JPH0735699A - Method and apparatus for detecting surface defect - Google Patents

Abstract

PURPOSE:To exactly detect a defect by simple method and apparatus by obtaining a processed image of a gross mean gradation difference from a difference of a gross mean of a gradation value of each pixel included in a predetermined area of the image and a gross mean of a gradation value when the area is deviated by one pixel. CONSTITUTION:A surface state of an object to be inspected is imaged by a CCD camera 3, and fetched by image fetching means 11 as an image for forming with an assembly of many pixels having information of brightness. Then, image processing means 12 obtains a gross means of a gradation value of each pixel of a predetermined area of the image and a gross mean of a gradation value when the region is deviated by one pixel, and inputs a difference of both the gross means as a gross mean gradation difference at a center of the region before it is deviated by one pixel. This process is repeated for all the area of the image, and a process image of the assembly of the gross mean gradation difference is obtained. Then, defect extracting means 13 decides an extracted level corresponding to a surface defect desired to be detected for the image, processes to extract it with it as a reference, and the surface defect can be exactly detected together with its position.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a surface defect detecting method and an apparatus thereof, and more particularly to a surface defect detecting method and an apparatus thereof for detecting microscopic defects occurring on the surface of a press work or the like by image processing.

[0002]

2. Description of the Related Art Generally, detection of a minute defect in a press work or the like often depends on the operator's visual inspection, and the detection work often depends on the operator's sensuality. Therefore, there is a problem that it is difficult for a skilled worker to detect the above-mentioned defects.

Therefore, development of a method and an apparatus capable of accurately performing the above defect detection without relying on the visual inspection of an operator is awaited.

To meet the above demands,
There is a method in which the state of the object to be inspected is captured as an image composed of a large number of pixels having brightness information and defect detection is performed based on the image. In this case, each pixel in the captured image is There is a problem that it is difficult to detect a defect because the difference in brightness is not clear in a minute defect.

It should be noted that the state of the object to be inspected is detected as image information by a plurality of sensors through the lens of the camera, and a plurality of image information signals corresponding to the number of the sensors are aggregated to obtain a lightness gradation. It has been proposed to precisely detect (see Japanese Patent Laid-Open No. 216768/1990).

[0006]

However, in the above-mentioned known example, since image information signals detected by a plurality of sensors are integrated, a very large number of sensors are required and the amount of information is also increased. There is a problem that the number of devices must be extremely large, and it is difficult to avoid complicated devices and complicated information processing work.

The present invention has been made in view of the above points, and an object of the present invention is to make it possible to accurately detect a minute defect of an object to be inspected by an extremely simple method and means.

[0008]

According to the method of claim 1, as a means for solving the above-mentioned problems, an image formed by a set of a large number of pixels indicating the surface state of an object to be inspected has brightness information. And the sum total average of the gradation values of each pixel included in a predetermined area of the image and the sum average of the gradation values of each pixel included when the area is shifted by one pixel are obtained. The process of inputting the average difference as a total average gradation difference in the center of the region before shifting one pixel is repeatedly performed on all the regions of the image, and the processed image including the set of the total average gradation differences is processed. A surface defect is detected by performing an extraction process based on a predetermined extraction level.

According to a second aspect of the present invention, as a means for solving the above-mentioned problems, an image capturing means for capturing the surface condition of the object to be inspected as an image composed of a large number of pixels having brightness information. And a sum average of gradation values for each pixel included in a predetermined area of the image and a sum average of gradation values for each pixel included when the area is shifted by one pixel, and both sum averages Image processing means for repeatedly performing the process of inputting as the sum total average gradation difference in the center of the region before shifting the difference of 1 pixel for all the regions of the image, and the processed image composed of the set of the sum total average gradation differences. On the other hand, there is provided a defect extraction means for performing an extraction process based on a predetermined extraction level.

[0010]

According to the method of claim 1 or the apparatus of claim 2,
The following actions are obtained by the above means.

