WO2000006997A1 - Processeur d'image - Google Patents
Processeur d'image Download PDFInfo
- Publication number
- WO2000006997A1 WO2000006997A1 PCT/JP1999/004096 JP9904096W WO0006997A1 WO 2000006997 A1 WO2000006997 A1 WO 2000006997A1 JP 9904096 W JP9904096 W JP 9904096W WO 0006997 A1 WO0006997 A1 WO 0006997A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- pixel
- mask
- image
- hole
- pattern
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
- G01N21/956—Inspecting patterns on the surface of objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30141—Printed circuit board [PCB]
Definitions
- the present invention relates to an image processing device.
- it can be used for pattern inspection equipment that can realize a wiring pattern inspection that eliminates power inspection omissions without erroneous detection by forming an appropriate mask for each pixel in a photo via hole that is image-recognized as an irregular shape. It relates to an image processing device. Background art
- An object to be inspected such as a printed wiring board may have a shape that is easily misidentified as a defect by the wiring pattern inspection device. For this reason, parts containing such shapes must be excluded from inspection. This corresponds to interlayer conductive holes such as through holes and photo via holes (hereinafter referred to as “via holes”).
- via holes interlayer conductive holes such as through holes and photo via holes (hereinafter referred to as “via holes”).
- the @ E @ pattern inspection apparatus described in, for example, Japanese Patent Laid-Open Publication No. Hei 6-2946426, separates the pattern image obtained by imaging the object to be inspected from the pattern image. By extracting true defect data based on the created hole mask image, the quality of the wiring pattern is inspected. That is, among the defect candidate data included in the pattern image, those that do not match the hole mask image are true defect data.
- This hole mask image was created by comparing and comparing the hole measurement signal indicating the part of the pattern image that has a shape considered to be a through hole with the hole recognition effective signal obtained from the specifications of the test object. You.
- the extraction of the hole measurement signal from the pattern image was performed as follows. In other words, as shown in Fig. 11, in the binarized pattern image, from the location where four pixels gather, the N, NE, E, SE, S, SW, W, and NW directions Count the number of dark pixels that continue. If the number of continuous dark pixels is within a predetermined range (upper and lower limits of the number corresponding to the through-hole radius) in any direction, the hole measurement signal is turned on for each pixel at that location. That is.
- the present invention has been made to solve the above-mentioned problems of the conventional technology. In other words, the challenge is to ensure the reliability of inspections around non-through holes by properly recognizing dark pixels in non-through holes and generating a minimum hole mask with high accuracy. It is an object of the present invention to provide an image processing apparatus that can be used for a wiring pattern inspection apparatus or the like that suppresses the occurrence of false alarms caused by non-through holes.
- An image processing apparatus for solving this problem includes: a pattern image creating means for creating a pattern image composed of pixel data obtained by reading a pattern of an object; For each pixel, the number of successive pixels of the same type from the pixel is counted for a plurality of directions, and a determination means for determining the magnitude of the count value and a predetermined value for each direction is provided. Means for generating a necessary-for-recognition-recognition signal indicating a range in which a required area may exist; and a recognition unit for determining whether or not the pixel is a required candidate pixel power based on the determination result of the determination means.
- the image processing apparatus further includes a processing-necessary candidate identifying unit and a processing-image creating unit that creates a processing image by comparing the processing-necessary pixel with the processing-necessary recognition effective signal.
- the wiring pattern inspection apparatus includes: a pattern image creating unit; a diameter determining unit; a mask recognition area creating unit; It has mask making means and inspection means.
- the pattern image creating means creates a pattern image consisting of pixel data obtained by reading the wiring pattern of the device under test. This pixel data may be binarized data.
- the diameter determination means measures, for each pixel included in the pattern image, the number of successive pixels of the same type from the pixel in a plurality of directions, and calculates the magnitude of the count value and a predetermined value for each direction. Is determined.
- the mask recognition area creating means creates a mask recognition effective signal indicating a range in which the non-inspection area can exist in the inspection object.
- the mask candidate certifying unit certifies whether or not the pixel is a mask candidate pixel based on the result of the determination by the diameter determining unit.
- the mask creating means creates a mask image by comparing the mask candidate pixel with the mask recognition valid signal.
- the inspection means extracts a defect shape from a portion other than the mask image in the pattern image.
- this wiring pattern inspection apparatus first, a wiring pattern of a device to be inspected is read. Then, a pattern image based on each pixel data included in the read wiring pattern image is created. If the pixel data is binarized, each pixel in the pattern image is either a dark pixel or a bright pixel. Usually, the pixels on the wiring pattern are bright pixels, and the pixels between the patterns are dark pixels. Ideally, all pixels in the via hole are ⁇ pixels. However, in practice, the brightness value before binarization is often a random value, and some parts become bright pixels due to the binarization, resulting in an irregularly shaped dark pixel group. The operations so far are performed by the pattern image creation means.
