JP2001317919A - Method and apparatus for measuring circuit pattern line width - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and apparatus for measuring a circuit pattern line width, which enable reduction in measuring errors of line width and shorten ing of the processing time, without requiring a large capacity memory. SOLUTION: The apparatus for measuring a circuit pattern line width has an image acquisition circuit 21 for taking in an image from an input device, a binary coding circuit 22, a fine line processing circuit 23, a segment extraction circuit 26, a measuring straight line generation circuit 27 for calculating a straight line passing through a coordinate value subjected to the measurement of line width, an edge point searching circuit 28 for searching contour edge points, a local subpixel processing circuit 29 which performs a subpixel processing, by division into N×N which is confined to the contour edge points and the vicinity thereof and further, a smoothing processing of subpixel image data for the removal of noises and a line width measuring circuit 30, which determines a position on a straight line and indicating a gray value exceeding a binary coded threshold from output data of the local subpixel processing circuit to calculate the line width from the position.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a circuit pattern line width measuring method and a circuit pattern line width measuring device, and more particularly to a pattern line width measuring method and a circuit pattern line which can be used for inspecting the appearance of a wiring pattern such as a printed circuit board. The present invention relates to a width measuring device.

[0002]

2. Description of the Related Art When inspecting the appearance of a wiring pattern on a printed circuit board or the like, an image of the wiring pattern is fetched by an input device such as a camera, and the image of the wiring pattern is subjected to a thinning process until the width becomes one pixel. A center line in which one pixel at the center is arranged is obtained, and image processing for obtaining a width for the center pixel is performed with one center pixel as a measurement point.

FIG. 11 shows the outline. One of the arranged central pixels (central points by thinning) 10 is set as a measurement point 11, and a contour line 12 for line width measurement is searched from this measurement point, thereby obtaining a line width b. However, in this case, since the line width is measured only by information on a pixel basis, in the line width measurement of the oblique line as shown in FIG. 10, the line width a and the center point (measurement point) 11 detected by thinning are determined. The error between the line width b measured by searching for the contour 12 based on the reference becomes large.

[0004] This conventional technique in pixel units will be described with reference to FIG. 11 using threshold values. Since the measurement is performed on a pixel-by-pixel basis, in a situation where the gray value of a pixel on the contour is almost the same as the fixed binary threshold set for extracting the contour, the threshold is slightly larger than the threshold. A line width (measurement line width A1) determined in the case and a line width (measurement line width A2) determined when the line width is slightly smaller causes a measurement error ΔA of one pixel width.

On the other hand, if the resolution of the input device is reduced to reduce the measurement error, a large-capacity image memory is required.

Further, since the image size becomes large and enormous information must be processed, there is a problem that the processing time is enormous and the inspection cannot be performed in a practical time.

[0007]

As described above, in the prior art, the line width measurement error increases, and if it is attempted to reduce it, a large-capacity memory is required, and the processing time is prolonged. Having.

Accordingly, an object of the present invention is to provide an effective circuit pattern line width measuring method and an effective circuit pattern line width measuring device which solve the above problems.

[0009]

A feature of the present invention is that a first step of capturing image data of a circuit pattern line, a second step of binarizing the image data, and a second step of thinning the image data. Step 3, and calculating a straight line having a slope orthogonal to the slope of the center line obtained in the third step and passing through the measurement point on the center line, and using the straight line, both end points of the line width A fourth step of obtaining a first start point and a first end point from the data obtained in the second step by comparing the first start point and the first end point with a predetermined edge determination threshold value; and determining the first start point and the first end point. A fifth step of restrictively converting only the vicinity area including the sub-pixel into a sub-pixel, and comparing a second start point and a second end point on the straight line by the sub-pixel with the predetermined edge determination threshold value The sixth step to find And flop, in said second starting point and the circuit pattern line width measurement method of calculating the line width of the circuit pattern lines from the second end point.

Here, it is preferable to perform a smoothing process only on the sub-pixel region. In addition, it is preferable that the region to be subpixel is a region of pixels forming the first start point and the first end point and pixels adjacent to these pixels from each direction. When a wide area outside the measurement target is included, it is preferable to perform a process of removing the wide area after the third step. In this case, the removing process is performed by the third process.
The image data of the wide area is obtained by performing the one-pixel width reduction processing on the image data obtained in the step (i). Next, the image data of the wide area only is subtracted from the image data obtained in the third step. Processing can be.

