JP3175347B2 - Inspection method for foreign substances in sheet products - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for inspecting foreign matter mixed in a sheet-like product for inspecting foreign matter mixed in the sheet-like product.

[0002]

2. Description of the Related Art If foreign matter is mixed in a sheet-like product such as paper, the value of the product becomes low, and it is necessary to inspect the product.

[0003]

In the method of inspecting foreign matter, the visual inspection is inaccurate due to individual differences.
Further, there is a problem that even if an attempt is made to optically detect a foreign substance, if the foreign substance is buried in the sheet, it cannot be detected merely by looking at the surface.

An object of the present invention is to solve the above-mentioned problems and to provide a method for inspecting foreign matter in a sheet-like product, which can surely optically detect the presence of foreign matter.

[0005]

In order to achieve the above object, the present invention illuminates a sheet-like product with light from the back side, images the surface with a CCD camera, and sets a level on an image taken by the CCD camera. Set the area of the blacker part
Extract with two large and small area thresholds, small area threshold
The following foreign substance detection is performed according to the small area threshold.
Determine the filter size and match it with the filter size
The minimum value filter is applied to the image resulting from applying the maximum filter
Image leaving a larger area than the small area threshold
Is obtained from the image and the image taken with the above CCD camera
Is subtracted to obtain the gray level difference of each pixel, and this gray level difference is set.
Extract only the part that exceeds the threshold compared to the threshold
To extract small foreign matter by shading morphology.
Images from images of good sheet products taken with a CD camera
The threshold for binarization is determined based on the image statistics,
With the threshold for binarization, the image to be inspected for foreign matter is defined as 2
Value and extract the area above the large area threshold.
Images and images obtained by the above-mentioned shading morphology processing.
The image obtained by the image statistic processing is synthesized to obtain the foreign object image.
The number and area of the parts recognized in this foreign matter image
To determine the quality of the product .

[0006]

According to the above arrangement, by irradiating the sheet-like product with light and imaging the transmitted light with the CCD camera, the difference in light and shade between the portion having the foreign matter and the portion having no foreign matter appears remarkably. Image morphology processing
Large foreign matter is the field along with be out looking for a significant part of the shading difference Te
Detect foreign matter easily by searching by image statistics processing
By calculating the number and area of these,
The quality of goods can be judged.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the accompanying drawings.

In FIG. 1A, a sheet 11 for detecting foreign matter is disposed on a light projecting device 10, and a monochrome CCD camera 12 for imaging the sheet 11 together with light transmitted through the sheet 11 is disposed above the sheet 11. Is done. The signal of the monochrome CCD camera 12 is input to the image processing device 13, and the computer 14 inspects the foreign matter for contamination.

FIG. 1 (b) shows an algorithm for detecting foreign matter.
As shown in (1), first, inspection is started, and when an image is input, foreign matter is extracted from the image based on density morphology and image statistics, the area and number of the foreign matter are measured, and the quality of the product is determined. End the inspection. The aperture of the lens attached to the monochrome CCD camera is adjusted in advance so that an appropriate inspection can be performed.

The inspection of foreign matter contamination based on the density morphology and image statistics will be described in detail with reference to FIG.

First, it is assumed that an image 20 of one screen taken by a monochrome CCD camera is composed of 512 horizontal pixels and 512 vertical pixels as shown in FIG. For example, when the signal of this pixel is expressed in 8-bit gray scale, the level of the signal is variable up to 0 for black (no transmitted light) and 255 for white (100% transmission). Portions 21a, 21b, 22c to 21e of the image 20 which are darker than the set level are indicated by oblique lines.

Each of the parts 21a, 21b, 22c to 21
The foreign substance of e is determined by two large and small area thresholds S1 and S
2 and the parts 21a, 21b, 22c to 21e
Area of, the shading morphology if small area threshold S1 below, is extracted by the use of large area threshold S2 higher if it images statistics.

Next, detection of foreign matter by shading morphology will be described. The size of the filter is determined in accordance with the area threshold value S1, and the maximum value filter is applied to the image 20 in accordance with the filter size. 2
An image 24 leaving only 1a and 22d is obtained. As a result, the level of each pixel at the position of the AA line of the image 20 is represented by the curve 26 of (d), but the level of each pixel at the position of the AA line is (d) in the image 24. Curve 25)
It is represented by By subtracting the curve 26 from the curve 25 in (d), a gray level difference is obtained as shown in (e). Similarly, the gray level difference is calculated for all the vertical directions as described above, that is, the image 20 is subtracted from the image 24 to obtain the gray level difference of each pixel,
Compared with the set threshold value T1, if only the portion exceeding the threshold value T1 is left as the pixel value level 255 and the other portions are left as the pixel value level 0, as shown in FIG. S1 or less, and the difference in density from the surrounding
Image 2 obtained by binarizing portions 22c and 22e larger than 1
7 is obtained.