- the diameter of each pixel included in the pattern image is determined by the diameter determining means. Diameter judgment is performed as follows. In other words, starting from one pixel of interest, the number of consecutive dark pixels from that pixel is counted for a plurality of directions (4, 8, 12, and 16, etc.). That is, as shown in FIG. 1, the measurement start pixel in each direction is one pixel. In this point, it differs from the conventional one where the measurement start pixel is not always one as shown in Fig. 11. Therefore, the count value is 0 or a natural number. Then, the count value in each direction is compared with a predetermined value (upper limit value). If it is a natural number within that value, it is judged as good, and if it is 0 or exceeded, it is judged as bad. This is the diameter determination, which is performed for each pixel included in the pattern image.
- a predetermined value upper limit value
- the upper limit corresponds to the size of the area not to be inspected such as a via hole, but is not necessarily the same for all directions. This is because there is a possibility that the upper limit value will differ between the direction that is the same as the X or Y direction of the image and the oblique direction that is not the same. Note that the counting in each direction does not necessarily have to be performed until the dark pixel is interrupted, and may be stopped when the counted value exceeds the upper PSf.
- the mask recognition effective signal is generated by the mask recognition effective area creation means, and the mask recognition effective signal indicates the range where the non-inspection target area can exist in the inspected object.
- a standard position where a non-inspection area such as a via hole exists in the device under test is separately input. This can be obtained from the design data of the test object, or a measurement reference plate with only via holes, etc. can be created and obtained from the measurement results. Since the position of the non-inspection area slightly deviates from the standard position due to variations in individuals, a mask recognition effective signal is often created by enlarging this to an appropriate width. The count at this time may be determined according to the required accuracy of the test object.
- the result of the diameter determination is sent to mask candidate recognition means, and based on the determination result, whether or not the pixel is a mask candidate pixel is determined.
- This certification is preferably made as follows. That is, the number of directions in which the judgment result is good is counted. Then, the counted number of directions is compared with a predetermined threshold value, and if it is equal to or greater than the threshold value, the pixel of interest is identified as a mask candidate pixel. If the threshold is the total number of directions, the pixels that are identified as mask candidate pixels are pixels in the area that exists only within the diameter corresponding to the ⁇ pixel power upper limit. The dark pixels in the via hole This is an example. Pixels on the SH ⁇ pattern are bright pixels and are not mask candidate pixels. Although the pixels between the patterns are dark pixels, they are usually continuous over a long distance and do not fall within the diameter corresponding to the upper limit, so they are not considered as mask candidate pixels. The pinhole defect is a mask candidate pixel at this point, but is eliminated later.
- the above operation is performed for all the dark pixels included in the pattern image.
- those belonging to the non-inspection area such as via holes, those caused by defects such as pinholes, and the like are the force s mask candidate pixels.
- the predetermined threshold value to be compared with the number of counted directions is usually the same as the total number of directions (4 directions, 8 directions, 12 directions, 16 directions, etc.).
- the reason is as follows. Since a non-inspection area such as a via hole is usually surrounded by the land of the wiring pattern, even if it counts in any direction from the pixel of interest, it always hits the bright pixel of the land and stops counting within the upper limit. That is because it should be completed. If a dark pixel continues over a long distance beyond the above, it is usually considered that it is not a via hole or the like but a part between patterns. However, this is not always the case.
- the threshold value In order to identify this as a mask candidate pixel, the threshold value must be smaller than the total number of directions.
- the threshold value is variable and the variable range includes the total number of directions so that both cases can be handled.
- the mask candidate pixels and the mask recognition valid signal are obtained, they are compared by the mask creation means, and a mask image is created.
- mask candidate pixels that correspond to the mask recognition valid signal are extracted and used as a mask.
- mask candidate pixels other than those belonging to the non-inspection area such as via holes are excluded. Those caused by defects such as pinholes are repelled here and do not become masks.
- a mask candidate pixel that is equal to 1 "1” in the mask recognition effective signal is extracted, only the pixel itself may be used as a mask, or a slightly larger number may be used.
- the mask may be multiplied by a rate.
- the mask created in this way covers the minimum along the ⁇ pixel part of the binarized pattern image.
- the final inspection is performed by the inspection means. That is, a defect shape is extracted from a portion other than the mask image in the pattern image. Even if a defect shape exists in the mask image, it is ignored because it is thought to be due to via holes and the like.
- the mask created by the mask creation means is minimal, false alarms are prevented without impairing the reliability of the inspection near the non-inspection area such as via holes. .
