JPH04147045A - Surface inspection device - Google Patents

Abstract

PURPOSE:To effect a window processing on a screen at high speed by judging that there is a defect if the number of digital signals exceeding a predetermined value in the window or an area of a mass is large when a memory is scanned by the window. CONSTITUTION:An image signal where a contract of brightness and darkness exceeds a predetermined value is extracted from the image signal of a video camera 10 by an image signal extractor 13, is stored into a memory 16 through an A/D converter, and at the same time a histogram of entire digital signal of the image is created 18. With this histogram signal, a waveform signal exceeding a threshold which is set by a threshold setter 20A is selected by a signal extractor 20 through an LPF 19 and a central line for vertical and horizontal directions for the image is obtained by a position setter 21. Also, a small window which is small enough for enabling a scratch to be screened is set 22 for the screen of the image. Then, the screen is scanned vertically by the window in a window address setter 23 and it is determined that there is defect 24 if the number of digital signals exceeds a predetermined value or an area of a mass is large.

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to an object surface inspection device.

[Conventional technology]

In recent years, in the industrial field, automatic inspection equipment that uses video cameras and electronic processing machines has increasingly been installed to replace human visual inspection of parts. The objects to be inspected include raw materials, raw materials, semi-finished products, finished products, etc., and defects on the surface,
For example, various surface inspection devices for detecting differences in external shape, size, etc., flaws, etc. have been proposed and put into practical use.

Conventional surface inspection equipment for objects irradiates the object to be inspected with illumination light, captures the transmitted light or reflected light of the object with a video camera, converts it into an image signal, and converts the signal into an image signal using a computer, etc. Defects are detected through processing analysis using an electronic processing machine.

In this case, the analog image signal captured by the video camera is generally converted into a digital signal using an A/D converter or the like to facilitate electronic processing. Furthermore, processing all the data regarding the degree of brightness and darkness of an image, that is, the gray scale, would be impractical as it would involve processing a huge amount of data, making the equipment complex and requiring a large amount of processing time. .

Therefore, most of the inspection devices currently in use have been put to practical use by dividing the brightness of the image into two values by thresholding and simplifying the processing.

Alternatively, there are some devices that digitize and process a limited number of grayscale levels.

For example, when attempting to detect flaws on the surface of an object to be inspected, various problems are encountered. The objects to be inspected include metal plates, metal strips, glass, and plastic.

There are various materials such as plastic containers, paper, and plastic rolls, but it is difficult to extract and detect only scratches from the surface. When photographing an object to be inspected with a television camera, it not only captures scratches as changes in brightness, but also detects slight wrinkles, dirt, etc. that may not be defects or the rough surface of the object being inspected. One of the reasons is that it is perceived as a change in brightness and darkness at the same time, making it difficult to distinguish it from a scratch.

Roughness, wrinkles, dirt, etc. on the surface of these surfaces can be said to be "noise" on the screen for a video camera, and from such noise, it is easy to identify only the defects, such as scratches. Extracting and detecting them has been difficult and a challenge.

Various methods can be used to solve the problems mentioned above. One effective method is to divide the screen into many small areas, measure the changes in brightness within each small area, and if the change exceeds a predetermined value, it is possible to detect defects there. There is a way to determine if there is.

The number of noise-like brightness changes is small in each small area, and there are many areas with brightness changes in areas with scratches, so it is easier to distinguish between scratches and noise. Of course, the size of the area to be divided may be determined as appropriate depending on the size of the flaw to be determined.

[Problem to be solved by the invention]

However, although this method of inspecting defects by dividing the screen into many small areas is very effective, it requires repeating the discrimination process while scanning small areas across the entire screen. It takes a lot of processing time to make a determination.

[Means and effects for solving the problem] According to the present invention, an illumination device emits illumination light onto an object to be inspected, the transmitted light or reflected light is captured by a video camera, and the video camera outputs an image. In an inspection device that processes signals using an electronic processor to detect defects in an object to be inspected, an image signal whose contrast between light and dark exceeds a predetermined value is selected from image signals from a video camera (10). The image signal is extracted by an image signal extraction means (13), the image signal is converted into a digital signal by an A/D converter (14), and the digital signal is stored in a memory (16). A histogram of all digital signals is created by a histogram creation means (18), the histogram signal is supplied to a low-pass filter (19) to remove its high frequency components to form a waveform signal, and a predetermined waveform signal is formed. A signal extraction means (20
), and for the selected waveform signal, the center line in the vertical or horizontal direction of the image is determined by the vertical or horizontal position setting means (21), and the center line is determined in advance to the extent that scratches can be detected on the screen of the image. setting a small window having a specified size by a window setting means (22);
Adjust the screen vertically using the above window relative to the memory in which the digital signal is stored, so that the center of the above window coincides with the vertical or horizontal center line of the image obtained from the selected waveform signal. Wind address setting means (
23), and when the memory is scanned by the window, scratches etc. may occur if there are a large number of digital signals exceeding a predetermined value in the window, or if the area of the mass of digital signals is large. Means for determining if there is a defect (24)
A surface inspection device is obtained which makes a determination based on the following.

