JP2998518B2 - Pattern inspection equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern inspection apparatus, and more particularly, to a pattern inspection apparatus for detecting a thin line width and a wide line width defect in an appearance inspection of a wiring pattern such as a printed circuit board.

[0002]

2. Description of the Related Art Conventionally, for example, "Printed circuit board inspection using radial matching", Journal of Precision Engineering, Vo1.
56, No. 8, pp. 30 to 33, 1990, there is a pattern inspection apparatus using a dedicated length measuring sensor and a pattern code.

Hereinafter, a conventional pattern inspection apparatus will be briefly described with reference to the drawings.

FIG. 11 is an explanatory diagram for explaining an inspection method of a pattern inspection apparatus using the above-described length measuring sensor. FIG. 11 shows a length measurement sensor. The length measurement sensor is composed of 16 length measurement pixels 102 extending radially from the inspection center 101. Each measurement pixel 102 is the inspection center 10
A target pattern is applied to a pair of symmetrically measured pixels 102 with respect to 1, and data measured by each of the measured pixels 102 is encoded using the length and symmetry of the measured pixels as elements. Next, the coded data is compared with a coded dictionary created in advance to determine whether the code of the target pattern is that of a normal pattern or that of a defective pattern.

FIGS. 12 (a), 12 (b) and 12 (c) are schematic diagrams showing examples of coding judgment by a length measuring sensor. FIG.
(A) shows the case of a normal pattern, and (b) shows the case of a thin line defect. In this case, when compared to the normal pattern,
The length-measuring pixel pairs 103 in the vertical (90 °) direction of the pattern are equal, but the symmetry of the length-measuring pixel pairs 104 in the left-right (0 °) direction is preserved, but the length becomes shorter,
The pixel pair 105 in the (°) direction becomes asymmetric. On the other hand (c)
Is a case of a line thick defect, but as compared with a normal pattern, the vertical (90 °) direction and the oblique (45 °) direction are equal to the normal pattern, but the horizontal (0 °) direction is asymmetric.

[0006]

In the above-described conventional pattern inspection apparatus, the pattern state at the center of the inspection is coded using a radially extending length measuring sensor, and the defect is detected by comparing it with a dictionary code prepared in advance. Therefore, there is a problem that it is necessary to execute complicated processing such as coding and to create a strict dictionary code for a complicated pattern without omission.

[0007]

A pattern inspection apparatus according to the present invention scans a target wiring pattern and outputs a video signal corresponding to reflected light of the target wiring pattern, and converts a video signal from the photoelectric conversion scanner into two. A binarizing circuit for binarizing a value image signal, a thinning circuit for sequentially performing thinning processing on binary image data obtained from the binarizing circuit to generate thin line image data, and a thin line of the thinning circuit A thin line processing control unit that controls the number of processing stages based on the line width reference tolerance to generate thin line image data for detecting a thin line defect and thin line image data for detecting a thick line defect; and the thin line controlled by the thin line processing control unit. A thin defect image memory for storing thin defect detection image data generated by the thinning circuit, and a thick defect detection memory controlled by the thin line processing control unit and generated by the thinning circuit. A defect detecting fine line image memory thickening and stores the image data, the thinning defect detecting fine line image picture data in the memory is read this for the line thinning section detects 3
A thin defect detecting section for extracting a portion having a line width of 1 pixel or more detected by applying a × 3 mask as a thin defect, and reading image data from the thin line image memory for detecting a thick defect to obtain a thick line. A thick defect detecting section for extracting a part having a line width of 2 pixels or more detected by applying the part detection 3 × 3 mask as a thick defect part.

