including a semiconductor component such as an LSI or an IC chip, a chip capacitor, an antenna switch module, or the like. contains.
[0003]
As an image processing method used for the appearance inspection, various methods such as a black and white binarization processing method using a monochrome processing and a color image processing method have been studied.
Further, in the appearance inspection using color image processing, it is possible to enhance the ability to discriminate subtle color differences in appearance compared with the case of using black and white binarization processing. For this reason, it has been recognized that the color image processing method is a particularly useful method in the appearance inspection of electronic circuit components that are likely to generate subtle color differences due to their constituent materials.
[0004]
[Patent Document 1]
JP 2000-046651 A
[Patent Document 2]
Japanese Patent Laid-Open No. 10-311713
[0005]
[Problems to be solved by the invention]
Therefore, when the appearance inspection of electronic circuit components is performed using color image processing, the following problems with regard to processing accuracy become particularly apparent as compared with the case where black and white binarization processing is used. A color image is represented by coordinate values of three independent coordinate components in a three-dimensional color space coordinate system. For this reason, the ability to discriminate subtle color differences in appearance is enhanced as compared with the shade of black and white expressed by black and white binarization. However, due to the improvement of the identification capability, the position shift area caused by the position accuracy when the inspection image and the reference image are aligned is a defective area when color image processing is performed by, for example, difference processing or pattern matching. May be overextracted. More specifically, referring to the schematic diagram of FIG. 5, first, an inspection image 50 corresponding to a predetermined inspection region of the electronic circuit component is acquired. This inspection image 50 represents one coordinate component of three independent coordinate components forming a three-dimensional color space coordinate system. Further, it is assumed that the inspection image 50 includes a wiring pattern 61 formed on a surface (hereinafter also simply referred to as a substrate surface) 60 on the substrate and a defective region 62 to be extracted. Then, after aligning the preset reference image 51 corresponding to the coordinate component representing the inspection image 50 and the inspection image 50, color image processing is performed using difference processing, pattern matching, or the like. At this time, the color difference of the shift area 63 caused by the alignment accuracy is clarified with respect to the substrate surface 60 and the defective area 62 by displaying the color image as compared with the black and white binarized display. For example, when the color difference between the deviation area 63 and the substrate surface 60 is close to the color difference between the defective area 62 and the substrate surface 60, a difference value representing the color difference between the substrate surface 60 and the deviation area 63 is obtained in the difference process. The difference is reduced with respect to the difference value representing the color difference between the substrate surface 60 and the defective area 62, and it is easily determined that the difference is outside the threshold value to be set. As a result, the deviation area is excessively extracted as the defective area. Will increase the probability. Of course, the processing accuracy for extracting the defective area itself is improved by performing color image processing. However, in order to further increase the processing accuracy, it is essential to suppress such excessive extraction.
[0006]
In addition, since the color difference of the above-described misalignment region 63 with respect to the substrate surface 60 is further clarified, the misalignment region 63 is within a range that has been missed in accuracy in the black and white binarization process in pattern matching or more. Therefore, it will be excessively extracted as a defective area. As described above, when the color misalignment area due to the alignment is displayed as a color image by a three-dimensional color space coordinate system as compared with the black and white binarized display, the identification ability is enhanced, so that it is excessively extracted as a defective area. The case where it is done is raised.
[0007]
As described above, the appearance inspection using color image processing has higher processing accuracy than the case of using black and white binarization processing, but corresponds to the recent increase in density and integration of electronic circuit components. In order to further improve the accuracy, it is important to reduce the excessive extraction of the unreasonable region due to the above-described deviation region. The present invention has been made in view of this problem. That is, the present invention makes it possible to improve processing accuracy when an appearance of an electronic circuit component is inspected using color image processing. It is an object of the present invention to provide an electronic circuit component appearance inspection method and an electronic circuit component manufacturing method including a step of performing an appearance inspection by the appearance inspection method.
[0008]
[Means for solving the problems and actions / effects]
  The method of manufacturing an electronic circuit component of the present invention for solving the above problems is as follows.
  Forming a wiring pattern on the substrate;
  The color light receiving unit receives the inspection light from the inspection surface on the substrate on which the wiring pattern is formed by the process,
  Based on the detection output of the color light receiving unit, generates inspection surface color information composed of three coordinate components of a three-dimensional color space coordinate system at each position in the inspection surface;
  The first inspection plane data representing the three coordinate components of the inspection plane color information of the generated three-dimensional color space coordinate system is transformed into a three-dimensional color space coordinate system different from the three-dimensional color space coordinate system. The coordinate component of the first image processing data that includes the first inspection surface data represented by the three coordinate components and at least one coordinate component among the six coordinate components based on the second inspection surface data while creating the second inspection surface data And data representing
  Of the preset good standard color information at each position in the inspection surface, the maximum value filtering process is performed on the good standard color information data corresponding to the data representing the coordinate component of the first image processing data. Data of the maximum value non-defective product reference data and the minimum value non-defective product reference data subjected to the minimum value filter processing,
  By subtracting each, the first defect candidate area is selected,
  Identifying the defective area on the inspection surface based on the first defective candidate area thus selected,
  A determination step of determining pass / fail of the wiring pattern based on the specified defective area;
  A method for manufacturing an electronic circuit component including:
  The subtraction process for selecting the first defect candidate area is:
  Using the maximum value non-defective product reference data and the minimum value non-defective product reference data of the same coordinate component as the coordinate component of the first image processing data,Coordinate component of the first image processing dataWhenCoordinate components processed using the maximum value non-defective product reference dataBetween,The coordinate component of the first image processing data andCoordinate components processed using the minimum value non-defective product reference dataPerformed between
Maximum value non-defective product reference data for one coordinate componentSubtraction processing using coordinate components processed usingOrFor one coordinate componentMinimum value good standard dataCoordinate components processed usingUsedSubtractionWhen the process is a unit of the number of processes, the number of processes is two.