- FIG. 1 is a diagram showing a method of counting the number of pixels in each direction in the wiring pattern inspection device of the present invention
- FIG. 2 is a diagram showing a block configuration of the wiring pattern inspection device according to the embodiment
- FIG. Fig. 4 shows a binarized image of a photo via hole
- Fig. 5 shows a binarized image of a pinhole
- Fig. 6 shows a binary image of the insulation between the wiring
- FIG. 8 illustrates the counting of the number of pixels in each direction in the wiring pattern inspection apparatus according to the embodiment
- FIG. 8 illustrates a diameter determination process using SRAM
- FIG. 9 is a diagram for explaining a mask determination process using SRAM
- FIG. 10 is a diagram for explaining breakout.
- FIG. 11 shows the method for counting the number of dark pixels in each direction using a conventional wiring pattern inspection device.
- Fig. 12 shows the photo via hole measurement using a conventional wiring pattern inspection device.
- Fig. 13 shows the conventional method.
- FIG. 4 is a diagram showing measurement of an insulated portion between lines by the wiring pattern inspection device of FIG.
- the wiring pattern inspection apparatus performs a quality inspection of a wiring pattern formed on a printed wiring board.
- this wiring pattern inspection apparatus has an imaging system 2, an inspection unit 31, and a mask processing unit 4 between them.
- a master CPU 32 for overall control and a terminal 33 for operator operation and display of inspection results are provided. Inspection unit 3 1 and master C P U
- the imaging system 2 has a function of reading a pattern 11 formed of a metal such as copper on the printed wiring board 10 and acquiring a pattern image.
- a CCD camera 21 that captures the ⁇ ⁇ pattern 11
- an AZD converter 22 that converts the analog image into a digital image
- a binarization circuit 23 that binarizes the digital signal.
- the pattern image PI output from the binarization circuit 23 is a binary image composed of dark pixels and bright pixels.
- the pixel at the location of the wiring pattern 11 is generally a bright pixel
- the pixel at the other location is a ⁇ pixel.
- Other locations include gaps between patterns, portions in via holes, and defects such as pinholes.
- the mask processing unit 4 has a function of creating a mask image indicating a portion of the pattern image to be excluded from the inspection target and supplying the mask image to the inspection unit 31 together with the pattern image. For this function, the mask processing unit 4 has the following blocks.
- the diameter determination unit 41 that receives the input of the pattern image PI from the binarization circuit 23 of the imaging system 2 and determines the diameter of the dark pixel for each pixel included therein, the via hole of the printed wiring board 10
- the hole information memory unit 4 2 that stores the data of the hole, the hole recognition effective based on the data DH of the via hole
- the hole recognition area creation unit 4 that creates the signal ME
- Hole determination unit 44 that compares with signal ME
- mask creation unit 45 that creates mask image MI based on the judgment result
- timing generation unit 46 that supplies a drive clock to each block
- a delay circuit 47 which delays the pattern image PI by the time required to create the mask image ⁇ Ml and supplies it to the inspection unit 31.
- Reference numeral 10 is a block for determining the diameter of a dark pixel as described above.
- the diameter judgment is the pixels obtained by excluding dark pixels included in the pattern image PI from those caused by gaps between patterns. This is because the gap between the patterns does not need to be used as a mask because it does not cause false information, and can be clearly distinguished from via holes and the like because it extends over a long distance along the wiring pattern 11.
- the judgment result MC is sent to the hole judgment unit 44.
- the hole information memory section 42 is a block for storing the data DH of the via hole of the printed wiring board 10 as described above. This is because via holes are likely to cause false information in the quality inspection of the printed wiring board 10, but their possible locations are determined for each type of printed wiring board 10.
- a method of preparing this data for example, there is a method of obtaining the data from the design data of the printed wiring board 10.
- a method of preparing a measurement reference plate and creating it from the pattern image PI for this Either method may be used, but here the method of obtaining from design data is adopted.
- s «in which only via holes are formed with the same specifications as the printed wiring board 10 can be used as the basis.
- the first one of a large number of printed wiring boards 10 of the same type manufactured in the same manufacturing lot may be used as the reference board.
- Data DH is? This is sent to the knowledge area creation unit 43.
- the hole recognition area creation unit 43 is a block for creating the hole recognition enable signal ME as described above.
- the data DH supplied from the hole information memory section 42 may be used as it is as the hole recognition effective signal ME, but here, the hole portion is enlarged by a predetermined width (number of pixels or magnification) to enable hole recognition.
- Signal ME The actual position of the via hole in the printed wiring board 10 is not always exactly as shown in the data DH, but is a force that varies slightly depending on the processing accuracy.
- the hole recognition effective signal ME covers pixels within a range where a via hole can exist in the printed wiring board 10. If there is a dark pixel that is a mask candidate in this range ⁇ , it is considered that it represents an actual via hole. Conversely, even if there is a ⁇ pixel which is a mask candidate outside this range, it has no relation to the via hole of the printed wiring board 10, so the mask should not be generated based on this.