〔Example〕

The present invention will be described below with reference to the drawings.

The present invention will now be described in detail with reference to FIGS. 1A to 1E.

In the present invention, first, a portion of the screen of a video camera showing an image of the surface of the object to be inspected, where the brightness changes, is converted into a digital signal. To do this, analog image signals of areas with changes in brightness and darkness on the west face may be binarized and converted into digital signals using a predetermined threshold, or analog signals of bright and dark areas may be differentially processed to differentiate between bright and dark areas. It is also possible to convert the signal into a waveform and similarly binarize it into a digital signal.

This digital signal is stored in memory over the entire screen.

Figure 1A shows the above-mentioned digital signals scattered as black dots in the area or screen (1) corresponding to the video camera screen of the memory storing the digital signals. . Roughly scattered black dots (2) on the screen (1)
) is the difference in brightness caused by factors other than defects on the surface of the inspected object, such as roughness or light dirt, which are converted into digital signals, and are so-called noise-like items that interfere with defect determination. It is.

On the other hand, the dense black spots (3) and (3A) in the screen (1) are
This is a digital signal corresponding to a defect such as a scratch that should be detected other than the nose, and many black spots are gathered here.

It is necessary to detect this and distinguish it from other noise.

Figure 1B shows a digital signal on the x-x' line of screen (1), that is, on the IA scanning line.For fear, if this digital signal is detected as it is, it will be difficult to distinguish between scratches and noise. However, as shown in Figure 1A, the screen (1) is divided into small sections, and in this example, 4. It is clear that noise and scratches can be easily distinguished by checking the number of black dots in each subdivision.

In the present invention, first, a histogram of digital signals (black dots) in the horizontal direction of the screen (1) is created over the entire vertical direction of the screen (1). This histogram is displayed on the screen (1
) shows the distribution state of all digital signals appearing in the horizontal direction. FIG. 1C illustrates this histogram.

Next, the histogram created from the digital signal shown in FIG.

The first diagram shows this waveform signal obtained by converting the digital signal.

In order to distinguish defects from noise such as dirt, the waveform signal shown in the first diagram is cut by a predetermined threshold value (TI()), the position of the part of the waveform signal exceeding the threshold value is determined, and the Vertical center line (CN) to the image
seek. In the case of the first diagram, there are two parts of the waveform signal that exceed the threshold (TH), so the center line above is (CNI)
, (CN2). Note that the reason why the histogram created from the digital signal is rectified into a waveform signal using a filter is to make it easier to find the center line of the mass of the digital signal. Furthermore, the number of center lines (CN) determined as described above will of course vary depending on the distribution state of the digital signals in the histogram and how the threshold value is determined.

As shown in FIG. 1E, for example, from the top to the bottom of the screen (1) of the memory that stores digital signals, a predetermined small section window (6) is centered m (CN
) to scan the screen (1), and if there are more than a predetermined number of digital signals in the window (6), or if the area of a block of digital signals is large, an abnormality will be detected. Or it is determined that there is a defect. Wind (6)
is a single object with a constant size.

FIG. 2 shows, in block form, an example of implementing the principles of the invention described above. In the example of FIG. 2, an image signal (a) of an object to be inspected (not shown) captured by a video camera (10) is transmitted to a preamplifier (11)! I will supply this. The preamplifier (11) supplies the amplified image signal (b) to the image signal extractor (13) and the synchronization signal separator (12). The image signal extractor (13) generates a portion of the image signal exceeding a threshold value (THI) predetermined by the threshold value setter (13A) as an extracted signal (C).

The image signal extractor (13) is, for example, a comparator, and extracts a specific image signal (C) based on the above threshold (TI(1)).
Extract. This extracted image signal (C) is supplied to an A/D converter (14) and converted into a digital signal (d).

The converted digital signal from the A/D converter (14) (
d) is supplied to the memory (16) through the fixed contact (X) and the movable contact (^) of the conversion switch (15).