[0009]
  Of the present inventionConfigure the main partThe appearance inspection method for an electronic circuit component is performed by performing image processing on the color image of the appearance of the inspection surface of the electronic circuit component, which is the appearance inspection target. First, a color image of the appearance of the inspection surface is obtained by receiving inspection light from the inspection surface by the color light receiving unit, and the color light receiving unit outputs a detection output for the input signal of the color image. Then, based on the detection output, inspection surface color information including three coordinate components of a three-dimensional color space coordinate system at each position in the inspection surface is generated. The inspection surface color information includes coordinates of three independent coordinate components forming a three-dimensional color space coordinate system at each position in the inspection surface, that is, at each pixel position where the color image of the appearance of the inspection surface is partitioned. It consists of values. As described above, the inspection surface color information includes the position information of each position in the inspection surface and the color image information at the position.
[0010]
After generating the inspection surface color information, the first inspection surface data representing the three coordinate components of the inspection surface color information and the three-dimensional color space coordinates different from the own three-dimensional color space coordinate system. First image processing data including at least one coordinate component in the second inspection plane data coordinate-converted by the system is prepared. The first inspection surface data is composed of the same three-dimensional color space coordinate system as the inspection surface color information, while the second inspection surface data is composed of a three-dimensional color space coordinate system different from the first inspection surface data. . When the appearance of the electronic circuit component is inspected, the coordinate components of the three-dimensional color space coordinate system that are useful depending on the inspection object change. Therefore, the first image processing data is composed of at least one coordinate component of the first inspection surface data and the second inspection surface data in a form corresponding to the useful coordinate components as appropriate. This first image processing data is image data that forms an inspection image when image processing is performed. Then, as data used as a reference image for the inspection image, data formed as follows is used. Maximum value non-defective product reference data subjected to maximum filter processing and minimum value non-defective product reference subjected to minimum value filter processing for data corresponding to each coordinate component of the first image processing data among preset good product standard color information At least one of the data is used as data serving as a reference image. First, the non-defective product reference color information includes non-defective product reference values set in advance for each coordinate component of the color image information at each position in the inspection surface. That is, a predetermined reference value within a non-defective acceptable range is set in advance as a coordinate value for each coordinate component of a three-dimensional color space coordinate system that is appropriately selected when image processing is performed as an appearance inspection.
[0011]
Then, for the data corresponding to each coordinate component of the first image processing data in the non-defective color reference information, the maximum non-defective product reference data subjected to the maximum value filter processing and the minimum non-defective product reference data subjected to the minimum value filter processing are obtained. prepare. This data may be from the viewpoint of processing time created in advance together with the good quality standard color information. Here, the maximum value filtering process and the minimum value filtering process will be specifically described with reference to the schematic diagram of FIG. By scanning each pixel 70 with respect to the reference image 51 representing a coordinate component corresponding to one coordinate component of the first image processing data in the good product reference color information, for example, 8 pixels which are the smallest unit surrounding itself The value of the pixel to be scanned (the pixel in the shaded area in FIG. 4) is rewritten to the maximum or minimum coordinate value value among the nine pixels including. At this time, the process of rewriting to the maximum coordinate value is the maximum value filter process, and the process of rewriting to the minimum coordinate value is the minimum value filter process. At this time, as shown in FIG. 4B, the area of the pixel for searching for the maximum value and the minimum value may be 5 pixels including 4 pixels closest to itself, or as shown in FIG. Alternatively, a region of 13 pixels including the 4 pixels closest to the next one shown in FIG. Thus, the area of the pixel for searching for the maximum value and the minimum value is not particularly limited, and when it is necessary to precisely filter a narrower area of the reference image 51, the maximum value and the minimum value are determined. This is selected as necessary, for example, by reducing the area of the pixel to be searched for. In addition, the pixel region for searching for the maximum value and the minimum value, which is set when performing these filter processes, serves as a reference at the time of scanning each pixel. For example, FIG. When 9 pixels including 8 pixels, which are the smallest unit that surrounds itself, are set as areas as shown in FIG. 4, in the pixel that forms a vertex on the reference image, the smallest unit area that surrounds itself is 4 pixels. It is something like that.
[0012]
As described above, the maximum value non-defective product reference data and the minimum value filter processing in which the maximum value filter processing is performed on the data forming the reference image corresponding to each coordinate component of the first image processing data in the non-defective product reference color information. The minimum value non-defective product reference data to which is applied is created. The filter reference images formed by the maximum value non-defective product reference data and the minimum value non-defective product reference data are, for example, as shown in the schematic diagram of FIG. Here, the chromaticity of the wiring pattern 61 is smaller than that of the substrate surface 60, that is, a schematic diagram when the wiring pattern 61 is smaller when comparing the average coordinate values in the respective image regions. By performing the maximum value filtering process on the reference image 51 in this way, the image area of the wiring pattern 61 is reduced, while by applying the minimum value filtering process, the image area of the wiring pattern 61 is increased. Of course, when the chromaticity of the wiring pattern is larger than that of the substrate surface, the image subjected to the minimum value filter processing is the same as the filter reference image subjected to the maximum value filter processing shown in FIG. The image subjected to the value filter process is the same as the filter reference image subjected to the minimum value filter process shown in FIG.
[0013]
Then, the data representing each coordinate component of the first image processing data is subtracted from the data corresponding to each coordinate component of at least one of the maximum value non-defective product reference data and the minimum value non-defective product reference data, respectively. To process. At this time, for example, when subtracting the maximum value non-defective product reference data or the minimum value non-defective product reference data from the first image data, the value obtained by subtracting the maximum value non-defective product reference data is positive, Extract the value that is negative when subtracting. Of course, when the first image processing data is subtracted from the maximum value non-defective product reference data or the minimum value non-defective product reference data, the value to be extracted may be reversed. A set of positions in the inspection surface corresponding to the data extracted in this way is set as a first defect candidate region on the inspection surface. Here, the effect of using the maximum value non-defective product reference data and the minimum value non-defective product reference data will be described with reference to the schematic diagram of FIG.