That is, the surface state of the object to be inspected is taken in, and the sum average of the gradation values of each pixel included in a predetermined area of the image formed by a set of many pixels having the brightness information, A process of obtaining a sum average of gradation values for each pixel included when the region is shifted by one pixel and inputting the difference between the two sum averages as the sum average gradation difference in the center of the region before shifting by one pixel Since it is repeated for all regions of the image, the sum average gradation difference in the obtained processed image is weighted for each pixel in the captured image, and then the predetermined extraction is performed. When the extraction process is performed with the level as a reference, the defective portion is accurately extracted.

[0012]

According to the method of the first aspect or the apparatus of the second aspect, the predetermined area of the image which is obtained by incorporating the surface state of the object to be inspected and which is composed of a large number of pixels having the brightness information. Before calculating the sum average of the gradation values of each pixel included in the above and the sum average of the gradation values of each pixel included in the case where the area is shifted by one pixel, the difference between the two sum averages is shifted by one pixel. Since the process of inputting the total average gradation difference in the center of the region is repeatedly performed for all the regions of the image, the total average gradation difference in the obtained processed image is Since each pixel is weighted, if the extraction process is performed with a predetermined extraction level as a reference, the defective portion will be accurately extracted, and the inspected object will be extracted by an extremely simple method / apparatus. Missing things An excellent effect that it is possible to accurately detect. Moreover, since it is possible to set an area for obtaining the total average gradation difference depending on the size of the defect, there is an effect that the processing for obtaining the total average gradation difference can be performed by adapting to the size of the defect.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

In this embodiment, the present invention is applied to detect a minute defect in a press work W, and FIG.
As shown in FIG. 3, the press work W positioned on the jig 1 is irradiated with the fiber illumination 2 at an angle of 10 to 20 °, and the CCD positioned on the opposite side of the fiber irradiation 2 is irradiated.
An image obtained by shooting with the camera 3 is subjected to image processing by the method described below to detect defects.

As shown in FIG. 2, the surface defect detecting apparatus of the present embodiment takes a photograph of the surface state of the press work W, which is the object to be inspected, by the CCD camera 3. An image capturing means 11 for capturing as an image A (see FIG. 3) composed of a set of pixels a, a ... And a gradation for each pixel a included in a predetermined area S (see FIG. 4) of the image A. The sum average of the values and the sum average of the gradation values of each pixel a included when the region S is shifted by one pixel are calculated,
An image processing unit 12 that repeatedly performs a process of inputting a total average gradation difference T (described in detail later) in the center of the region before shifting the sum total average by one pixel for all regions of the image A.
And a defect extraction means 13 for performing an extraction process using a predetermined extraction level as a reference for the processed image A'consisting of the set of the total average gradation differences. Reference numeral 14
Is a display means (for example, CRT or hard copy) for displaying the defects extracted by the defect extraction means 13.

A method of detecting a surface defect of the press work W by the surface defect detecting device having the above structure will be described.

An image is taken by the CCD camera 3 and is taken in as an image A (see FIG. 3) composed of a set of a large number of pixels a, a ... Having brightness information. The image A is a gray level raw image, and has a difference in brightness based on, for example, 256 gradation values.

Next, as shown in FIGS. 4 and 5, the sum total average of the gradation values for each pixel a included in a predetermined area S (for example, an area including i × j pixels) of the image A and , The sum average of the gradation values of each pixel a included when the region X is shifted by one pixel in the direction of the horizontal axis X, and the difference between both sum averages (that is, the sum average gradation difference) T = P {X + (j / 2), Y + (i /
2)} is calculated by the following equation.

[0019]

[Equation 1]

Here, P: gradation value for each pixel, Y: vertical axis, the sum average gradation difference T obtained as described above is shown in FIG.
As shown in, the input is made in the center of the area before shifting by one pixel.

The above-described calculation of the total average gradation difference T and the input processing are repeatedly performed for all pixels in the image A by shifting by one pixel in the horizontal axis X direction, and the processing for all pixels in the horizontal axis X direction is completed. Then, as shown in FIG. 7, by shifting one pixel in the Y direction of the vertical axis, the same process as described above is performed, and the calculation and the input process of the total average gradation difference T are performed for all the regions in the image A. The above processing is executed by the image processing means 12.