The sync signal separator (I2) is a vertical sync signal separator (12
v) and a horizontal synchronization signal separator (12h), which separates the vertical synchronization signal (v) and horizontal synchronization signal (0-1) from the image signals captured and amplified by the video camera (10), respectively. do. These vertical and horizontal synchronization signals (V) and (
h) to the image address generator (17). An image address generator (17) generates an address signal (e). This address signal (e) is connected to the conversion switch (15A).
is applied to the memory (16) through the fixed contact (X') and the movable contact (^), so that the digital signal (d) is stored in synchronization with the image signal (a) captured by the video camera (10). , controls the vertical and horizontal addresses of the memory (16).

On the other hand, the digital signal (d) output from the digital converter (14) is supplied to a horizontal histogram generator (18), and a horizontal histogram signal (f) of the digital signal (d) is generated.
(see Figure 1C) and apply it to the low-pass filter (19)
At the same time, the vertical synchronization signal (V) and the horizontal synchronization signal (5) are supplied to the horizontal histogram generator (18) and the low-pass filter (1) to obtain the waveform signal (6).
9) respectively to correctly create histogram signal (grain) and waveform signal (barrel).

This waveform signal (2) is fed to a signal extractor (20), which extracts the waveform signal (from the chain to a threshold setter (2OA
), the signal (i) exceeding a predetermined threshold (TH2) is extracted (see first diagram). This signal (i) is supplied to a vertical position setter (21), and the waveform signal (vertical center point of the odor image) of this signal (1) is determined.

Vertical synchronization signal (■
) and a horizontal synchronizing signal (5), the center point is controlled to be set at the correct position on the screen, and a vertical position signal (j) is output.

Reference numeral (22) denotes a window setting device, which generates a window setting signal tentatively for setting a window with a size small enough to include the flaw desired to be detected.

(23) is a window address generator, which receives the vertical position signal (j), the window setting signal (2), and the vertical and horizontal synchronization signals (■) and (5), and performs vertical scanning with a defined window. At this position, a window address signal (1) that vertically scans the digital signal (d) stored in the memory (16) is output to the other fixed contact (Y') of the transfer switch (15A). Switch (15) and (
15A) shall be interlocked, and fixed contacts (X) and (X
') is in the signal acquisition mode of the memory (16), and the other fixed contacts (Y) and (Y') are in the judgment mode.

In this way, the switches (15) and (15A) are set to the signal acquisition mode of the memory (16), the digital signal (d) of all images is stored in the memory (16), and the switch (15) and (15A) are switched to the judgment mode, the window is scanned only at a specific vertical position by the operation shown in FIG. 1E, and the result is judged by the judgment circuit (24). ) is detected.

The above has explained an example in which the operation is performed by obtaining a horizontal histogram, but it is clear that the same operation and effect can be obtained by obtaining a vertical histogram instead of the horizontal histogram and scanning the window horizontally. Dew. In this case, all you need to do is change the horizontal histogram generator (18) to the vertical histogram generator and the vertical position setter (21) to the horizontal position setter. Omit what you do.

Furthermore, FIG. 3 shows a block door according to another embodiment of the present invention. In FIG. 3, the same reference numerals as in FIG. 2 indicate the same parts. The difference between the example shown in FIG. 3 and the example shown in FIG. 2 is that in the example shown in FIG. 3, two sets of histograms, horizontal and vertical, are created simultaneously.

Therefore, in the example of FIG. 3, a vertical histogram generator (18B) and a low-pass filter (19
B), signal extractor (20B), and threshold setter (20C)
).

The example in Figure 3 is different in operation from the example in Figure 2, in that the vertical position setter (21) in the example in Figure 2 controls the horizontal and vertical positions simultaneously. It is a horizontal and vertical position setter (21B) that performs
) is a point that generates a window address signal that addresses the intersection of the horizontal and vertical positions.

Figure 4 illustrates this state. In the figure, the vertical center line (CNI) and (CN2) obtained from the horizontal histogram, and the horizontal center line (CN3) and (CN2) obtained from the vertical histogram. A determination is made by sequentially setting a window (6) at each intersection of CN4).

According to this method, the window is not scanned and the window is simply moved to the above-mentioned intersection points in order to make the determination, so the processing time for comprehensive determination is extremely short.

A description of other parts of the bronc in FIG. 3 will be omitted since the operations are exactly the same as in FIG. 2.

Although the present invention has been explained in detail using a block diagram, in any case, the present invention is generally implemented by computer processing, and the above-described operation of the present invention cannot be performed by software programming. This will be obvious to those skilled in the art.