FIG. 7 shows the case where the substrate surface 60 is larger when the average coordinate values in the image areas of the substrate surface 60 and the wiring pattern 61 are compared. Here, the subtraction process considers only the case where the maximum value non-defective product reference data or the minimum value non-defective product reference data is subtracted from the first image data. FIG. 7 is an overlay of the inspection image (broken line in the figure) formed by one coordinate component in the first image data and the filter reference image (solid line in the figure) corresponding to the coordinate component of the inspection image. is there. 7 is used, the average coordinate value of the defective area 62 to be extracted is larger than that of the wiring pattern 61 and the substrate surface 60. Even if the defective area 62 and the substrate surface 60 are Even when the difference value of the average coordinate value of the first image processing data is close to the difference value of the average coordinate value between the wiring pattern 61 and the substrate surface 60, the maximum value non-defective product reference data Since the deviation region due to the alignment accuracy with the filter reference image formed by is at least zero or negative in the subtraction process, only the defective region 62 is extracted as a positive value in the subtraction process. Further, the average coordinate value of the defective area 62 to be extracted is smaller than that of the wiring pattern 61, and the difference value between the average coordinate values of the defective area 62 and the substrate surface 60 is the same as the wiring pattern 61 and the substrate surface 60. Even when the difference between the average coordinate values of the two is close, by using the minimum value non-defective product reference data in the lower diagram of FIG. 7, the deviation region due to the alignment accuracy between the inspection screen and the filter reference image is subtracted. Therefore, only the defective area 62 is extracted as a negative value in the subtraction process.
[0014]
Further, for example, when the average coordinate values in the substrate surface and the image area of the wiring pattern are compared, if the wiring pattern is larger, the maximum value non-defective product reference data in FIG. In addition, considering the case where the average coordinate value of the defective area 62 to be extracted is smaller than the wiring pattern 61 or the substrate surface 60, only the defective area 62 is extracted as a negative value in the subtraction process for the same reason. Will be. On the other hand, the minimum value non-defective product reference data in the lower diagram of FIG. 7 is set as the maximum value non-defective product reference data, and the average coordinate value of the defective area 62 to be extracted is smaller than the wiring pattern 61 and the substrate surface 60. For this reason, only the defective area 62 is extracted as a positive value in the subtraction process.
[0015]
As described above, by performing image processing by subtraction processing using the maximum value non-defective product reference data and the minimum value non-defective product reference data, wiring is caused by the alignment accuracy between the inspection image and the reference image, which has been a problem in the past. It is possible to effectively suppress the deviation area generated in the peripheral portion of the pattern from being excessively extracted as a defective area. Then, a set of positions in the inspection surface corresponding to the data extracted by the subtraction process is set as a first defect candidate region on the inspection surface, and on the inspection surface based on the selected first defect candidate region. The defective area is specified. As a result, the specified defective area is based on the first defective candidate area selected using the subtraction process, and thus the specific accuracy is effectively improved.
[0016]
As described above, according to the present invention, it is possible to effectively increase the processing accuracy of appearance inspection of electronic circuit components performed using color image processing. When color image processing is used, as a matter of course, it is better that a defective area to be extracted as an inspection target has a larger color difference on the image than other areas, and such coordinate components of the three-dimensional color space coordinate system are useful. The first image processing data of the present invention is configured from the coordinate components as appropriate. However, due to the excessive extraction of defective areas due to the alignment accuracy between the inspection image and the reference image, the actual problem is that the defective area to be extracted as the inspection object has a larger color difference on the image than the other areas. Although there are aspects that cannot be said to be good, it is possible to further increase the usefulness of a color image by using image processing by subtraction processing shown in the present invention. Further, as shown in FIG. 7, the effect of the subtraction process is enhanced as the overlapping area between the wiring pattern of the filter reference image used in the subtraction process and the wiring pattern of the inspection screen increases. It is desirable to appropriately adjust the pixel region for searching for the maximum value and the minimum value when performing the filtering process in consideration of the amount of deviation due to.
[0017]
  the aboveVisual inspection of electronic circuit componentsTo reviewIn the subtraction process, the coordinate component processed using the maximum value non-defective product reference data and the coordinate component processed using the minimum value non-defective product reference data are included in each coordinate component of the first image processing data.It is good.
[0018]
Defect areas that are subject to visual inspection of electronic circuit components include, for example, many inspection items such as adhesion of foreign matter, discoloration, cracks, and surface peeling on the substrate surface on which the wiring pattern is formed. Therefore, it is often difficult to unambiguously limit the chromaticity relationship and color difference between the defective area and the other non-defective areas. Of course, from various preliminary experiments and experimental data obtained therefrom, it is possible to limit to some extent the items that are likely to occur as defective areas and the chromaticity relationship between the non-defective areas and other non-defective areas. . Therefore, the subtraction process is performed only on the coordinate component processed using the maximum value non-defective product reference data or only the coordinate component processed using the minimum value non-defective product reference data in each coordinate component of the first image processing data. As a process, the effect of the subtraction process can be obtained. However, in particular, when the subtraction process is performed using both the maximum value non-defective product reference data and the minimum value non-defective product reference data, the effect can be further enhanced. In other words, it is possible to accurately identify the defective areas of the two, which are roughly classified according to the chromaticity relationship between the defective area and the other non-defective areas, and to further increase the identification accuracy. Become.
[0019]
  Also,Appearance inspection of electronic circuit componentsTo reviewIn the subtraction process, when the processing using the maximum value non-defective product reference data or the minimum value non-defective product reference data performed for one coordinate component is used as a unit of the number of processes, the number of processes is two.May be.The number of subtraction processes referred to here is, for example, the processing performed when subtraction processing is performed on the data of one coordinate component in the first image processing data and the data in the maximum value non-defective product reference data corresponding to the coordinate component. The number is one.
