JPH0627037A - Flaw inspecting apparatus - Google Patents

Abstract

PURPOSE:To detect only flaw by making a flaw decision after substituting an image value with a limit value when an inputted image value is larger than a threshold, i.e., a limit value, for discriminating a non-edge part from an edge part thereby excluding a sharp edge on the surface of an inspection object from flaw detection. CONSTITUTION:When a cylindrical object 6 having a cut face is inspected, image signals obtained through a camera 1 have extremely high image values at an edge part and image values at a flaw part are only subtly different from those at other parts. Lower limit of threshold for discriminating a non-edge part from an edge part is then determined based on an input image of an object 6 having no flaw and an image value higher than the limit value is substituted with the limit value thus smoothing the edge part. When differentiation processing is performed based on the image data thus smoothed, differentiated value only at a defective part is increased and compared with a reference value thus detecting a flaw.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect inspection apparatus for determining the presence or absence of a defect on a surface portion to be inspected by using an image processing technique.

[0002]

2. Description of the Related Art FIG. 6 is a diagram showing a structure of a defect inspection apparatus described in, for example, Japanese Utility Model Laid-Open No. 3-34155.
In the figure, 1 is a signal source (camera), 2 is an A / D converter for converting the image signal from the signal source 1 into a digital signal, 3 is an image memory for storing the image value converted into the digital signal, and 4 is A data processing device 5 for determining the presence or absence of a defect on the surface of the inspection object based on the image value, 5 is an illuminating device,
Is an inspection target, and 7 is a transfer device that transfers the inspection target 6.

Next, the operation will be described with reference to the flowchart of FIG. First, the surface of the inspection target 6 is illuminated by the illumination device 5, and the surface of the inspection target 6 is scanned by the signal source 1 to input an image signal (S1). The input image signal is converted into a digital signal by the A / D converter 2 (S2), and the image memory 3 stores the image value for each pixel.
(S3). The image value of each pixel stored in the image memory 3 is subjected to a difference calculation (differential process) of image values between adjacent pixels by the data processing device 4 (S).
4). Finally, the difference calculation result is compared with a predetermined reference value (S5), and when the difference calculation result is larger than the reference value, it is determined that the surface portion corresponding to the pixel is defective (S6). ). If the difference calculation result for all pixels is less than or equal to the reference value, it is determined as a non-defective product (S
7). Note that the reference value is a value at which only a defect to be detected can be detected, and is determined by inputting an image of a sample to be inspected having such a defect.

[0004]

Since the conventional defect inspection apparatus is constructed as described above, when there is a point (edge) which changes sharply on the surface to be inspected, that portion is generally used. The image value of is extremely larger than that of the surroundings, and the difference calculation result between the adjacent pixels becomes larger accordingly, so that there is a problem that the edge portion that is not a defect is determined as a defect. On the contrary, if the edge is taken into consideration, the reference value cannot be reduced, and the defect may not be detected.

The present invention has been made in order to solve the above-mentioned problems, and an apparatus capable of detecting only a defect, not a point (edge) on the surface of an inspection object that changes at an acute angle as a defect. Aim to get.

[0006]

A defect inspection apparatus according to claim 1 of the present invention is a defect-free inspection in advance when inspecting an inspection object having a point (edge) that changes sharply on the surface. Based on the target image signal, set a limit value as a threshold that can distinguish the non-edge part and the edge part, and replace the image value with the limit value when the input image value is larger than the limit value. The defect is determined from the above.

In the defect inspection apparatus according to claim 2 of the present invention, there are points (edges) on the surface that change at an acute angle, and the edges are located symmetrically with respect to a reference line set on the surface. When inspecting an object to be inspected, the grayscale image signal is input by scanning in the direction perpendicular to the reference line, and the density is maximized, and the density is distributed in a symmetrical position with respect to the reference line. The parts to be checked are excluded, and the defect is determined as the inspection range between those parts.

[0008]

In the defect inspection apparatus according to the first aspect of the present invention, when the image value of the inspection object is input, even if the image value of the pixel at the edge portion becomes extremely large, it is replaced with the limit value, so that the difference between the adjacent pixels is changed. The calculation result does not become large, and it is determined that the edge portion is not a defect.