In addition, the shape of the window is not limited to a rectangle, but may be made into a special shape depending on the purpose of discrimination, and the determination process within the window may be performed by measuring the quantity of extracted signals, or determining the degree of agglomeration or shape of extracted signals. Of course, it is possible to set it freely depending on the purpose of detection, such as making a determination based on the following criteria.

[Effects of the Invention] The present invention eliminates the drawbacks of the conventional method in terms of processing time, makes it possible to perform window processing on the screen at extremely high speed using a unique method, and improves the speed of a high-speed production line. The effect of making it possible to use such a device as described above is extremely significant.

Another major feature is that the device can be easily manufactured without requiring any particularly complicated configuration.

[Brief explanation of drawings]

Figure 1 is a route map for explaining the principle of the present invention, Figures 2 and 3 are block diagrams showing embodiments of the present invention, and Figure 4 is a route diagram for explaining the principle of the present invention.
It is a route map for explaining the operation of the example shown in the figure. In the figure, (10) is a video camera, (12) is a sync signal separator, (13), (20) is a signal extractor, and (14) is a sync signal separator.
) is an A/D converter, (15) and (15A) are conversion switches, (16) is a memory, (17) is an address generator, (18) and (18B) are horizontal and vertical histogram setters, (19 ), (19B) is a low-pass filter,
(20), (,20B) is a signal extractor, (20A)
, (20C) is a threshold value setter, (21B) is a horizontal/vertical position setter, (22) is a window setter, (23) is a window address generator, and (24) is a determination circuit, respectively. Agent Hidemori Matsukuma No. 1 N1 N2 30th 8th motion Iil: f) Explanation 1z I6 Roughly Yoneizumi Diagram Figure 4

Claims (1)

[Claims] 1. Illumination light is irradiated onto the object to be inspected from an illumination device, the transmitted light or reflected light is captured by a video camera, and the image signal outputted by the video camera is image-processed by an electronic processor. ,
In an inspection device for detecting defects in an object to be inspected, an image signal extracting means extracts an image signal whose contrast between light and dark exceeds a predetermined value from among image signals from a video camera, and converts the image signal into A. /D converter to convert it into a digital signal, store the digital signal in a memory, create a histogram of all the digital signals in the horizontal or vertical direction of the image by a histogram creation means, and supply the histogram signal to a low-pass filter. forming a waveform signal by removing high frequency components thereof, selecting a waveform signal exceeding a predetermined threshold value of the waveform signal using a signal extracting means, and using the selected waveform signal in the vertical or horizontal direction of the image. The center line in the direction is determined by the vertical or horizontal position setting means, and the window setting means sets a small window with a predetermined size that is large enough to screen out flaws on the image screen, and the selected waveform signal is means for setting the screen vertically in the window with respect to the memory in which the digital signal is stored, so that the center of the window coincides with the center line in the vertical or horizontal direction of the image obtained from the image. When the above memory is scanned by the above window, defects such as scratches may occur if there are a large number of digital signals exceeding a predetermined value in the window, or if the area of a chunk of digital signals is large. A surface inspection device characterized in that a determination means determines whether the presence exists. 2. Illumination light is irradiated onto the object to be inspected from an illumination device, the transmitted light or reflected light is captured by a video camera, and the image signal outputted by the video camera is image-processed by an electronic processor;
In an inspection device for detecting defects in an object to be inspected, an image signal extracting means extracts an image signal whose contrast between light and dark exceeds a predetermined value from among image signals from a video camera, and converts the image signal into A. A/D converter converts the digital signal into a digital signal, stores the digital signal in a memory, and creates histograms of all the digital signals in the horizontal and vertical directions of the image, respectively, using a histogram creating means, and generates the histogram signals in the horizontal and vertical directions. is supplied to a low-pass filter and its high-frequency components are removed to form horizontal and vertical waveform signals, and a waveform signal exceeding a predetermined threshold of the two waveform signals is selected by a signal extraction means. , for the two selected waveform signals, determine the center lines in the vertical and horizontal directions of the image using the vertical and horizontal position setting means, respectively, and set a predetermined size on the image screen to the extent that scratches can be identified. Set a small window with a certain angle using the window setting means, find the position where the vertical and horizontal center lines of the image obtained from the two selected waveform signals intersect, and set the center of the window at the position where the center lines of the image intersect. The determination means determines that there is a defect such as a scratch when there is a large number of digital signals exceeding a predetermined value in the window, or when the area of a block of digital signals is large, etc. A surface inspection device characterized by what it has done.