[0020]
As described above, the subtraction process corresponds to each coordinate component in the first image processing data and each coordinate component of the first image processing data of at least one of the maximum value non-defective product reference data and the minimum value non-defective product reference data. This is a process of subtracting data from each other. In this case, as the number of coordinate components constituting the first image processing data increases, the effect of the subtraction process and thus the accuracy of specifying the defective area to be specified increases. However, the appearance inspection of electronic circuit components must be sequentially processed for a large number of electronic circuit components, and shortening the processing time is an important problem together with processing accuracy. Therefore, the subtraction processing is preferably performed when the processing using the maximum value non-defective product reference data or the minimum value non-defective product reference data performed for one coordinate component is a unit of the processing number. By setting the number of subtraction processes to two, it is possible to sufficiently obtain the effects of the processes, suppress an excessive increase in the processing time, and shorten the processing time. Here, when the number of subtraction processes is two, for example, for each of the two coordinate components of the first image processing data, the subtraction process is performed using the maximum value non-defective product reference data, or each of the minimum value non-defective products. Subtraction processing is performed by subtraction processing using the reference data, or by using one of the maximum value non-defective product reference data and the other of the subtraction processing using the minimum value non-defective product reference data. As described above, it is often difficult to unambiguously limit the chromaticity relationship and color difference between a defective area to be subjected to appearance inspection and another non-defective area. Therefore, when the subtraction process using the maximum value non-defective product reference data or the subtraction process using the minimum value non-defective product reference data for each of the two components of the first image processing data, When the chromaticity relationship with other non-defective regions is biased in either direction, the effect of the subtraction process is sufficiently enhanced. In addition, when the subtraction process using the maximum value non-defective product reference data for one of the two components of the first image processing data and the other is the minimum value non-defective product reference data, the defective area to be extracted is another non-defective product. When there is no bias in the chromaticity relationship with the area, the effect of the subtraction process is sufficiently enhanced. As described above, the coordinate component and the type of data used in the subtraction process are appropriately selected according to the inspection target when the appearance inspection of the electronic circuit component is performed. Of course, a method of subtracting one coordinate component of the first image processing data using the maximum value non-defective product reference data and the minimum value non-defective product reference data may be used. In this case, it is effective for one coordinate component in which the color difference between the defective area to be identified and the other non-defective area is easily clarified.
[0021]
  Appearance inspection of electronic circuit partsIn accordance with the size of the data representing the coordinate components of the maximum value non-defective product reference data or the minimum value non-defective product reference data, the data representing the coordinate components of the first image processing data corresponding to the coordinate components is The level is correctedMay be.
[0022]
In the inspection surface to be subjected to the appearance inspection of electronic circuit components, even when the same inspection surface in the same type of product is targeted, variation in color difference that occurs on the substrate surface is likely to occur. In some cases, the variation in color difference between a non-defective area such as a wiring pattern and a defective area becomes excessively large. This leads to the effect of subtraction processing being suppressed. Therefore, the size of the data representing the coordinate components of the maximum value non-defective product reference data or the minimum value non-defective product reference data, that is, the coordinates of the coordinate components of the first image processing data corresponding to the coordinate components in accordance with the coordinate values. The level of the value is corrected. As a result, it is possible to effectively suppress variations in color difference between a non-defective region such as a substrate surface and a wiring pattern and a defective region, and it is possible to make the effect of subtraction processing more useful. As the level correction process, for example, in the coordinate component corresponding to each coordinate component forming the first image processing data among the first inspection surface data and the second inspection surface data, data representing a predetermined region of the inspection surface, The data of each coordinate component of the maximum value non-defective product reference data or the minimum value reference data corresponding to the region is compared based on the coordinate value for the same coordinate component, and the first image processing data is matched to the comparison result. This can be done by a method of correcting the level of the coordinate value of each coordinate component forming Here, the predetermined area of the inspection surface for level correction includes, for example, a combination of an area on the substrate surface (for example, 100 × 100 pixels) and an area on the wiring pattern (for example, 50 × 50 pixels). Thus, it can also be configured from a plurality of individual areas. In that case, it is also possible to use a method in which individual level correction is performed on individual areas. In addition, the comparison process using data corresponding to the predetermined area of the inspection surface performed for level correction may be performed by averaging the coordinate values of the data corresponding to the predetermined area, or by averaging with a predetermined weight. There is no particular limitation such that the comparison processing is performed using a converted or Fourier-transformed one, and a known one can be used. As described above, the level correction processing method and the predetermined area (pixel area) on the inspection surface for level correction are not particularly limited, and the first image processing data is obtained from the non-defective area and the defective area such as the substrate surface and the wiring pattern. It is first important that the variation in color difference is effectively suppressed.
[0023]
  Next, in the present invention, when specifying the defective area on the inspection surface,
  Do not configure the first image processing data,For first inspection surface data and second inspection surface dataAccordingCoordinate componentInOne coordinate componentConsist ofSecond image processing dataThroneData representing the standard component;
  Among the non-defective product reference color information set in advance at each position in the inspection surface, the non-defective product reference color information data corresponding to the data representing the coordinate component of the second image processing data,
  The second defect candidate area is selected by differential processing or pattern matching processing,
  A defect area on the inspection surface is specified including the second defect candidate area and the first defect candidate area.
[0024]
In addition to the first defect candidate area selected in the above subtraction process, further including the second defect candidate area selected in the difference process and pattern matching process, by specifying the defect area on the inspection surface, Furthermore, the identification accuracy can be increased. Here, the inspection image data used in the difference processing and the pattern matching processing includes second image processing data including coordinate components that do not constitute the first image processing data among the first inspection surface data and the second inspection surface data. It is important to. That is, a coordinate component having a large color difference between the defective area and the non-defective area on the inspection surface is appropriately selected from the first inspection surface data and the second inspection surface data as the coordinate component of the first image processing data. Therefore, by constructing the second image processing data from coordinate components other than the coordinate components constituting the first image processing data, for example, at the periphery of the wiring pattern caused by the alignment accuracy between the inspection image and the reference image Excessive extraction of misaligned areas as defective areas is naturally suppressed, and defective areas that could not be selected by subtraction processing can be effectively extracted by other processing means such as difference processing and pattern matching processing. And Thus, by performing image processing by difference processing or pattern matching, it is possible to effectively select, for example, a defective portion such as a defective region on the wiring pattern or a defective region on the substrate surface that could not be selected by the subtraction processing. However, in particular, the coordinate component to be employed as the second image processing data is selected as appropriate, taking into account the proportion of the over-extraction of the misalignment area around the wiring pattern as a defective area.