- the created hole recognition valid signal ME is sent to the hole judgment unit 44.
- the hole determination unit 44 is a block that compares the determination result MC of the diameter determination unit with the hole recognition valid signal ME.
- the purpose is to select only those pixels that are truly caused by the via holes in the printed wiring board 10 from the dark pixels included in the pattern image PI. For this reason
- the block selects a dark pixel that is a mask candidate based on the decision result: 3 ⁇ 4MC, and compares it with the hole recognition enable signal ME. In other words, it is determined that the pixel ⁇ in the range of the hole recognition effective signal ME is a signal truly caused by the via hole.
- a signal outside the range of the hole recognition effective signal ME is judged not to be a signal due to a via hole.
- the signal MR of the pixel determined to be truly caused by the via hole is sent to the mask creation unit 45.
- the mask creation unit 45 is a block that creates a mask image MI based on the determination result of the hole determination unit 44, as described above.
- this block basically, for each pixel included in the hole recognition signal MR sent from the hole determination unit 44, one pixel is used as the mask image Ml. As a result, a mask that covers the via hole almost completely is obtained. Then, a certain magnification (9 times, etc.) may be applied to each pixel included in the signal MR according to the required inspection accuracy. In that case, a slightly larger mask (9 pixels larger by one pixel width) can be obtained.
- the created mask image Ml is sent to the inspection unit 31.
- the inspection unit 31 detects a shape that is ultimately considered to be a pattern defect from the pattern image PI, and outputs the detection signal DD to the master CPU 32. Therefore, the mask image MI is input to the inspection unit 31 from the mask creation unit 45, and the pattern image PI is input via the delay circuit 47 to match the timing with the input of the mask image Ml. It has become. Then, the inspection unit 31 extracts a defect shape such as a disconnection, a short circuit, or a chip in a range of the pattern image PI other than a range not to be inspected by the mask image Ml. Defect shape data for that purpose may be stored in the inspection unit 31 in advance, or may be supplied from the master CPU 32.
- the operation of the printed circuit board 10 for quality inspection by this wiring pattern inspection apparatus will be described.
- the via hole data DH obtained from the design data of the product type is stored in the hole information memory unit 42 in advance.
- the printed board 10 to be inspected is placed under the CCD camera 21 and the wiring pattern 11 is imaged.
- the analog image is converted into a digital signal by the A / D converter 22 and further binarized by the binary converter 23.
- a pattern image PI of a binary image consisting of dark pixels and bright pixels is obtained.
- the obtained pattern image PI is sent to the mask processing unit 4 and input to the diameter determination unit 41.
- the pattern image PI has a shape composed of ⁇ pixels, such as through holes (Fig. 3), photo via holes (Fig. 4), pinhole defects (Fig. 5), and line spacing (Fig. 6). include.
- the photovia hole in Fig. 4 has an irregular shape for the reason explained in the section of the prior art.
- the diameter determination unit 41 performs the following operation on the dark pixels included in the pattern image PI. That is, as shown in Fig. 7, we focus on a certain dark pixel A, and from that dark pixel A, a number of ⁇ pixels continue in each of the eight directions N, NE, E, SE, S, SW, W, and NW. To count. At this time, the pixel of interest A is counted as one. In Fig. 7, the count value for the direction E is “2” and the count value for the direction W is “4”, as indicated by the numbers in parentheses to the right of each direction symbol. The count values in all other directions are “1”.
- the count values for each direction are obtained, the count values are calculated based on the number of pixels corresponding to the design value of the diameter of the through hole and the photo via hole included in the printed wiring board 10 (in this example, the photo via hole).
- the diameter is compared with the number of pixels “1 1”) to determine whether it is OK or NG.
- a lookup table as shown in Table I is prepared in the diameter determination unit 41.
- the “count value” in Table I is expressed in binary in the actual lookup table.
- the output value “0” is assigned to the count values “0” and “1 2” or more, and the output value “1” is assigned to the count values “1” to “1 1”. Assigned.
- An output value of “0” means that the count result in that direction is NG, and an output value of “1” means that the count result is OK.
- the maximum count “1 1” to which the output value “1” is assigned is the number of pixels corresponding to the design value of the diameter of the via hole. From this, if the count value is a positive number less than or equal to the maximum diameter of “1 1”, the direction is determined to be OK, and if the count value is out of this range, the direction is determined to be NG. You. Multiple types of such look-up tables are prepared for each size of via hole.