A pattern inspection apparatus according to the present invention scans a target wiring pattern and outputs a video signal corresponding to the reflected light, and converts the video signal from the photoelectric conversion scanner into a binary image signal. A binarization circuit for performing binarization, a thinning circuit for sequentially performing thinning processing on binarized image data obtained from the binarization circuit to generate thin line image data having a line width of one pixel, and a thinning circuit generated by the thinning circuit. A first line width enlarging circuit configured to perform an enlarging process on the thin line image data to reconstruct a thin line width pattern that is a lower limit line width pattern serving as a reference for line width thinning determination; A second line width enlargement circuit configured to perform an enlargement process on the generated thin line image data to reconstruct a thick line width pattern that is a line width pattern of an upper limit value of the line width thickening determination; Binarization circuit A thinned portion extraction circuit for extracting a portion protruding from the obtained binary image as a thinned portion, and a thinned portion extracting a portion where the binary image obtained from the binarization circuit protrudes from the thick line width pattern. And a fat part extraction circuit for extracting the fat part.

[0009]

Next, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing the configuration and processing flow of an embodiment of the present invention. Hereinafter, the operation of this embodiment will be described along the flow of FIG.

The video signal d obtained by scanning the wiring pattern to be inspected by the photoelectric conversion scanner 1 is converted into a binary signal by the binarization circuit 2 so that the pattern portion has a value of "1" and the other portions have a value of "0". To obtain binary image data e.

Next, in the thinning circuit 3, the thin line processing controller 4 performs thin line processing of the number of thin line processing stages specified based on the line width reference tolerance. First, the thin line processing controller 4 sets the line width pattern matching the line width reference tolerance lower limit value to the line width "1" until the line width pattern matches the line width pattern tolerance and the line width reference tolerance upper limit value to the line width "line width". Step number setting data f1 and f2 are set, each of which indicates the number of fine line processing steps up to 1 ”.
The thinning circuit 3 first stores the thin line image data g in the thin line image memory 5 for detecting a thinned defect portion when the number of thin line processing stages reaches the established stage number f1. Then, the processing is terminated, and the thin line image data h at that time is stored in the thin line image memory 6 for detecting a thickened defect. Each thin line image data is a binary image in which the pattern portion is “1” and the others are “0”.

FIGS. 2 and 3 are schematic diagrams showing examples of the operation of the binarization circuit 2 and the thinning circuit 3. FIG.

FIGS. 2A and 3A show a binary image 10 based on the basic binary image data e, in which a thin line portion a and a thick line portion b exist, respectively. This 2
A thin line image 11 of thin line image data g for thin defect detection and a thin line image 12 of thin line image data h for thick part detection obtained by performing thin line processing by the thin line processing circuit 3 on the value image 10 by the thin line processing control unit 4 are shown. 2 (b) and FIG. 3 (b). This corresponds to a case where a thin line portion a and a thick line portion b exist, respectively.

A characteristic feature of the thin line image 11 for detecting a thin defect portion is that the line width of the thin line image 11 is "2" or more in the number of pixels in the normal pattern line width portion. The line width of the thin line image is "1" in the portion of the thin defect portion a having a line width smaller than the lower limit value of "1". In one thick line defect detecting thin line image 12, the line width of the thin line image 12 is “1” in the normal pattern line width portion, but the line width is larger than the upper limit of the line width reference tolerance. In the defective portion b, the line width of the fine line image is "2" or more. The subsequent thin defect detecting section 7 and thick defect detecting section 8 extract the line thin defect and the thick defect using the above-described features.

Next, the thin defect portion extracting section 7 reads the thin line image data o of the thin line image memory 5 for thin defect detection, applies a 3 × 3 line thin portion mask thereto, and extracts the thin defect portion. Will be On the other hand, in parallel with this, the fat defect extracting section 8 reads the thin line image data p of the fat defect image memory 6 and applies a line fat detection 3 × 3 mask to extract the fat defect. Will be

FIGS. 4 and 5 are schematic diagrams showing the processing operation of the thin defect extracting section 7 and the thick defect extracting section.

FIG. 4A shows a thinned line detecting 3 × 3 mask 13 applied to the thinned line image data 11 for thinning defect detection.
4A is an explanatory diagram of a 3 × 3 mask 14 for detecting a line thickening applied to the thin line image data 12 for detecting a thick line defect. FIG. 4B is a diagram illustrating a line width “1” extracted by the 3 × 3 mask 13 for detecting a line thinning portion. "
FIG. 5B is an explanatory diagram of a thick defect portion having a line width of “2” or more extracted by the line thick portion detection 3 × 3 mask 14.