[0025]
So far, in the present invention, it has been described that the accuracy of specifying the defective area is effectively increased by performing the subtraction process at least on the data subjected to the filter process. Therefore, next, a three-dimensional color space coordinate system used in this subtraction processing and further in image processing by difference processing and pattern matching will be described. The first is not particularly limited as long as it is a known three-dimensional color space coordinate system. Among them, the first inspection image data, that is, the three-dimensional color space coordinate system forming inspection surface color information is also included. Is preferably an RGB coordinate system having the coordinate components of the three primary colors of light, red R, green G and blue B. This is because a CCD camera is often used as a color light receiving unit for color imaging of the external appearance, and the detection output signal corresponds to the RGB coordinate system. As described above, the inspection surface color information is made up of the RGB coordinate system, so that when the inspection surface color information is generated based on the detection output data from the color light receiving unit, there is no need to perform coordinate conversion, Processing time can be shortened. Next, the second inspection surface data created by coordinate conversion of the first inspection surface data is preferably an HSI coordinate system having hue H, saturation S, and lightness I as coordinate components. This is because non-defective standard color information consisting of non-defective standard values for electronic circuit components must be set based on a judgment standard obtained by visual observation, and the HSI coordinate system is close to human color sensation. It is. In other words, by setting the second inspection surface data in the HSI coordinate system, the determination standard of the non-defective product reference color information, which is the non-defective product reference value, can be set with higher accuracy. Can be increased.
[0026]
Of course, the first inspection surface data and the second inspection surface data are not limited to the above-described three-dimensional color space coordinate system. For example, the RGB coordinate system and the HSI coordinate system include three independent coordinate components. However, it may be composed of four or more coordinate components obtained by adding coordinate components subordinate to the coordinate components. As described above, the three-dimensional color space coordinate system that forms the first inspection surface data and the second inspection surface data is appropriately selected in consideration of processing accuracy and processing time required in the appearance inspection processing. What is necessary is just to select from a well-known thing.
[0027]
  As described above, in the electronic circuit component manufacturing method of the present invention,First, in the process of forming a wiring pattern on a substrate, a wiring pattern corresponding to the element function of the electronic circuit component is formed. It is necessary to inspect the appearance of defective areas such as adhesion of foreign matter, discoloration, cracks, and peeling of the surface with respect to the appearance on the substrate surface on which such a wiring pattern is formed.FromThe visual inspection of the inspection surface on the substrate surface on which the wiring pattern is formed is performed in the determination process using the visual inspection method as described above.It has become.
In the determination step, first, the above-described appearance inspection methodThe lawThe defect area on the inspection surface is specified by using this. Further, the final pass / fail determination is performed by evaluating the shape such as the area and length of each part of the wiring pattern corresponding to the specified defective area.
By performing the determination process for the appearance of the inspection surface on the substrate surface in such a work process, the processing accuracy of the appearance inspection can be improved. As a result, it is possible to improve the quality as well as the reliability when the manufactured electronic circuit component is a product.
[0028]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic configuration diagram showing an example of an appearance inspection apparatus that can be used when performing the appearance inspection method for electronic circuit components of the present invention. The appearance inspection apparatus 100 includes a stage 20 on which the electronic circuit component 1 is placed. Further, the stage 20 can automatically position and convey the electronic circuit component 1 by the electric power from the movement control device 50. In addition, the appearance inspection apparatus 100 includes an illumination device 40 for illuminating light from the light source 30 onto the inspection surface of the electronic circuit component 1, a color light receiving unit 2, an image processing device 11, and the like. . The color light receiving unit 2 captures a color image of the appearance of the inspection surface of the electronic circuit component 1 to be inspected, and detects a color image input signal at each pixel corresponding to each position on the inspection surface. In addition, it has a function of outputting detection output signals corresponding to the three coordinate components of the three-dimensional color space coordinate system representing the color image in each pixel to the image processing apparatus 11. Next, as shown in FIG. 3, the image processing apparatus 11 includes a CPU 13 that performs functions such as arithmetic processing, a memory 14 that functions to store data, a hard disk 15 that functions to store and store data, and a color light receiving unit. 2 captures detection output signals corresponding to each of the three coordinate components of a three-dimensional color space coordinate system representing a color image, and captures an image having a function of A / D converting the analog signal into a digital signal. And a monitor 12 for displaying a captured image, a processing state, and the like. Such an image processing apparatus 11 can create image data, perform image processing, and the like. The form of the image processing apparatus 11 in FIGS. 1 and 3 is mainly a personal computer (PC), but a general-purpose computer such as a workstation can be used instead of the PC. In addition, an image processing apparatus specialized in image capture and processing functions can be made to function in a form of being externally connected to a computer. Furthermore, although the hard disk 15 is positioned as an auxiliary storage device here, the hard disk as the auxiliary storage device is an external storage device connected via a removable storage medium such as MO, CD-R, CD-RW, or LAN. It can be replaced with a storage medium. On the other hand, the memory 14 is used for storing image data, image processing, arithmetic processing, and the like. Of course, the function of the memory 14 can be performed by the hard disk 15.