Another feature of the present invention is that an image capturing circuit for capturing an image from an input device, a binarizing circuit for binarizing image data acquired by the image capturing circuit with a threshold value,
A thinning circuit for performing thinning processing on the binary image data by the binarization circuit with a specified number of lines, and coordinate values in units of line segments as a set of points where the image data by the thinning circuit are continuously located And a line segment extraction circuit that calculates the inclination of the center line passing through the coordinate value for which the line width is measured, calculates the inclination perpendicular to the inclination, has the orthogonal inclination, and performs the line width measurement. A measurement straight line generation circuit for calculating a straight line passing through the coordinate values; and an edge point search for searching for a contour edge point along the straight line, which is an end point of the line width measured from the coordinate data of the binary image by the binarization circuit. A local subpixel circuit for subpixel-forming only the contour edge point and the area in the vicinity thereof into N × N divisions per pixel, and smoothing the subpixel image data to remove noise; The local sub A line width measurement which calculates a line width at a coordinate value for performing the line width measurement based on a position which is on the straight line and has a gray value equal to or higher than the binarization threshold value from the output data of the cell conversion circuit. And a circuit pattern line width measuring device having a circuit.

Here, a reduction / enlargement circuit for performing one-pixel reduction / enlargement processing on the processing result data of the thinning circuit, and an AND operation of the processing result data of the reduction / enlargement circuit and the processing result data of the thinning circuit And an area difference circuit for using the result in the line segment extraction circuit.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a circuit pattern line width measuring apparatus according to an embodiment.

Referring to FIG. 1, an image capturing circuit 21 for capturing an image from an input device such as a camera or a scanner, a first image memory 31 for storing image data acquired by the image capturing circuit 21, Binarization circuit 2 for binarizing the image data of the image memory 31 with the set threshold value Th
2, a second image memory 32 for storing binary image data output from the binarization circuit 22, and a second image memory 32
And a third image memory 33 for storing the processing result data of the thinning circuit 23.

Further, a reduction / enlargement circuit 24 for performing one-pixel reduction / enlargement processing on the image data of the third image memory 33, a fourth image memory 34 for storing the processing result data of the reduction / enlargement circuit 24, and a third An AND operation is performed between the image data in the image memory 33 and the image data in the fourth image memory 34, and the result of the operation is stored in the fifth image memory 35. A line segment extraction circuit 26 outputs a coordinate value in units of line segments as a set of points where image data of a certain fifth image memory 35 is continuously located, and a line segment data memory 37 for storing the result.

Further, with respect to the line segment data output from the line segment data memory 37 which is the result of the line segment extraction circuit 26 which is a data search circuit, coordinate values Pi (0,...) For performing line width measurement according to FIG. , N) to calculate the gradient G of a straight line passing through two coordinate values at positions spaced at equal intervals in the opposite direction from each other,
_A measurement straight line generation circuit 27 that calculates a slope G _A orthogonal to the slope G and calculates a straight line L of the slope GA passing through Pi, and from the coordinate data of the binary image in the image memory 2 along the straight line L described above. Contour edge point Ps which is the end point of the line width to be measured
(Start point), edge point search circuit 2 for searching Pe (end point)
8 and only the contour edge points Ps and Pe and the area in the vicinity thereof are stored in the sixth image memory 36.
6 is divided into N × N sub-pixels for each pixel, and a local sub-pixel conversion circuit 39 for smoothing the sub-pixel image data to remove noise.
From the output data of the local subpixel conversion circuit 29 on the straight line L and the above-described binarization threshold Th
The positions Ps 'and Pe' which are the above gray values are measured, and | P
a line width measuring circuit 30 that calculates s′−Pe ′ | as a line width at Pi.

That is, it has a function of resampling the vicinity of the contour on a computer, performing a smoothing process in a sub-pixel state, and measuring the line width using the result.

Next, a circuit pattern line width measuring method according to the embodiment will be described.