The thin defect detecting section 7 detects the thin line detecting portion 3 × under the following conditions for all the pattern portions “1” of the thin line image 11 of the thin line image data o for detecting the thin line defect detected.
The thinned portion 15 having a line width of “1” is extracted by applying the three masks 13 and sent to the result output unit 9 as thinned portion data q.

Condition of Line Width "1" When two or more pairs of "0" are both facing pixel pairs in a 3 × 3 mask On the other hand, the fat defect detecting part 8 detects the fat defect read in. All the pattern portions “1” of the thin line image 12 of the thin line image data p are subjected to a line thick portion detection 3 × 3 mask 14 under the following conditions to extract a line thick portion 16 having a line width “2” or more. Is sent to the result output unit 9 as the thickened portion data r.

Condition where the line width is equal to or more than "2". In the 3 × 3 mask, when both pairs of pixels facing each other have no more than one set of "0". Finally, in the result output unit 9, the thin defect extraction unit 7 and the line-thickness part data 8 obtained from each of the thick-defect-part extraction units 8 are read out, and the center position of each of the nearby defect parts is calculated and output as defect position data s. I do.

FIG. 6 is a block diagram showing the configuration and processing flow of another embodiment of the present invention. Hereinafter, the operation of this embodiment will be described along the flow of FIG.

The video signal D obtained by scanning the wiring pattern to be inspected by the photoelectric conversion scanner 21 is converted into a binary signal by the binarizing circuit 22 so that the pattern portion is "1" and the other portions are "0". To obtain binary image data E. This binary image data E is used in the next thinning circuit 23 and the thinned portion extracting circuit 26 and the thickened portion extracting circuit 27 later.

Next, the thinning circuit 23 performs thinning on the binary image data E obtained by the binarizing circuit 22 to obtain thin line image data F in which the line width of the pattern portion is "1" in terms of the number of pixels. . The thin line image data F is sent to the next line width expansion circuit 24 and line width expansion circuit 25. FIG. 7 and FIG.
FIG. 3 is a schematic diagram showing the operation of the value conversion circuit 22 and the thinning circuit 23.

FIGS. 7 (a) and 8 (a) show the basic 2
5 shows a binary image 29 based on value image data E, in which a thin line portion R and a thick line portion S exist, respectively. FIGS. 7B and 8B show a thin line image 30 based on thin line image data F obtained by performing thin line processing on the binary image 29 by the thin line circuit 23. The thin line image 30 is the center line corresponding to the skeleton of the pattern, corresponding to the case where the thin line portion R and the thick line portion S exist, respectively.

Next, the line width enlarging circuit 24 and the line width enlarging circuit 25 use the thin line image data F
, And for each thin line image data F, a line width pattern (thin line width pattern) data G of a lower limit as a reference for line width thinning determination, and a line width pattern of an upper limit of line width thickening determination on the other hand (Thick line width pattern) Data H is generated. For each line width pattern, the pattern portion is “1”,
Others are binary images having a value of “0”.

FIGS. 9 and 10 are schematic views showing the defect extraction processing operation after the thinning processing in this embodiment.

FIG. 9A shows a thin line image 30 shown in FIG.
FIG. 10A shows a thick line width pattern 32 obtained by enlarging the thin line image 30 shown in FIG. 8B (hatched portion). What is characteristic here is that each line width pattern 31 enlarged from the normal pattern is 2
The value image 29 is completely included, and the thick line width pattern 32 completely contains the binary image 29 inside. However, if there is a thin defect portion R, the thin line width pattern 31 is formed at that portion. It protrudes from the binary image 29. Conversely, fat defect S
If there is, the binary image 29 will protrude from the thick line width pattern 32 at that portion. In the following thinned portion extraction circuit 26 and thickened portion extraction circuit 27, a thin line defect portion and a thick line defect portion are respectively extracted using the above-described features.