[0029]
Next, the appearance inspection method according to the present invention will be described with reference to FIGS. FIG. 2 shows one work procedure of the appearance inspection method. The inspection surface of the electronic circuit component 1 placed on the stage 20 is irradiated with light from the light source 30 via the illumination device 40, and a color image of the appearance of the inspection surface of the electronic circuit component 1 is received in color. In addition to imaging by the unit 2, detection output signals corresponding to the three coordinate components of the three-dimensional color space coordinate system representing the color image at each pixel corresponding to each position on the inspection surface are sent to the image processing device 11. Output. Then, the image capturing board 16 of the image processing apparatus 11 performs A / D conversion on the input detection output signal and generates inspection surface color information as a digital signal. This inspection surface color information corresponds to each of three independent coordinate components of the three-dimensional color space coordinate system at each position on the inspection surface, that is, at each pixel position that partitions the color image of the appearance of the inspection surface. It consists of two color space coordinate values. Thus, the inspection surface color information includes the position information of each position on the inspection surface and the color image information at each position. Such inspection surface color information is taken into the image processing apparatus 11 as color image information on the inspection surface via the image capture board 16. This work corresponds to Step 1.
[0030]
The inspection surface color information input to the image processing apparatus 11 in step 1 is stored as a first inspection image in the memory 14 of the image processing apparatus 11. Then, the image processing apparatus 11 calculates the amount of deviation when there is a positional deviation in the first inspection image, compared with a reference image previously read into the memory 14. Thereafter, the position of the first inspection image is corrected based on the amount of deviation, and stored in the memory 14 as first inspection surface data. This work corresponds to Step 2.
[0031]
The first inspection plane data stored in the memory 14 in step 2 is coordinate-converted into a three-dimensional color space coordinate system different from its own three-dimensional color space coordinate system by arithmetic processing using the CPU 13 and the memory 14. The processing result is stored in the memory 14 as second inspection surface data. Then, first image processing data composed of at least one coordinate component in the first inspection surface data and the second inspection surface data is created using the image processing device 11 and stored in the memory 14. In addition, for each coordinate component of the color image information at each position of the inspection surface constituting the first inspection surface data and the second inspection surface data, a non-defective product reference value within a non-defective product allowable range is set in advance. Non-defective product reference color information including these non-defective product reference values is read in the memory 14 in advance. Further, each of the coordinate components in the non-defective product reference color information is subjected to a maximum value filter process and a minimum value filter process by an arithmetic process using the CPU 13 and the memory 14, respectively. The maximum value non-defective product reference data and the minimum value non-defective product reference data are stored in the memory 14 in advance. The CPU 13 and the data representing each coordinate component of the first image processing data and the data corresponding to each coordinate component of the first image processing data of at least one of the maximum value non-defective product reference data and the minimum value reference data, Subtraction processing is performed using the memory 14, and image processing is performed in which data having a value exceeding a predetermined threshold (for example, positive) or a value less than (for example, negative) is extracted. A set of positions in the inspection surface corresponding to the data extracted by this image processing is set as a first defect candidate area. This operation corresponds to the selection of the first defect candidate area in step 3. Further, the data extracted here is stored in the memory 14 as first extracted data.
[0032]
Then, after the operation of step 3, a defective area is specified based on the selected first defective candidate area. At this time, for example, the second image processing data composed of the coordinate components that do not constitute the first image processing data among the first inspection surface data and the second inspection surface data is used to perform the second processing by performing difference processing and pattern matching processing. When a defect candidate area is selected, the defect area is specified including this second defect candidate area. That is, in the case of only the first defect candidate area, the first defect candidate area and the defect area are in the same area. Including the image processing for selecting the second defect candidate area, the defect area is specified in step 4.
[0033]
Next, after the work in step 4, using the first extracted data stored in the memory 14 corresponding to the selected first defect candidate area, the defect area is identified in the inspection surface of the electronic circuit component 1. The image processing 11 analyzes the area and length of each of the parts, and creates shape evaluation data. If the shape evaluation data is extracted that exceeds a threshold set for each part of the wiring pattern formed in the inspection surface of the electronic circuit component 1, the electronic circuit component 1 is regarded as defective. judge. This work corresponds to pass / fail determination by shape evaluation in step 5.
[0034]
Although not described in the above work procedure, the non-defective reference image formed by each coordinate component constituting the non-defective reference color information is previously subjected to position correction by comparing with the reference image used in the position correction described above. It was made. As a matter of course, a mask process is performed in advance on a region of the electronic circuit component 1 that is not a target for appearance inspection. Now, the appearance inspection of the electronic circuit component 1 is performed according to the above work procedure, but this work procedure is merely an example. Known methods and procedures for position correction, data creation, calculation processing, and analysis in the image processing apparatus 11 can be applied. In addition, when specifying a defective area in step 4, it is also possible to use the second defective candidate area selected by difference processing or image processing by pattern matching.
[0035]
Next, a specific example of the three-dimensional color space coordinate system used for the first inspection surface data and the second inspection surface data and a specific example when performing difference processing and pattern matching in step 4 will be described below. Examples of the appearance inspection method of the invention will be described.
[0036]
(Example 1) As the color light receiving unit 2, one that can capture a color image for each of the three primary colors of light (red R, green G, and blue B) is used. For example, a CCD type three-plate color line sensor camera or the like is used. As shown in FIG. 8, by using the line sensor camera 2 as described above, the stage 20 on which the electronic circuit component 1 is mounted is moved in the X direction perpendicular to the camera scanning direction. A color image of the appearance of the inspection surface of the component 1 is taken into the image processing apparatus 11 as inspection surface color information for each of R, G, and B. In addition, as shown in FIG. 8, each captured image for each of R, G, and B is configured with a pixel as a minimum unit, and the size of each pixel is determined by the type of the line sensor camera 2 to be used and the type of lens. Defined by resolution. On the other hand, the magnitude of the luminance that represents the output value of the detection output that is output from the line sensor camera 2 to the image processing device 11 for each of R, G, and B is, for example, 1 byte of 0 to 255. Converted to a digital value. In this way, inspection surface color information composed of coordinate values corresponding to R, G, and B coordinate components at each pixel position corresponding to each position on the inspection surface is taken into the image processing apparatus 11. The inspection surface color information includes position information of each position on the inspection surface and color image information represented by the RGB coordinate system at each position.
[0037]
Next, the inspection surface color information input to the image processing apparatus 11 is stored in the memory 14 of the image processing apparatus 11 as a first inspection image. Then, it is compared with a reference image previously read into the memory 14, and if there is a positional shift in the first inspection image, the amount of shift is calculated in the image processing device 11. Thereafter, the position of the first inspection image is corrected based on the amount of deviation and stored in the memory 14 as first inspection surface data.
