JPS5952378B2 - Defect determination method for sheet materials - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for detecting and determining the type of defects present in a sheet-like material.

Conventionally, the methods used to identify defects in sheet-like objects, such as iron plates, paper, films, cloth, etc., were as follows: (1) Scanning a light spot over the surface to be inspected and photoelectrically converting the reflected or transmitted light. A method to identify defects by analyzing electrical signal waveforms.

(2) Transmit the transmitted light or reflected light from the surface to be inspected through a Fourier transform lens to obtain a correlation image of the defect pattern via a matched filter and collimating lens;
A method of determining defects based on the correlation image (Japanese Patent Laid-Open No. 49-6
0782).

However, in method (1), the electrical signal waveform obtained is complex, and especially when the object to be inspected is cloth, the cloth itself does not form a perfect grid, so the light used in method (1) is In the spot, the unique irregularities of the cloth are extracted as the electrical signal waveform of the normal part, and the S of the signal in the defective part is extracted.
Since the /N ratio is poor and the signal waveform is complicated, it is often practically impossible to analyze the signal waveform and determine the defect because it takes a long time to perform the determination. In addition, method (2) uses matched filters that display Fourier transformed images of various defect patterns, so many photoelectric conversion elements corresponding to various defects are required on a plane, and at the same time, the fabrication of matched filters is extremely difficult. It's difficult. As mentioned above, it is extremely difficult to put the above method into practical use, but the present invention improves these drawbacks, eliminates the complexity of the device, and makes it possible to identify defects in sheet materials at high speed. It is something.

The configuration for achieving this purpose includes a detection means that irradiates the surface to be inspected with coherent light to obtain a Fraunhofer diffraction image of the transmitted light or reflected light of the defective area, and obtains an electrical signal using a plurality of photoelectric conversion elements; The present invention is characterized by comprising a discriminating means for converting the received electrical signal into a binary code and discriminating defects based on the binary code pattern. Embodiments using the transmission method of the present invention will be described in detail below with reference to the drawings. In Fig. 1, 1 is a laser light source for irradiating coherent light, 2 and 2' are collimating lenses for increasing the laser beam, and a scanning light spot for inspecting a sheet-like object is set to a predetermined size. .

Reference numeral 3 denotes a rotating mirror for optically scanning the surface to be inspected, which is driven by a high-speed motor (not shown).

4 is a sheet-like material running in the longitudinal direction (in the direction of the arrow).

5 is a scanning light spot irradiated onto the surface to be inspected from the laser light source 1; 6 is a Fourier transform lens whose front focal plane is on the surface to be inspected, and its rear focal plane is at the position of the light receiving surface 7. It is located.

Coherent light emitted from the laser light source 1 passes through the surface to be inspected, passes through the Fourier transform lens 6, and forms a Fraunhofer diffraction image of the surface to be inspected on the light receiving surface 7 on the focal plane on the rear side. FIG. 2 is a schematic diagram showing a Fraunhofer diffraction image when the object to be inspected is a woven fabric, in which a shows the Fraunhofer diffraction image of a normal part of the woven fabric, and b shows a Fraunhofer diffraction image of a defective part. Due to the nature of Fraunhofer diffraction, the shape of the diffraction image does not change as the object to be inspected moves. The present invention has the following features. That is, the diffraction image is point symmetrical with respect to the origin (0th order diffraction point). Therefore, changes in the diffraction image due to various defects can be avoided by receiving only 1/2 of the image plane of the diffraction image.

As a result, the number of photoelectric conversion elements for receiving light can be reduced and the device can be extremely simplified.

The light-receiving surface 7 shown in FIG. 1 receives 1/2 of the entire Fraunhofer diffraction image 8 shown in FIGS. As shown in Figure 3, a plurality of photoelectric conversion elements Dl, D-2,...
・D-n is placed. FIG. 4 shows a block diagram of means for processing electrical signals from the light-receiving surface 7 of the Fraunhofer diffraction image to determine defects. In the binary code conversion section, a plurality of photoelectric conversion elements Dl in the previous stage diffraction image detection section,
The outputs of D-2, . . . D-n are divided into a plurality of levels of level values, and each level value signal is converted into a binary code (pulse signal). The defect discriminating section discriminates the type of defect by comparing it with a binary code pattern associated with each defect to be discriminated in advance.
At the same time, the defect position detection unit extracts the synchronization signal from the rotating mirror 3, determines the position of the defect, and inputs the defect type and the type into the storage unit for recording. The process of processing electrical signals input to the photoelectric conversion element to determine defects will be explained in more detail with reference to FIGS. 5 and 6.