When the sum total average gradation difference T obtained as described above is classified into a color difference expression or a bird's-eye view by classifying the gradation difference into several stages, a processed image A'as shown in FIG. 8 is obtained. can get.

From the above results, when the extraction level corresponding to the surface defect to be detected is determined and the surface defect is extracted with the extraction level as a reference, the surface defect N is accurately detected together with the position as shown in FIG. . The processing is carried out by the defect extracting means 13.

As described above, according to the present embodiment, the surface state of the press work W is taken in, and the predetermined area S of the image A formed by a set of a large number of pixels a, a ... Having brightness information. Of the gradation values of the respective pixels a included in the above and the average of the gradation values of the respective pixels included in the case where the area S is shifted by one pixel, and the difference between the two total averages is calculated by one pixel. Since the process of inputting the total average gradation difference T in the center of the region before shifting is repeatedly performed for all the regions of the image A, the total average gradation difference in the obtained processed image A ′ is obtained. T is weighted for each pixel a in the captured image A. Therefore, after that, when the extraction process based on the predetermined extraction level is performed, the defective portion is accurately extracted, and the defect N in the press work W as the inspection object is detected by the extremely simple method and device. It can be detected accurately. Moreover,
Since it is possible to set the area S for obtaining the total average gradation difference T according to the size of the defect N, the processing for obtaining the total average gradation difference T can be performed by adapting to the size of the defect.

[Brief description of drawings]

FIG. 1 is a side view showing an image pickup state in a surface defect detecting apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram showing the contents of a surface defect detection device according to an embodiment of the present invention.

FIG. 3 is a plan view showing an image captured by the surface defect detection apparatus according to the embodiment of the present invention.

FIG. 4 is an explanatory diagram illustrating a process of obtaining a total average gradation value in a predetermined area for an image captured by the surface defect detection apparatus according to the embodiment of the present invention.

FIG. 5 is an explanatory diagram illustrating a process of obtaining a total average gradation value when a predetermined region is shifted by one pixel in an image captured by the surface defect detection apparatus according to the embodiment of the present invention.

FIG. 6 is a sum average gradation value in a predetermined area obtained from an image captured by the surface defect detection apparatus according to the embodiment of the present invention and a sum average gradation when the area is shifted by one pixel in the horizontal axis direction. FIG. 9 is an explanatory diagram illustrating an input state of a difference from a value (that is, a total average gradation difference).

FIG. 7 is an explanatory diagram illustrating an input state of a sum average gradation difference when the area is shifted by one pixel in the vertical axis direction.

FIG. 8 is an explanatory diagram showing a processed image obtained as a set of the total average gradation differences.

FIG. 9 is an explanatory diagram showing a result of performing an extraction process on the processed image based on a predetermined extraction level.

[Explanation of symbols]

3 is a CCD camera, 11 is an image capturing means, 12 is an image processing means, 13 is a defect extracting means, A is an image, a is a pixel,
A'is a processed image, N is a defect, S is a region, T is a total average gradation difference, and W is an object to be inspected (press work).

Claims (2)

[Claims] 1. A surface state of an object to be inspected is captured as an image composed of a large number of pixels having brightness information, and a sum average of gradation values of respective pixels included in a predetermined region of the image is acquired. And the sum average of the gradation values for each pixel included in the case where the area is shifted by one pixel, and the difference between the two sum averages is input as the sum average gradation difference in the center of the area before shifting by one pixel. The processing is repeatedly performed on the entire area of the image, and the surface defect is detected by performing the extraction processing based on a predetermined extraction level on the processed image including the set of the total average gradation differences. And a method for detecting surface defects. 2. An image capturing means for capturing a surface state of an object to be inspected as an image composed of a large number of pixels having brightness information, and a floor for each pixel included in a predetermined area of the image. The sum total average of the tone values and the sum average of the gradation values for each pixel included when the region is shifted by one pixel are obtained, and the difference between the two sum averages is shifted by one pixel. An image processing unit that repeatedly performs the process of inputting as a tone difference on the entire region of the image, and performs an extraction process with a predetermined extraction level as a reference for a processed image composed of a set of the total average gradation differences. A surface defect detecting apparatus comprising: a defect extracting means.