In the defect inspection apparatus according to the second aspect of the present invention, when an image value to be inspected is input, a portion symmetrically with respect to the reference line is excluded from the portions having the maximum density. Since the defect is determined as the inspection range between the portions, only the true defective portion can be detected.

[0010]

【Example】

Example 1. Embodiment 1 of the present invention will be described below with reference to the drawings. Since the operation principle of inputting an image of an inspection object as a grayscale image signal, converting the image signal into a digital signal and performing data processing is the same as the conventional one, the configuration is the same as in FIG. FIG. 1 is a flowchart showing a process for setting an image limit value of an edge portion according to an embodiment of the present invention,
FIG. 2 is a flowchart showing a process of inspecting an inspection object having an edge using the limit value set in FIG. Further, FIG. 3 is a diagram for explaining changes in image data to be processed.

Next, the operation will be described. First, the process of obtaining the limit value of the edge portion from the inspection target having no defect will be described with reference to FIG. A defect-free inspection target is placed on the transport device 7 (T1), and an image signal of the surface of the inspection target is input (T2). The input image signal is converted into a digital signal by the A / D converter 2 (T3),
The image data is stored in the image memory 3 (T4).
With respect to the stored image data, the grayscale image value is sequentially increased by 1 from the lowest value within the input value range, and binarization processing is performed on each grayscale image value (T5). By the binarization process, a pixel having an image value larger than the grayscale image value at that time becomes "1", and a pixel having a smaller image value becomes "0".
Only the pixel of the edge part is "1" and the other pixels are "0". The threshold value of the grayscale image value and the lower limit value thereof are calculated as much as possible, and the value is stored as the limit value (T6). This limit value becomes the upper limit value of the non-defective portion of the input image value in the defect inspection by the processing of FIG. 2 (T10, T11) described later.

Next, the process of inspecting an inspection object having an edge will be described with reference to FIG. An image signal of the surface to be inspected is input (T7), the input image signal is converted into a digital signal by the A / D converter 2 (T8), and stored in the image memory 3 as image data (T9). The stored image data of each pixel is compared with a preset limit value (T10), and when the image data is larger, it is replaced with the limit value (T11). If the image data is smaller or equal, there is no need to replace it with the limit value. As in the conventional case, the image data that has been subjected to the process of replacing with the limit value is differentiated between adjacent pixels (T12), and the differential value and the reference value are compared (T12).
13) If there is a pixel having a larger differential value, it is determined as a defective product (T14). If the differential value is smaller for all the pixels, it is determined to be non-defective (T15).

The change of the image data by the above operation is shown in FIG.
It is shown using. As shown in FIG. 3A, it is assumed that a cylindrical inspection object 6 having a cut surface is inspected by this apparatus. Since there is a cut surface, there are edges on the surface.
At this time, the image signal on the straight line shown in FIG.
As shown in (c), the image value of the edge portion becomes extremely large, and the image value of the defective portion is only slightly different from the other portions. Conventionally, the input image value is directly subjected to the differential processing, and as a result, the differential value at the edge portion becomes large as shown in FIG. In this case, if the reference value is set so that the defect can be detected, the edge is also determined to be a defect, and if the reference value is increased so as not to detect the edge, the defect cannot be detected. On the other hand, in the first embodiment of the present invention, the limit value which is the lower limit value of the threshold value for distinguishing the non-edge portion and the edge portion is calculated based on the input image of the inspection target having no defect,
If the image value is larger than that, the limit value is replaced, so that the edge portion is smoothed with other portions as shown in FIG. As a result of the differential processing based on the image data smoothed in this way, only the differential value of the defective portion becomes large as shown in FIG. 3 (f), and only the defective portion is obtained by comparison with the reference value. Can be detected as a defect. Moreover, it is possible to set the reference value to a smaller value, and even smaller defects can be detected.

Example 2. Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 4 is a flowchart showing a process of inspecting an inspection object having an edge, and FIG. 5 is a diagram for explaining a change in image data to be processed. In this embodiment, the edges existing in the inspection target are distributed at positions symmetrical with respect to the reference line, as shown in FIG.