- the judgment result of NG when the count value is ⁇ 0 ” means that if the target pixel is a bright pixel, it cannot be recognized as a mask candidate. If the count value is " 1 2 ”If the judgment result is NG in this case, it means that the pixel is not considered to be a via hole because the dark pixel continues over a long distance. As a result, if the target pixel is A in Fig. 7, it is determined that all directions are OK. Similarly, if the target pixel belongs to a through-hole (Fig. 3) or a pinhole defect (Fig. 5), it is determined that all directions are OK. On the other hand, if the target pixel belongs to the space between the lines (Fig. 6), the direction N and the direction S are determined to be NG.
- the diameter determination unit 41 performs this determination using SRAM as shown in FIG. That is, in the system of FIG. 8, the count value is input to the address line (A0 to A3) of the SRAM via the multiplexer, and the output value (0 ( NG) or 1 (OK)) is output to the SRAM data lines (I / O 0 to I / O 3). Since the count value is the number of directions shown in Fig. 7, that is, eight, the circuits of Fig. 8 for these directions are prepared, and this process is performed in parallel. In Fig. 8, since the SRAM data line has four bits, it is possible to set four different criteria for determining whether it is OK or NG. However, only one setting is made here.
- the four bits can be used as a look-up table that simultaneously determines four types of diameters with different hole diameters.
- the output value of each direction (hereinafter referred to as “diameter judgment value MC”) is sent to the hole judgment unit 44. This process is performed for each pixel of the pattern image PI.
- the CPU address is connected to the CPU data bus, and the specified data is input. The above is the processing in the diameter determining unit 41.
- the hole information memory section 42 and the hole recognition area creating section 43 create a hole recognition effective signal ME.
- the hole information memory section 42 is provided with data DH relating to the position and size of the via hole in the printed wiring board 10 to be inspected, which is extracted from the design data.
- Data DH has a value of “1” for pixels in the via hole and a value of “0” for other pixels. If there are a maximum of eight types of via holes, the data width of the hole information memory section 42 only needs to be 8 bits. Then, the hole recognition area creating section 43 may create the mask recognition effective signal ME by expanding the data DH by a predetermined pixel width.
- the created validity signal ME is sent to the hole judgment unit 44.
- the hole judgment value MC of each direction for each pixel of the pattern image PI is input from the diameter judgment unit 41 to the hole judgment ⁇ 44.
- the hole recognition valid signal ME is input from the hole recognition area creation unit 43.
- a signal of the number of continuous OK lines is input from the mask processing CPU unit 49.
- the hole determination unit 44 performs the hole determination for each dark pixel of the pattern image ⁇ I.
- This hole judgment is performed using SRAM as shown in Fig. 9. That is, in the system of Fig. 9, the diameter judgment value MC of each direction for a certain dark pixel of the pattern image PI is input to the SRAM address line. In addition, the signal of the continuous OK number and the hole recognition valid signal ME are also input. As a result, the hole recognition signal MR is created and output from the data line as follows. First, among the diameter judgment values MC, the value is “1”, that is, the number of directions in which OK is continuous is counted. For example, if the pixel of interest is A shown in Fig. 7, the judgment value is OK for all directions, and the count value is “8”. Similarly, if the pixel of interest belongs to a through-hole (Fig.
- the count value is “8”.
- the direction N and the direction S are unsuccessful. “NE, E, SE” and “SW, W, NWJ”, and the count value is “3”.
- this count value is compared with the continuous OK number. If the count value is equal to or greater than the number of continuous OK lines, the pixel of interest is determined to be a mask. If the count value is less than the number of consecutive OKs, the pixel of interest is not a mask candidate. For example, if the target pixel is A shown in Fig. 7, the count value is "8", and it is considered as a mask candidate. Similarly, if the pixel of interest belongs to a through-hole (Fig. 3) or a pinhole defect (Fig. 5), it is also considered as a mask candidate. On the other hand, if the pixel of interest belongs to the space between lines (Fig.
- the count value is “3”, which is less than the number of consecutive OK lines, and is not considered a mask candidate.
- the reason why the number of continuous OK lines is the same as the total number of directions is that there should be no break in the land around the via hole in the normal printed wiring board 10. However, if the grade of processing accuracy is low due to the specifications of the printed wiring board 10, some breakage may be acceptable. In such a case, as shown in Fig. 10, the land is cut off at some point, and the direction passing through it may be NG. In this case, such a pixel is not considered as a mask candidate. 997
- the number of consecutive OKs slightly smaller than the total number of directions, or by fixing the result of a specific direction to “OK”.
- the number of continuous lines is variable within a range including the total number of directions, can be set within the variable range, and the result of the predetermined direction can be fixed for a predetermined hole.