Image 1 shown in FIG. 2 is a diagram showing a binarized image. Areas 1 to 4 where line width measurement is performed (area 1 is an area having a different line width, areas 2 to 4 are uniform line widths) An image whose area (w is w) can be extracted by the binarization processing is acquired by the image acquisition circuit 21, and the image acquired by the image acquisition device 21 is stored in the first image memory 31.

Then, the first image memory 31 is used as an input to the binarization circuit 22, and the image data in the first image memory 31 is binarized with the set threshold value Th to measure the line width. The region is extracted and stored in the second image memory 32.

Then, thinning processing is performed on the regions 1 to 4 in FIG. 2 by the thinning circuit 23 with the number of stages c for thinning the line width w, and the result is stored in the image memory 33. here,
The number c of steps is set to a value equal to the line width of the areas 2 to 4 in FIG.

The output of the thinning circuit 23 is an image 2 shown in FIG. 3, that is, a diagram showing the thinning result of the designated line width. Since the areas 2 to 4 have a uniform line width, the entire area is thinned. As a result, an area (area a) having a width of one pixel is obtained. On the other hand, the result of the region 1 includes a region (region b) having a width of one pixel or more.

Next, in order to extract only the target area of the line width measurement, that is, the area of one pixel width, the third image memory 33 is used.
Is input, and a process of deleting the area b from the data of the image 2 is performed by the reduction / enlargement circuit 24.

First, a one-pixel width area (area a) is deleted from the image data (image 2 in FIG. 3) in the third image memory 33 by one-pixel width reduction processing. Next, the result data is subjected to one-pixel width expansion processing.

The result is only the area b as shown in FIG. 4 showing the result of the one-pixel reduction / enlargement of the image 3, and this data is stored in the fourth image memory 34.

Next, an image 2 shown in FIG.
By subtracting the image data of the image 3 of FIG. 4 from the image data of FIG. 4, only the one-pixel width area (area a) is extracted, and the result is stored in the fifth image memory 35. The content of the fourth image memory 34 is a diagram of the area difference result which is the image 4 in FIG.

Next, the line segment extraction circuit 26 searches the fifth image memory 35 for image data, and stores continuous coordinates in the line segment coordinate data memory 37 in units of line segments.

Referring to FIG. 6, line segment coordinate data memory 3
The coordinate data of 7 is the center point Pi (Xi, Yi) of the pixel 11 of the measurement point for which the line width is measured.

Next, the inclination G of the center line passing through the center point Pi is calculated by the measurement straight line generation circuit 27 using two equally spaced points in the opposite direction from Pi on the thinned data.

Next, there is a gradient G _A orthogonal to the gradient G,
And the formula of a straight line L passing through the center point Pi (Xi, Yi), Y−
Yi = G _A (X−Xi) is calculated.

Next, the end point and the starting point Ps (X
s, Ys) and the end point Pe (Xe, Ye) are compared with a predetermined edge determination threshold Th and detected from the image data in the second image memory 32.

Next, only the Ps and Pe of the image memory 1 and the vicinity thereof are subpixelized into N × N divisions (for example, N = 10) by the local subpixel conversion circuit 29, and the data is stored in the image memory 6 ( Local sub-pixel image n). FIG. 7 shows a data structure resampled to sub-pixels, and shows a state where a pixel located at the upper left in the figure among pixels arranged in 3 rows and 3 columns is re-sampled to sub-pixels. Data 100 is 100-
11, 100-12 ... 100-1N ... 10
It is subdivided into 0-NN.

In addition, Ps and Pe, which are resampled to the sub-pixels in a limited manner, and the vicinity thereof are minimized in consideration of the high-speed measurement and the cost reduction of the apparatus.
It is preferable that the pixel region is a region of s, Pe, and one pixel adjacent to the pixel in each direction. That is, it is preferable that only pixels constituting Ps and Pe and pixels surrounding the pixel in a single layer are resampled to sub-pixels.

However, depending on the measurement speed and the circumstances of the apparatus, the pixel constituting Ps and Pe and the pixel region surrounding the pixel in a single or double manner, or the pixel constituting Ps or Pe and the pixel constituting the Ps or Pe in one time It is also possible to convert the area of the pixel that surrounds double, triple, into sub-pixels.