As shown in FIGS. 5A and 5B, the camera 1 scans the surface of the inspection object in the direction (X) perpendicular to the reference line and inputs an image signal (U1). The input image signal is converted into a digital signal by the A / D converter 2 (U2) and stored in the image memory 3 as image data (U3). For the stored image data, search for two pixels (pixels) that have the maximum density value for each line scanned by the camera, identify these pixels as points at which the surface shape changes sharply, and specify the center of this pixel. Is calculated and connected to form a reference line symmetrical to the image (U4).

The image of the two blocks existing symmetrically with respect to the reference line calculated in the above processing is excluded and differential processing is performed in the range between both images (U5), and the differential value and the reference value are compared ( U6) If there is a pixel having a larger differential value, it is determined as a defective product (U7). If the differential value is smaller for all pixels, it is determined to be non-defective (U8).

FIGS. 5C to 5F show the image data obtained by the above processing. That is, FIG. 5C shows the image signal input by the camera, but there is a large maximum value portion corresponding to the edge portion at a position symmetrical to the reference line, and if this is directly subjected to the differential processing, FIG. The output waveform is as shown in d), and the change amount of the edge portion of the cut surface is higher than the change amount (differential value) of the defect portion, so that the edge portion of the cut surface is extracted as a defect, not a true defect. Will be.

On the other hand, in the second embodiment of the present invention, as shown in FIG. 5 (e), the symmetrically distributed maximum value portions are excluded and the portion between them is subjected to the differential processing. Therefore, as shown in FIG. 5F, only the truly defective portion is extracted.

[0019]

As described above, according to the invention of claim 1, a limit value is set as a threshold value capable of distinguishing between the non-edge portion and the edge portion of the surface to be inspected, and the image value is less than the limit value. Since the image value of the edge portion appears as a maximum value which is symmetric with respect to the reference line, the maximum value portion is excluded. Since the inspection range is set between them, the defect can be accurately determined without erroneously determining the edge as the defect.

[Brief description of drawings]

FIG. 1 is a flowchart showing a process for setting an image limit value of an edge portion according to the first embodiment of the present invention.

FIG. 2 is a flowchart showing a process of inspecting an inspection object having an edge according to the first embodiment of the present invention.

FIG. 3 is a diagram for explaining changes in image data due to the processing shown in FIG.

FIG. 4 is a flowchart showing a process of inspecting an inspection object having an edge according to the second embodiment of the present invention.

FIG. 5 is a diagram for explaining changes in image data due to the processing shown in FIG.

FIG. 6 is a diagram showing a configuration of a conventional defect inspection apparatus.

FIG. 7 is a flowchart showing an inspection process by a conventional defect inspection apparatus.

[Explanation of symbols]

 1 signal source (camera) 2 A / D conversion device 3 image memory 4 data processing device 5 light source 6 inspection target 7 transport device

Claims (2)

[Claims] 1. An image to be inspected is input as a gray-scale image signal and stored as an image value of each pixel, and a means for calculating a change amount between each pixel and an image value of a surrounding is calculated. When the above-mentioned change amount is larger than a predetermined reference value, in the defect inspection apparatus that determines that the surface portion of the inspection target corresponding to the pixel has a defect, the point (edge) that changes sharply on the surface is In the case of inspecting a certain inspection target, based on the image signal of the inspection target having no defect in advance, a limit value is set as a threshold value capable of distinguishing the non-edge portion and the edge portion, and the input image is input. A defect inspection apparatus characterized in that when the value is larger than the limit value, the image value is replaced with the limit value and then the defect is judged. 2. An image to be inspected is input as a grayscale image signal and stored as an image value of each pixel, and a means for calculating a change amount between each pixel and the surrounding image value is calculated. When the above-mentioned change amount is larger than a predetermined reference value, in the defect inspection apparatus that determines that the surface portion of the inspection target corresponding to the pixel has a defect, the point (edge) that changes sharply on the surface is When inspecting an inspection object that exists and whose edges are distributed in symmetrical positions with respect to the reference line set on the surface, scan in the direction perpendicular to the reference line and input the grayscale image signal. A defect inspecting apparatus, characterized in that out of a portion having the maximum density, a portion distributed at a position symmetrical with respect to a reference line is excluded and a defect is determined as an inspection range between these portions.