[0038]
After performing the above-described position correction, the first inspection plane data stored in the memory 14 is coordinate-converted into an HSI coordinate system different from its own RGB coordinate system by arithmetic processing using the CPU 13 and the memory 14, The processing result is stored in the memory 14 as second inspection surface data. Here, the first inspection surface data has three coordinate components of R, G, and B coordinate components as color image information, while the second inspection surface data is H (hue) and S (saturation). , I (lightness) coordinate components. The hue H represents a hue. As shown in FIG. 9A, the hue H is R (red), Y (yellow), G (green), C (cyan), B ( Blue) and M (magenta) are defined, and the hue is defined by the angle with respect to the center. Therefore, the coordinate value of the H coordinate component is obtained by assigning an angle in the range of 0 to 360 degrees indicating the hue to a value of 0 to 255 (1 byte). For example, if R is 0, G becomes 85, B becomes 170, and 255 becomes R again. Next, as shown in the HSI coordinate system of FIG. 9B, the saturation S representing the vividness of the color is defined by the length from the center. Therefore, a value obtained by assigning the length to a value of 0 to 255 (1 byte) is set as a coordinate value of the S coordinate component. Finally, the brightness I representing the brightness of the color is defined by the position on the central axis. Therefore, the position value assigned to 0 to 255 (1 byte) is used as the coordinate value of the I coordinate component.
[0039]
First image processing data including at least one coordinate component in the first inspection surface data and the second inspection surface data, in which the coordinate values of the respective coordinate components are defined as described above, is created using the image processing device 11. At the same time, it is stored in the memory 14. In addition, for each coordinate component of the color image information at each position of the inspection surface constituting the first inspection surface data and the second inspection surface data, a non-defective product reference value within a non-defective product allowable range is set in advance. Non-defective product reference color information including these non-defective product reference values is read in the memory 14 in advance. Then, based on the reference image used for the position correction of the first inspection image, the non-defective reference color information is assumed to have been subjected to position correction. Of course, the non-defective product reference color information can also be used as a reference image. Next, with respect to each coordinate component in such non-defective product reference color information, a maximum value filter process and a minimum value filter process are performed by arithmetic processing using the CPU 13 and the memory 14, respectively. The maximum value non-defective product reference data and the minimum value non-defective product reference data are stored in the memory 14 in advance. Then, data representing each coordinate component of the first image processing data and data corresponding to each coordinate component of the first image processing data of at least one of the maximum value non-defective product reference data and the minimum value non-defective product reference data are stored in the CPU 13. Subtraction processing is performed using the memory 14 and, for example, data using a maximum value non-defective product reference data is extracted as a positive value, and data using a minimum value non-defective product reference data is extracted as a negative value. Image processing is performed. A set of positions in the inspection surface corresponding to the data extracted by this image processing is set as a first defect candidate area. The data extracted here is stored in the memory 14 as first extracted data.
[0040]
In addition, before the subtraction process, the level (coordinate value) of each coordinate component of the first image processing data is level-corrected using the maximum value non-defective product reference data or the minimum value non-defective product reference image data. You can also. By performing the level correction in this way, it is possible to improve the processing accuracy of the first defect candidate area selected in the subtraction process. Here, as a method for level correction, for example, the data of each coordinate component of the first image processing data corresponding to a predetermined area of the inspection surface, and the maximum value non-defective product reference data or the minimum value non-defective product reference data corresponding to the region. A method is used in which the average value of these coordinate values is compared with the data, and the size of the first image processing data is corrected based on the difference value. Further, the level correction can be performed based on a comparison result between the first image processing data, the maximum value non-defective product reference data, and the minimum value non-defective product reference data. Furthermore, the first image processing data is individually compared with the maximum value non-defective product reference data and the minimum value non-defective product reference data, and the result of level correction based on each comparison result is stored in the memory as data. It may be left. In this case, each level-corrected data used in accordance with the maximum value non-defective product reference data and the minimum value non-defective product reference data used in the subtraction process is used. The level correction can also be performed by comparing the first image processing data and the good product reference color information.
[0041]
A defective area on the inspection surface is specified based on the first defective candidate area selected in the image processing by the subtraction process. When only the first defect candidate area is used, this first defect candidate area is set as a defect area. In this way, the defective area on the inspection surface is specified. In specifying the defective area, the coordinate components of the first image processing data to be used are appropriately selected as useful depending on the type of the electronic circuit component to be used and the item of the defective area. For example, when the appearance of the electronic circuit component is close to green (G), at least the G component is included in the first image processing data. Or, when the appearance is close to white, at least I (brightness) is included so that the types and number of coordinate components constituting the first image data are determined as appropriate. Further, at least one of the maximum value non-defective product reference data and the minimum value non-defective product reference data is appropriately selected in the subtraction process depending on the chromaticity relationship between the non-defective region and the defective region on the inspection surface.
[0042]
By performing visual inspection by image processing using the subtraction process as described above, it is possible to increase the processing accuracy of various inspection items in electronic circuit components in a way that appropriately corresponds to these inspection items. It becomes. Next, a case where a second defect candidate area other than the first defect candidate area is also used when specifying the defect area will be described.
[0043]
(Example 2) The procedure until the selection of the first defect candidate region is performed in the same manner as in Example 1. In addition to the selection of the first defect candidate area, the second defect candidate area is selected by image processing based on difference processing or pattern matching. First, the second image processing data including at least one coordinate component that does not constitute the first image processing data among the first inspection surface data and the second inspection surface data is created using the image processing apparatus 11 and stored in the memory 14. store. Then, non-defective product reference color information that has been read into the memory 14 in advance is created, and non-defective product reference data corresponding to the coordinate components of the second image processing data is created, and the non-defective product reference data and the second image processing data are the same. The CPU 13 and the memory 14 are used for the coordinate component, and image processing by difference processing or pattern matching is performed to extract data that exceeds a predetermined threshold value. A set of positions in the inspection surface corresponding to the data extracted by this image processing is set as a second defect candidate area. Further, the data extracted here is stored in the memory 14 as second extracted data.