- the target pixel is a mask candidate, it is further sorted by the hole recognition effective signal ME. That is, only a pixel that is recognized as a mask candidate and the position of the pixel is within the position range of the hole recognition enable signal ME is selected and determined as a true hole recognition signal. By this selection, mask candidate pixels outside the range of the hole recognition valid signal ME are excluded. Mask candidate pixels outside the range of the hole recognition effective signal ME cannot be based on via holes, but are thought to be based on defects such as pinholes. Therefore, it is inappropriate to use such a pixel as a mask.
- the above-described recognition of mask candidates and selection by the hole recognition valid signal ME are performed for each pixel included in the pattern image PI.
- the hole recognition signal MR sent from the hole judging unit 44 to the mask making unit 45 is information on the pixels that have finally been selected as the true reference for mask making.
- the mask creating unit 45 receives the input of the hole recognition signal MR and creates a mask image Ml. That is, one pixel is used as a mask pixel for each pixel included in the hole recognition signal MR. As a result, a mask image Ml that covers the via hole barely is obtained (in the case of FIG. 4, all the “0” pixels are used as masks as they are). However, each pixel included in the hole recognition signal MR may be multiplied to some extent (eg, 9 times). In this case, a slightly larger mask (9 pixels larger by one pixel width) can be obtained (in the case of Fig. 4, the mask image that is almost the same as the one covered above is enlarged by one pixel). Thus, the mask creation unit 45 can be said to be an enlargement filter for obtaining an arbitrary mask size.
- the created mask image Ml represents a part of the pattern image PI which is not to be subjected to the pass / fail inspection.
- the inspection unit 31 After the mask image Ml has been created in this way, the inspection unit 31 performs an inspection process. Therefore, the inspection unit 31 receives the mask image Ml and the pattern image PI. By the way, the pattern image PI input to the inspection unit 31 is subjected to the delay processing for the mask processing time by the delay circuit 47 in order to align with the mask image Ml. In addition, the inspection unit 31 is supplied with data on defect shapes such as disconnection, short circuit, and chipping from the master CPU 32 in advance and stores them. As a result, the inspection unit 31 determines the defect type in the range other than the mask image Ml in the pattern image PI. A process for extracting a shape or a shape similar thereto is performed.
- the data DD such as the extracted defect shape is sent to the master CPU 32 and totalized.
- the correct wiring pattern may include a shape similar to the above-mentioned defect. In such a case, it is preferable to prepare in advance the inspection unit 31 with data excluding this from the inspection target.
- the defect data collected by the master CPU 32 can be displayed and output through the terminal 33.
- the inspection of the printed wiring board 10 is performed as described above. Then, even when the pass / fail inspection is continuously performed on the same type of printed wiring board 10 belonging to the same manufacturing lot, the hole recognition signal MR is extracted from the pattern image PI of the printed wiring board 10 in the same manner as the first printed wiring board. Then, a mask image Ml is created again. This is because the pass / fail inspection is performed using a highly accurate mask image MI that matches the hole shape of the printed wiring board 10.
- the diameter determination unit 4 determines each dark pixel included in the pattern image PI obtained by binarizing the imaging data of the printed wiring board 10. At 1, the number of ⁇ pixels continuing from the dark pixel in every eight directions is counted, and a diameter judgment 6MC is output.
- the count value of the pixel of interest is counted as “1” in each direction, ⁇ pixels can be extracted without overlooking the one-pixel part. Therefore, even if a dark pixel group with a distorted shape like a via hole is formed, it can be extracted without fail and used as the hole recognition signal MR.
- through-hole extraction based on radius reference judgment is also possible as in the past.
- the mask making section 45 can make an appropriate mask that covers the hole shape barely. As a result, an efficiency pattern inspection apparatus has been realized while eliminating the occurrence of false alarms due to via holes and the like, and preventing the occurrence of inspection omissions.
- the value of the continuous OK number is used as a criterion for extraction.
- the value is variable within the range including the total number of directions, and the value can be set within the variable range. Even if they are used, the inspection can be performed using a mask image that is appropriate for the use, so that it is possible to perform inspections without leaks while suppressing the occurrence of false alarms.
- a mask image may be created at a higher speed by determining the diameter of the pattern image PI based on the hole recognition valid signal ME.
- the number of directions for diameter determination in the diameter determination unit 41 is not limited to “8”. For example, “4”, “1 2”, “1 6”, etc. are conceivable, and need not be a multiple of 4.
- the via hole data DH supplied from the hole information memory section 42 to the hole recognition area creating section 43 may be prepared by another means instead of being taken from the design data of the printed wiring board 10.
- a pattern image PI is created for a reference board in which the same holes are formed as the printed wiring board 10 to be inspected, and the processing performed by the hole determination unit 44 (FIG. 9) for the hole recognition valid signal ME is performed.
- a hole recognition signal MR is created using only the diameter judgment value MC and the number of continuous OKs without using, and this is fed back to the hole information memory unit 42 to become data DH.