Next, noise is removed by performing a smoothing process (for example, a mean filter using an average value of the M × M area) only on the straight line L of the local subpixel image n.
The error of the line width measurement by one line of the sub-pixel grayscale image is reduced. Then, the result is stored in the sixth image memory 36.

Next, the Ps and Pe converted into sub-pixels from the sixth image memory 36 by the line width measurement circuit 30 are set to Ps having a grayscale value not less than the edge determination threshold Th.
s ′ and Pe ′ are defined, and the distance between Ps ′ and Pe ′ is calculated as the line width.

The processing from the measurement line generation circuit 27 to the line width measurement circuit 30 is performed for the data amount of the line data memory 37.

FIG. 8 shows a search for a contour point along a search path in a 4.times.5 pixel matrix, thereby recognizing a pixel of a contour point A in pixel units, and sub-pixelating and smoothing the pixel and its neighborhood. A later contour point B is shown. As described in the prior art of FIG. 11, the error ΔA
Becomes However, the error when measured in sub-pixel units smaller than one pixel is as small as ΔB. That is, when a pixel is divided into N sub-pixels, ΔB is reduced to 1 / N of ΔA.

Referring to FIG. 9, in the measurement in pixel units, the line width determined when the width is slightly larger than the threshold value (measurement line width A1) and the line width determined when the width is slightly smaller (measurement line width). A large measurement error ΔA of one pixel width occurs in the width A2). However, in the measurement by subpixel conversion according to the present invention, the sub-width is determined when the line width is slightly larger than the threshold value (measurement line width B1) and the line width is determined when the width is slightly smaller (measurement line width B2). The measurement error ΔB having a small pixel unit width can be reduced.

Next, another embodiment of the present invention will be described. In the above embodiment, the straight line L of the route for searching the positions of the start point Ps and the end point Pe of the line width measurement is calculated. The method of calculating the gradient G in the process uses two points on the thinned data.

On the other hand, in the present embodiment, the number of points to be used is increased and the slope is calculated by a statistical method such as the least squares method. Replace with the method of calculating the slope.

In the above embodiment, the sub-pixel conversion of the image data is performed by resampling the limited area and smoothing it. In this embodiment, the image data on the search path is linearly interpolated. Or replace with curve interpolation method.

[0043]

As described above, the effect of the present invention is that high-speed and high-accuracy measurement can be performed by first limiting an area requiring high-accuracy processing.

The image memory required by minimizing the area to be subpixelized is smaller than when the entire image is subpixelized or when high-resolution image acquisition is performed. Therefore, the manufacturing cost is low.

Furthermore, since the vicinity of the contour can be resampled on a computer and smoothed in a sub-pixel state, the result can be used to reduce the effect of noise that may result in a quantization error. Can be eliminated, and high accuracy can be realized.

In the smoothing process, the processing time can be reduced by limiting the processing target to a sub-pixel area, and the speeding up is realized.

Further, if there is an area R (a line width to be measured: an area having a width larger than w) which is not to be subjected to the line width measurement in the image data to be measured, the center at which the line width is measured is determined. By terminating the thinning process for extracting a line at the stage where the designated width w is thinned, the region R becomes 1
Only the center line pixel for which the line width is to be measured can be automatically extracted by extracting only the region R by the one-pixel enlargement / reduction processing and deleting it from the result of the thinning processing. Therefore, it is possible to increase the speed.

The direction in which the line width is measured is calculated, for example, from the inclination using two points of the same center line which are equidistant from the center point to be measured. Since the calculation is performed on a straight line obtained from the orthogonal inclination and the center point, the speed can be increased.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a circuit pattern line width measuring apparatus according to an embodiment.

FIG. 2 is a diagram showing image data of a binarized image according to the embodiment.

FIG. 3 is a diagram showing image data as a result of thinning a designated line width in the embodiment.

FIG. 4 is a diagram showing image data as a result of one-pixel reduction and enlargement in the embodiment.

FIG. 5 is a diagram illustrating image data of a region difference result in the embodiment.

FIG. 6 is a diagram showing a line width measurement at a measurement position in the embodiment.

FIG. 7 is a diagram showing a data structure when pixels are resampled into sub-pixels in the embodiment.