[0044]
After extracting the second defect candidate area as described above, the area combined with the first defect candidate area is specified as the defect area. Then, by using the first extraction data representing the first defect candidate area and the second extraction data representing the second defect candidate area, the quality determination by the same method as the shape evaluation described above is performed, and the final determination is made. The quality of electronic circuit components is determined.
[0045]
As a coordinate component of the second image processing data used when selecting the second defect candidate area, of course, a coordinate component that can effectively select a defect area that cannot be selected in the first defect candidate area is appropriately selected. become. For example, referring to the schematic diagram of FIG. 10, in the captured image 70 that forms the inspection surface color information, the appearance of the substrate surface 60 is a green (G) component, and the substrate surface 60 has a defect mainly composed of a red (R) component. Consider the case where region 62 exists. First, if at least the R component is selected as the first image processing data, an area like the area 81 is selected as the first defect candidate area by the subtraction process. Next, by selecting the hue H as the second image processing data and performing image processing by difference processing or pattern matching, a region such as the region 82 is selected as the second defect candidate region. As shown in FIG. 10, in the second defect candidate area, a defect area that was not selected in the first defect candidate area may be selected, and a logical sum of the first defect candidate area and the second defect candidate area is calculated. By obtaining the area 83, the area 83 is specified as a defective area, so that the accuracy of specifying the defective area can be further increased. As a result, it is possible to further improve the processing accuracy of the appearance inspection.
[0046]
By using the appearance inspection method of the present invention described above, it is possible to increase the inspection accuracy of the appearance inspection performed on the inspection surface of the electronic circuit component. In addition, a determination process for inspecting the appearance on the substrate on which the wiring pattern corresponding to the element function of the electronic circuit component has been formed on the substrate is checked for adhesion, discoloration, cracks, surface peeling, etc. Also in the method for manufacturing an electronic circuit component, the inspection accuracy for the appearance can be improved by performing the determination step using the same method as the appearance inspection method of the present invention.
[0047]
The above-described embodiments and examples according to the present invention are merely examples, and any data that has been subjected to the maximum value filter processing and the minimum value filter processing in the appearance inspection of the inspection surface of the electronic circuit component is obtained. The concept of performing the subtraction processing used as image processing is included in the concept of the present invention.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram showing an example of an appearance inspection apparatus that can be used in an appearance inspection method of the present invention.
FIG. 2 is a schematic process diagram showing one work procedure of the appearance inspection method of the present invention.
FIG. 3 is a schematic configuration diagram of a main part of an appearance inspection apparatus according to the present invention.
FIG. 4 is a schematic diagram for explaining an appearance inspection method according to the present invention.
FIG. 5 is a schematic diagram for explaining a defect caused by a conventional appearance inspection method.
FIG. 6 is a schematic diagram for explaining an appearance inspection method according to the present invention.
FIG. 7 is a schematic diagram for explaining an appearance inspection method according to the present invention.
FIG. 8 is a schematic diagram for explaining an appearance inspection method according to the present invention.
FIG. 9 is a schematic diagram illustrating a three-dimensional color space coordinate system used in the appearance inspection method of the present invention.
FIG. 10 is a schematic diagram for explaining an appearance inspection method according to the present invention.
[Explanation of symbols]
1 Electronic circuit components
2 Color receiver (line sensor camera)
11 Image processing device
100 Visual inspection device

Claims (2)

Forming a wiring pattern on the substrate;
The color light receiving unit receives the inspection light from the inspection surface on the substrate on which the wiring pattern is formed by the process,
Based on the detection output of the color light receiving unit, generates inspection surface color information composed of three coordinate components of a three-dimensional color space coordinate system at each position in the inspection surface;
The first inspection plane data representing the three coordinate components of the inspection plane color information of the generated three-dimensional color space coordinate system is transformed into a three-dimensional color space coordinate system different from the three-dimensional color space coordinate system. The coordinate component of the first image processing data that includes the first inspection surface data represented by the three coordinate components and at least one coordinate component among the six coordinate components based on the second inspection surface data while creating the second inspection surface data And data representing
Of the preset good standard color information at each position in the inspection surface, the maximum value filtering process is performed on the good standard color information data corresponding to the data representing the coordinate component of the first image processing data. Data of the maximum value non-defective product reference data and the minimum value non-defective product reference data subjected to the minimum value filtering
By subtracting each, the first defect candidate area is selected,
Identifying the defective area on the inspection surface based on the first defective candidate area thus selected,
A determination step of determining pass / fail of the wiring pattern based on the specified defective area;
A method for manufacturing an electronic circuit component including:
The subtraction process for selecting the first defect candidate area is:
Using the maximum value non-defective product reference data and the minimum value non-defective product reference data of the same coordinate component as the coordinate component of the first image processing data, processing is performed using the coordinate component of the first image processing data and the maximum value non-defective product reference data. Between the coordinate component of the first image processing data and the coordinate component processed using the minimum value non-defective product reference data ,
Coordinates are processed using the minimum good reference data performed for subtraction or one coordinate component with coordinate components that are processed using the maximum value good reference data performed with respect to one coordinate component An electronic circuit component manufacturing method characterized in that when the subtraction processing using a component is a unit of processing number, the processing number is two. In identifying the defective area on the inspection surface,
Does not constitute the first image processing data, the data representing the coordinates component of the second image processing data consisting of a single coordinate component in the coordinate component by the first test surface data and the second test surface data,
Among the predetermined good standard color information at each position in the inspection surface, non-defective standard color information data corresponding to data representing coordinate components of the second image processing data,
The second defect candidate area is selected by differential processing or pattern matching processing,
2. The method for manufacturing an electronic circuit component according to claim 1, wherein a defective area on an inspection surface is specified including the second defective candidate area and the first defective candidate area.

Images