- the hole recognition signal MR may be fed back to the hole information memory unit 42 each time the printed wiring board 10 is inspected to have a learning effect.
- the present invention it is possible to appropriately recognize the dark pixel portion even in a non-through hole such as a photo via hole whose bottom is dull and shine, and to generate a minimum hole mask with high accuracy. As a result, it is possible to provide a wiring pattern inspection apparatus that suppresses the occurrence of false alarms caused by non-through holes while ensuring the reliability of inspection around the non-through holes.
- the present invention is an image processing apparatus capable of generating a minimum processed image with high accuracy
- its application is not limited to the above-described wiring pattern inspection apparatus, but may be applied to image processing in various industrial fields. Used for equipment.
Landscapes
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Immunology (AREA)
- Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Pathology (AREA)
- Analytical Chemistry (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Processing (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
- Internal Circuitry In Semiconductor Integrated Circuit Devices (AREA)
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP99933175A EP1109009B1 (en) | 1998-07-31 | 1999-07-29 | Wiring pattern inspecting apparatus |
AU49307/99A AU4930799A (en) | 1998-07-31 | 1999-07-29 | Image processor |
DE69935793T DE69935793T2 (de) | 1998-07-31 | 1999-07-29 | Vorrichtung zur bestimmung der leiterbahnstruktur |
US09/772,927 US6636632B2 (en) | 1998-07-31 | 2001-01-31 | Image processor and image processing method |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP10/216772 | 1998-07-31 | ||
JP21677298A JP4071866B2 (ja) | 1998-07-31 | 1998-07-31 | 配線パターン検査装置 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US09/772,927 Continuation US6636632B2 (en) | 1998-07-31 | 2001-01-31 | Image processor and image processing method |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2000006997A1 true WO2000006997A1 (fr) | 2000-02-10 |
Family
ID=16693664
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP1999/004096 WO2000006997A1 (fr) | 1998-07-31 | 1999-07-29 | Processeur d'image |
Country Status (6)
Country | Link |
---|---|
US (1) | US6636632B2 (ja) |
EP (1) | EP1109009B1 (ja) |
JP (1) | JP4071866B2 (ja) |
AU (1) | AU4930799A (ja) |
DE (1) | DE69935793T2 (ja) |
WO (1) | WO2000006997A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1302280C (zh) * | 2000-04-27 | 2007-02-28 | 精工爱普生株式会社 | 通孔内异物检查方法 |
Families Citing this family (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002163638A (ja) * | 2000-11-29 | 2002-06-07 | Ibiden Co Ltd | 画像データ検査装置および画像データ検査方法 |
US7813638B2 (en) * | 2004-06-07 | 2010-10-12 | Rudolph Technologies, Inc. | System for generating camera triggers |
KR100773332B1 (ko) * | 2006-04-11 | 2007-11-05 | 한국전자통신연구원 | 변조 장치, 복조 장치 및 무선 모뎀 |
KR100699899B1 (ko) * | 2006-05-08 | 2007-03-28 | 삼성전자주식회사 | 집적회로 장치 제조용 마스크 검사 장치 및 그 검사 방법 |
US8068674B2 (en) * | 2007-09-04 | 2011-11-29 | Evolution Robotics Retail, Inc. | UPC substitution fraud prevention |
US8935107B1 (en) * | 2011-02-18 | 2015-01-13 | The United States of America as Represented by the Adminstrator of National Aeronautics and Space Adminstration | Shock sensing system and method |
US9117290B2 (en) * | 2012-07-20 | 2015-08-25 | Samsung Electronics Co., Ltd. | Apparatus and method for filling hole area of image |
US10762618B1 (en) * | 2019-02-14 | 2020-09-01 | United Microelectronics Corp. | Mask weak pattern recognition apparatus and mask weak pattern recognition method |
US11631169B2 (en) * | 2020-08-02 | 2023-04-18 | KLA Corp. | Inspection of noisy patterned features |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06294626A (ja) * | 1993-04-08 | 1994-10-21 | Ibiden Co Ltd | プリント配線基板の検査装置 |