Next, in the thin defect portion extraction circuit 26, the thin line width pattern data G and 2
The binary image data E obtained by the binarizing circuit 2 is read to extract a thin defect portion. At the same time, on the other hand, the fat defect extracting circuit 27 reads the fat line width pattern data H generated by the line width enlarging circuit 25 and the binary image data E obtained by the binarizing circuit 22, and extracts the fat defect.

The thin defect portion extraction circuit 26 reads the thin line width pattern data G and the binary image data E that have been read, and extracts a thin defect portion from a portion satisfying the following condition.

(Binary image data E = 0) AND (thin line width pattern data G = 1) In the fat defect extracting circuit 27, the following conditions are satisfied.

(Binary image data E = 1) AND (thick line width pattern data H = 0) FIG. 9B shows a thin defect portion T extracted by the thin defect portion extracting circuit 26 in FIG. b) is a fat defect extracting unit 27
5 shows a fat defect U extracted by.

Finally, the result output unit 28 reads the thin line portion data O and the thick line portion data P obtained from each of the thin portion extraction circuit 26 and the thick portion extraction circuit 27, and for each defective portion in the vicinity. In summary, the center position is calculated, and this is output as defect position data Q.

[0034]

As described above, according to the pattern inspection apparatus of the present invention, a thinning process based on a line width reference tolerance and a simple 3 × 3 mask are applied to the image data resulting from the thinning process to obtain a thinning defect portion. By detecting a line-thickness defect portion or by detecting a defect by a combination of a thinning process and an enlargement circuit, a complicated coding process using a length measurement sensor like a conventional pattern inspection device, Defect detection can be performed accurately without the hassle of creating a coded dictionary in advance.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the configuration and processing flow of an embodiment of the present invention.

2A is a diagram showing a binarized image 10 in a case where there is a thin line portion a by the binarization circuit 2 in FIG. 1, and FIG. 2B is a diagram showing FIG.
Thin line image 1 subjected to thinning processing by the thinning circuit 3 in the middle
FIG.

3A is a diagram showing a binarized image 10 in a case where there is a thick line portion b by the binarization circuit 2 in FIG. 1, and FIG. 3B is a diagram showing FIG.
Thin line image 1 subjected to thinning processing by the thinning circuit 3 in the middle
FIG.

4A is a diagram showing a thinned line detection 3 × 3 mask 13 used by the thinned defect portion extraction unit 7 in FIG. 1, and FIG. 4B is extracted by the thinned line detection 3 × 3 mask 13; It is a figure which shows the thin defect part of line width "1".

5A is a diagram showing a line-thickness detection 3 × 3 mask 14 used by the line-thickness defect portion extraction unit 8 in FIG. 1, and FIG. 5B is a line width extracted by the line-thickness detection mask 14. It is a figure which shows the fat defect of 2 "or more.

FIG. 6 is a block diagram showing the configuration and processing flow of another embodiment of the present invention.

7A is a diagram showing a binarized image 29 in a case where there is a thinned portion R by the binarizing circuit 22 in FIG. 6, and FIG. 7B is a thinning process by the thinning circuit 23 in FIG. FIG. 5 is a diagram showing a thin line image 30 on which the following has been performed.

8A is a diagram showing a binarized image 29 in the case where there is a thick portion S by the binarizing circuit 22 in FIG. 6, and FIG. 8B is a thinning process by the thinning circuit 23 in FIG. FIG. 5 is a diagram showing a thin line image 30 on which the following has been performed.

9A is a diagram for explaining the operation of the line width enlarging circuit 24 in FIG. 6, and FIG. 9B is a diagram for explaining the operation of the thinned portion extraction circuit 26.

10A is a diagram for explaining the operation of the line width enlarging circuit 25 in FIG. 6, and FIG. 10B is a diagram for explaining the operation of the thick portion extracting circuit 27.

FIG. 11 is a diagram showing a length measurement sensor of a conventional pattern inspection apparatus.

12A and 12B are schematic diagrams of an example of coding determination by the length measurement sensor shown in FIG. 11, in which FIG. 12A shows a normal pattern, FIG. 12B shows a thin line defect, and FIG. 12C shows a thick defect.