Next, detection of foreign matter based on image statistics will be described. Before inspection, a non-defective image is input in advance, and the average gray level m and the standard deviation σ of the image are obtained.
From the standard deviation σ and the threshold value T2 for binarization (= m + k * σ
+ B k: arbitrary coefficient, b: bias value). When the image 20 is binarized by the threshold value T2 and a portion equal to or larger than the area threshold value S2 is left, as shown in FIG. An image 23 including the portion 21a above S2 is obtained.

Image 27 obtained by shading morphology processing
And the image 23 obtained by the image statistic processing are combined to obtain the foreign substance image 28 as shown in FIG. Then, the number and the area of the portions 21a, 22c, and 22e recognized as the foreign matter are obtained, and the quality of the product is determined from these.

In this embodiment, a monochrome CCD camera is used. However, even if a color CCD camera is used, the inspection can be performed by performing the same processing for each of the R, G, and B components.

In this case, the image size is set to 512 × 51.
Two pixels, the gradation of the image grayscale value is 256 gradations of 8 bits,
Although the gray level level 0 is described in detail as black and 255 as white, it is possible to inspect foreign matter even if an image size and a gradation level are selected without being limited to this.

As an example of the foreign substance extraction processing, both the density morphology and the image statistic are used. However, the foreign substance extraction and inspection can be carried out even when individually used according to the foreign substance to be extracted.

[0019]

In summary, according to the present invention, light is applied to a sheet-like product, and the transmitted light is imaged by a CCD camera. Therefore, strike both left looking for a significant portion of the shading difference from the image of the CCD camera at gray morphology processing
Large foreign objects can be easily found by using image statistics processing.
Foreign matter can be detected in the
Thus, the quality of the sheet product can be determined.

[Brief description of the drawings]

FIG. 1 is a diagram showing one embodiment of the present invention.

FIG. 2 is an explanatory diagram for inspecting a foreign substance in the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Floodlight device 11 Sheet-shaped product 12 CCD camera

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Masahiro Yagabe 2-1-1, Toyosu, Koto-ku, Tokyo Ishikawa Shima-Harima Heavy Industries Co., Ltd. Tokyo First Plant (72) Inventor Junichi Ito Toyosu, Koto-ku, Tokyo 2-1-1, Ishikawa Shima-Harima Heavy Industries Co., Ltd. Tokyo 1st Factory (56) References JP-A-58-173456 (JP, A) JP-A-61-89544 (JP, A) JP-A-3-156349 (JP, A) Boerner, "FEATURE EXTRACTION BY GR AYSCALE MORPHOLOGI CAL OPERATIONS-A COMPARISON TO DOG FILTERS. 112- 117 Philippe Salembie r, "MULTIRESOLUTION DECOMPOSITION AND ADAPTIVE FILTERIN G WITH RANK ORDER BASED FILTERS - AP PLICATION TO DEFEC T DETECTION -", 1991 International Conf erence on IEEE Aco ustics, Speech, and Signal Processing, 1991, Vol. 4, p. 2389-2392 (58) Field surveyed (Int. Cl. ⁷ , DB name) G01N 21/84-21/958 JICST file (JOIS)

Claims (1)

(57) [Claims] 1. A sheet-like product is irradiated with light from its back side, its surface is imaged by a CCD camera, and an area of a portion which is blacker than a set level on an image taken by the CCD camera is set.
Extract with two large and small area thresholds, small area threshold
The following foreign substance detection is performed according to the small area threshold.
Determine the filter size and match it with the filter size
The minimum value filter is applied to the image resulting from applying the maximum filter
Image leaving a larger area than the small area threshold
Is obtained from the image and the image taken with the above CCD camera
Is subtracted to obtain the gray level difference of each pixel, and this gray level difference is set.
Extract only the part that exceeds the threshold compared to the threshold
To extract small foreign matter by shading morphology.
Images from images of good sheet products taken with a CD camera
The threshold for binarization is determined based on the image statistics,
With the threshold for binarization, the image to be inspected for foreign matter is defined as 2
Value and extract the area above the large area threshold.
Images and images obtained by the above-mentioned shading morphology processing.
The image obtained by the image statistic processing is synthesized to obtain the foreign object image.
The number and area of the parts recognized in this foreign matter image
A method for inspecting for contamination of a sheet-like product by determining whether the product is good or bad .