JPH0720062A (ja) * | 1993-06-15 | 1995-01-24 | Nikon Corp | 画像検出装置 |
JPH11166903A (ja) * | 1997-12-03 | 1999-06-22 | Fujitsu Ltd | バイアホール検査装置 |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4578810A (en) * | 1983-08-08 | 1986-03-25 | Itek Corporation | System for printed circuit board defect detection |
GB2152658A (en) * | 1984-01-09 | 1985-08-07 | Philips Electronic Associated | Object sorting system |
KR900007548B1 (ko) * | 1985-10-04 | 1990-10-15 | 다이닛뽕스쿠링세이소오 가부시키가이샤 | 패턴 마스킹 방법 및 그 장치 |
DE3786699T2 (de) * | 1986-05-10 | 1993-11-11 | Fujitsu Ltd | System zur Untersuchung von Mustern. |
JP2696000B2 (ja) * | 1991-02-08 | 1998-01-14 | 大日本スクリーン製造株式会社 | プリント基板のパターン検査方法 |
JP3189515B2 (ja) * | 1993-07-09 | 2001-07-16 | イビデン株式会社 | プリント配線基板の検査装置 |
US5608816A (en) * | 1993-12-24 | 1997-03-04 | Matsushita Electric Industrial Co., Ltd. | Apparatus for inspecting a wiring pattern according to a micro-inspection and a macro-inspection performed in parallel |
-
1998
- 1998-07-31 JP JP21677298A patent/JP4071866B2/ja not_active Expired - Fee Related
-
1999
- 1999-07-29 WO PCT/JP1999/004096 patent/WO2000006997A1/ja active IP Right Grant
- 1999-07-29 AU AU49307/99A patent/AU4930799A/en not_active Abandoned
- 1999-07-29 EP EP99933175A patent/EP1109009B1/en not_active Expired - Lifetime
- 1999-07-29 DE DE69935793T patent/DE69935793T2/de not_active Expired - Lifetime
-
2001
- 2001-01-31 US US09/772,927 patent/US6636632B2/en not_active Expired - Lifetime
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06294626A (ja) * | 1993-04-08 | 1994-10-21 | Ibiden Co Ltd | プリント配線基板の検査装置 |
JPH0720062A (ja) * | 1993-06-15 | 1995-01-24 | Nikon Corp | 画像検出装置 |
JPH11166903A (ja) * | 1997-12-03 | 1999-06-22 | Fujitsu Ltd | バイアホール検査装置 |
Non-Patent Citations (1)
Title |
---|
See also references of EP1109009A4 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1302280C (zh) * | 2000-04-27 | 2007-02-28 | 精工爱普生株式会社 | 通孔内异物检查方法 |
Also Published As
Publication number | Publication date |
---|---|
EP1109009B1 (en) | 2007-04-11 |
AU4930799A (en) | 2000-02-21 |
US20010053242A1 (en) | 2001-12-20 |
US6636632B2 (en) | 2003-10-21 |
DE69935793T2 (de) | 2007-12-27 |
EP1109009A4 (en) | 2002-05-08 |
JP2000046745A (ja) | 2000-02-18 |
JP4071866B2 (ja) | 2008-04-02 |
DE69935793D1 (de) | 2007-05-24 |
EP1109009A1 (en) | 2001-06-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JPS6261390A (ja) | プリント基板検査方法およびその装置 | |
JP6143445B2 (ja) | ビアホールを検査する方法及び装置 | |
WO2000006997A1 (fr) | Processeur d'image | |
JP3589424B1 (ja) | 基板検査装置 | |
JP2003086919A (ja) | パターン検査装置 | |
JP2745778B2 (ja) | 配線パターン検査装置 | |
JP2007003494A (ja) | 配線パターンの検査方法およびその装置 | |
JP4084969B2 (ja) | パターン検査装置及びパターン検査方法 | |
JP3581040B2 (ja) | 配線パターン検査方法 | |
JPH04184244A (ja) | プリント基板のパターン検査方法 | |
JP2021056004A (ja) | 画像判定装置及び画像判定方法 | |
JPS6186638A (ja) | 配線パターン欠陥検査方法 | |
JP4010838B2 (ja) | パターン検査装置 | |
JP2676990B2 (ja) | 配線パターン検査装置 | |
JP2009236786A (ja) | 貫通孔の検査方法および貫通孔の検査装置 | |
JP2002163638A (ja) | 画像データ検査装置および画像データ検査方法 | |
JP2001183119A (ja) | パターン検査装置 | |
JP4009085B2 (ja) | パターン検査装置およびパターン検査方法 | |
JP3210713B2 (ja) | 所定の特徴及び許容差の識別のための画像化パターンの収縮、拡張及び処理を用いた幾何学的パターン検査方法及び装置 | |
JP2002372502A (ja) | プリント回路板の検査方法 | |
JP2677052B2 (ja) | スルーホール検査装置 | |
CN117115523A (zh) | 一种缺陷检测方法、装置、计算机设备、系统及空调器 | |
JP2003270171A (ja) | プリント回路板検査装置 | |
JPH0619252B2 (ja) | 印刷配線基板のはんだ付検査装置 | |
JPS5892869A (ja) | 配線パターンの欠陥判定方法およびその装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AU CA CN IN KR US |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
WWE | Wipo information: entry into national phase |
Ref document number: 09772927 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1999933175 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 1999933175 Country of ref document: EP |
|
WWG | Wipo information: grant in national office |
Ref document number: 1999933175 Country of ref document: EP |