[Explanation of symbols]

 Reference Signs List 1, 21 photoelectric conversion scanner 2, 22 binarization circuit 3, 23 thinning circuit 4 thin line processing control unit 5 thin line image memory for thinning defect detection 6 thin line image memory for thickening defect detection 7 thinning defect unit extraction unit 8 thickening defect unit Extraction unit 9 Result output unit 10, 29 Binary image 11, 30 Thin line image for thin defect detection 12 Thin line image for thick defect detection 13 Line thin part detection 3 × 3 mask 14 Line thick part detection 3 × 3 mask 15 Line width “ 1 "line thinning part 16 line width" 2 "or more line thickening part 24, 25 line width enlargement circuit 26 thin part extracting circuit 27 thick part extracting circuit 28 result output part 31 thin line width pattern 32 thick line width pattern 101 inspection Center 102 Length measurement pixel 103 Vertical (90 °) direction length measurement pixel pair 104 Left / right (0 °) direction length measurement pixel pair 105 Oblique (45 °) direction length measurement pixel pair d, D Video signal e, E 2 Value image data f1 Setting number of steps of thin line image for thin defect detection f2 Setting data of number of steps of thin line image for thick defect detection g Thin line image data for thin defect detection h Thin line image data for thick defect detection o Thin line image memory data for thin defect detection p Thin line image memory data for thick defect detection q, O line thin portion data r, P line thick portion data s Defect position data F Fine line image data G Thin line width pattern data H Thick line width pattern data Q Defect position data

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ identification code FI H05K 3/00 G06F 15/62 405A (58) Investigated field (Int.Cl. ⁷ , DB name) G01N 21/88 G01B 11 / 24 G01B 11/30 G06T 7/00 H05K 3/00

Claims (2)

(57) [Claims] A photoelectric conversion scanner for scanning a target wiring pattern and outputting a video signal corresponding to the reflected light;
A binarization circuit for binarizing a video signal from the photoelectric conversion scanner into a binary image signal, and a thin line for sequentially performing thin line processing on the binary image data obtained from the binarization circuit to generate thin line image data A thin line processing control unit that controls the number of thin line processing stages of the thin line circuit based on a line width reference tolerance and generates thin line image data for thin line defect detection and thin line image data for line thick defect detection, A thin defect image memory controlled by the thin line processing control unit and configured to store thin defect detection image data generated by the thinning circuit; and a thickening controlled by the thin line processing control unit and generated by the thinning circuit. A thin line image memory for thick defect detection for storing image data for defect detection, and image data of the thin line image memory for thin defect detection are read, and a line thin portion detection 3 × 3 mask And a thin defect detecting section for extracting a portion having a line width of 1 pixel or more as a thin defect, and reading image data from the thin line image memory for detecting a thick defect. A thick defect detecting section for extracting a portion having a line width of 2 pixels or more detected by applying a × 3 mask as a thick defect portion. 2. A photoelectric conversion scanner for scanning a target wiring pattern and outputting a video signal corresponding to the reflected light,
A binarization circuit for binarizing a video signal from the photoelectric conversion scanner into a binary image signal; and sequentially performing fine line processing on the binarized image data obtained from the binarization circuit, and having a line width of one pixel. A thinning circuit for generating thin line image data, and an enlargement process is performed on the thin line image data generated by the thinning circuit to regenerate a thin line width pattern that is a lower limit line width pattern that is a reference for line width thinning determination. A first line width enlarging circuit to be constituted and an enlarging process are performed on the thin line image data generated by the thinning circuit to reconstruct a thick line width pattern which is a line width pattern of an upper limit value of a line width thickening judgment. A second line width enlarging circuit, a thin portion extracting circuit for extracting a portion where the thin line width pattern protrudes from the binary image obtained by the binarizing circuit as a line thin portion, and a thin portion extracting circuit. The obtained binary image is the fat Pattern inspection apparatus which comprises a fat portion extracting circuit for extracting a portion protruding from the width pattern as lines thickened portion.