That is, in FIGS. 5 and 6, a shows a normal part, b shows a partial view of a defective part, and shows an example of a light receiving part of a Fraunhofer diffraction image.

The output of each photoelectric conversion element in FIG. 5a is converted into a binary code (1 or 0) as shown in FIG. 6 by a binary code converter, and the output of each photoelectric conversion element in FIG. b
Convert to binary code like .

Two binary codes Ll and L in the upper and lower rows in Fig. 6A and b
2, as shown in Figure 7, the output of each photoelectric conversion element is divided into a certain appropriate level 1 (L1) and a lower level 2 (L1).
2) represents the signal divided into and.

If necessary, it may be divided into multiple level values of 3 or 4 levels.

Next, in a defect discriminating section, the output of each photoelectric conversion element in FIG. 6a or 6b is combined by a logic circuit and compared with combinations of various defects known in advance to determine the defect type.

In FIGS. 5 and 6, 25 photoelectric conversion elements are used. However, the present invention is not limited to this. In the embodiment, when detecting a Fraunhofer diffraction image of a defective part, the Fraunhofer diffraction image is directly received by the photoelectric conversion element. However, it is also possible to place a spatial frequency filter obtained from the Fraunhofer diffraction image in front of the light-receiving surface to remove the Fraunhofer diffracted light that overlaps with the normal area and to receive only the changes in the Fraunhofer diffraction image due to the defective area. .

Next, specific examples of the present invention will be described. (Example)
A 100 cm wide taffeta fabric was used as the sheet material, the traveling speed was 300 m/min, the rotation speed of the rotating mirror (hexagonal mirror) was 2500 min, the spot diameter of the 5 mw laser was 20 mm, and Fraunhofer diffraction of the transmitted light from the cloth surface was performed. They took out the statue and inspected the fabric.

Compared to visual inspection, we were able to detect defects and identify their types much faster and more accurately. The running speed is 3
00 m/Min to 350 m/Min is suitable, and in the case of thick fabrics, similar results can be obtained by slightly increasing the output of the laser beam. As described in detail above, in the present invention, a sheet-like object is irradiated with coherent light, a Fraunhofer diffraction image of transmitted light or reflected light from a defective part is received by a plurality of photoelectric conversion elements, and the obtained electrical signals are converted into multiple stages. This is a defect discrimination method that divides each level value signal into a binary code and identifies various defects based on the binary code pattern, so the device is simpler, lower cost, and faster than conventional methods. This makes it possible to identify defects in sheet materials and is easy to put into practical use.

[Brief explanation of drawings]

Figure] is a layout diagram of a defect detection device for a sheet-like object showing an embodiment of the defect discrimination method according to the present invention, Figures 2A and b are schematic diagrams of Fraunhofer diffraction images of defective areas in normal areas, and Figure 3. 4 is a block diagram showing the signal processing process, and FIGS. 5A and 5B are arrangement examples showing the relationship between Fraunhofer diffraction images and photoelectric conversion elements in normal and defective areas, respectively. For example, FIGS. 6A and 6B are diagrams showing the binary code conversion output of each photoelectric conversion element output, and FIG. 7 is a diagram showing the two-division level for converting into a binary code.

Claims (1)

[Claims] 1. A detection means for obtaining a Fraunhofer diffraction image by irradiating a sheet-like object with coherent light and optically converting the transmitted light or reflected light, and obtaining an electrical signal using a plurality of photoelectric conversion elements provided on the diffraction image plane; It is characterized by comprising a discriminating means that divides the electrical signal obtained from the photoelectric conversion element into a plurality of levels of level values, converts each level value signal into a binary code, and discriminates various defects based on the binary code pattern. A method for determining defects in sheet-like materials.