FIG. 8 is a diagram showing a comparison of the appearance of contour points before and after sub-pixel conversion in the embodiment.

FIG. 9 is a diagram showing a comparison of a line width measurement state before and after sub-pixel conversion in the embodiment.

FIG. 10 is a diagram illustrating a line width measurement according to a conventional technique at a measurement position.

FIG. 11 is a diagram showing a state of line width measurement according to a conventional technique.

[Explanation of symbols]

 Reference Signs List 10 center pixel (center point by thinning) 11 pixel to be measured point 12 contour 21 image capture circuit 22 binarization circuit 23 thinning circuit 24 reduction / enlargement circuit 25 area difference circuit 26 line segment extraction circuit 27 measurement line generation Circuit 28 Edge point search circuit 29 Local subpixel conversion circuit 30 Line width measurement circuit 31 First image memory 32 Second image memory 33 Third image memory 34 Fourth image memory 35 Fifth image memory 36 Sixth Image memory of 37 line segment data memory

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 2F065 AA22 BB02 CC01 DD03 DD06 FF01 JJ03 JJ26 QQ04 QQ13 QQ24 QQ25 QQ32 QQ34 QQ36 UU05

Claims (7)

[Claims] A first step of capturing image data of a circuit pattern line, a second step of binarizing the image data, and a third step of thinning the image data.
A straight line having a slope orthogonal to the slope of the center line obtained in the third step and passing through a measurement point on the center line is calculated, and the straight line is used as a first end point of the line width. A fourth step of obtaining a start point and a first end point from the data obtained in the second step by comparing the data with a predetermined edge determination threshold value, and an area in the vicinity including the first start point and the first end point Fifth where only pixels are limited to sub-pixels
And a sixth step of comparing the edge determination threshold value with the edge determination threshold value to obtain a second start point and a second end point on the straight line and by the sub-pixel, and the second start point and the second A circuit pattern line width measuring method, wherein a line width of the circuit pattern line is calculated from an end point. 2. The circuit pattern line width measuring method according to claim 1, wherein a smoothing process is performed only on the sub-pixel region. 3. The sub-pixel-forming region is a region of pixels constituting the first start point and the first end point and pixels adjacent to these pixels from each direction. 2. The circuit pattern line width measurement method according to 1. 4. The circuit pattern according to claim 1, wherein when a wide area outside the measurement target is included, a process of removing the wide area is performed after the third step. Line width measurement method. 5. The process of removing the image data according to the third step by performing a one-pixel width reduction process to obtain image data of only the wide area, and then performing the third step. 5. The circuit pattern line width measuring method according to claim 4, wherein the process is a process of subtracting the image data of only the wide area from the image data. 6. An image capturing circuit that captures an image from an input device, a binarizing circuit that binarizes image data obtained by the image capturing circuit with a threshold value, A thinning circuit for performing thinning processing on the value image data with the number of stages of the designated line width, and a line segment extracting circuit for outputting coordinate values in line segments as a set of points where the image data is continuously located by the thinning circuit And calculating the inclination of the center line passing through the coordinate values for performing the line width measurement, calculating the inclination orthogonal to the inclination, calculating the straight line having the orthogonal inclination and passing the coordinate values for performing the line width measurement A measurement straight line generation circuit, an edge point search circuit for searching for a contour edge point along the straight line, which is an end point of the line width measured from the coordinate data of the binary image by the binarization circuit, and the contour edge point And only the area in the vicinity thereof are N × A local subpixel conversion circuit for converting the subpixel into subpixels, smoothing the subpixel image data to remove noise, and output data of the local subpixel conversion circuit on the straight line and the binary Position where the gray value is the same as the
A circuit width measurement circuit for calculating a line width from the position as a line width at a coordinate value for performing the line width measurement. 7. A reduction / enlargement circuit that performs one-pixel reduction / enlargement processing on the processing result data of the thinning circuit, and performs an AND operation on the processing result data of the reduction / enlargement circuit and the processing result data of the thinning circuit. 7. The circuit pattern line width measuring apparatus according to claim 6, further comprising: an area difference circuit that uses a result of the operation as the line segment extracting circuit.