JP2000035321A - Inspection device for surface of metal zone - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a surface inspection device capable of conducting grading of metal zone automatically and sufficiently exactly. SOLUTION: For a surface inspection device provided with an image sensor for obtaining the image of surface of metal zone, a multiralued data means and a data binarizing means for obtaining multiralued data and binary data by using the detected signal of the image sensor, the degree of a flaw is judged by the first judging means 41 based on the multiralued data, the kind of the flaw is judged by the second judging means 42 based on the binary data and a flaw kind and degree judgment means 43 judges the kind and degree of the flaw from these judged results. With a representing flaw determining means 48, the maximum flaw of weighted value is selected as a representing flaw, total sum of weighted value of all flaw in a flaw evaluation range is obtained with a weighted value operation means 49 and a grading judgment means 50 grades from a grading judgment table 46 based on the kind of the representing flaw and the total sum of weighted value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a surface inspection of a metal strip, such as a stainless steel strip, which passes through a continuous processing line such as bright annealing, for automatically inspecting a surface flaw of the metal strip, and determining a grade. Related to the device.

[0002]

2. Description of the Related Art For example, in a production line for a metal strip containing stainless steel or the like, a metal steel strip is manufactured through various processes such as rolling, annealing, and pickling. After rolling, the deformed crystal grains of the metal band are heated in an annealing furnace to recrystallize and soften, and descaling is performed in an acid washing process to remove scales and the like adhered by the annealing, and further washed with water and dried. The final product is obtained by winding it as a coil and passing through a number of processes.At each stage of these processes, the state of the front and back surfaces of the metal band is visually inspected, and Grades are graded for each unit length of the final product according to the extent of such defects (heavy, medium, and light) such as flaws and scales.

As described above, inspection of defects on the front and back surfaces of a metal strip is conventionally performed by human eyes, which requires human cost and that the inspection results are affected by the physical condition of the inspector. I can't avoid it. In addition, this inspection place is not always in a favorable working environment, and it is required to avoid personnel arrangement as much as possible.

Accordingly, the present applicant has proposed a surface inspection apparatus disclosed in Japanese Patent Application Laid-Open No. 9-89803, which can automatically perform a grade rating of a metal strip. This surface inspection apparatus uses scanning means for scanning the surface of the metal band with laser light, light receiving means for receiving regular reflection light and irregularly reflected light of the laser light from the metal band surface, and light receiving detection signals of the light receiving means. Means for obtaining multi-valued data for each pixel of the quantized image, and binary data generating means for obtaining binary data for each pixel of the quantized image using a light reception detection signal of the light receiving means. First determining means for determining the degree of the flaw based on the obtained multi-value data, second determining means for determining the form of the flaw based on the obtained binary data, first and second
And a weight value determining means for determining a flaw weight value for each flaw from the flaw weight table based on the judgment result of the determining means. The weight value determined by the weight value determining means is added for each type of flaw form within the flaw evaluation range, for example, for linear flaws, point flaws, and planar flaws, and the sum of the weight values for each form of flaw is calculated by the calculating means. Required by The sums of the weight values obtained by the calculating means are compared with each other, and the flaw type determining means determines the flaw having the largest weight value sum as a flaw existing in the flaw evaluation range. Then, a grade rating for the flaw evaluation range is determined from the rating evaluation table based on the sum of the weight values of all the flaws within the flaw evaluation range.

[0005]

In the above-described surface inspection apparatus, the grade of the product is automatically determined by using the form of the flaw and the weight value of the flaw. The flaws within the range are divided into linear flaws, point flaws and planar flaws,
The flaw type within the flaw evaluation range is determined based on the sum of the weight values of the flaws of each mode. Therefore, the grading obtained by the grading judgment means based on the grading evaluation table is not sufficiently accurate depending on the kind of the flaw, and particularly when the grading includes a small and inferior one, the evaluation by the operator and the surface inspection device are required. Was not consistent with the rating.

It is an object of the present invention to provide an apparatus for inspecting the surface of metal strips, which is capable of automatically and sufficiently accurately grading metal strips.

[0007]

SUMMARY OF THE INVENTION The present invention provides an image sensor for obtaining an image of the surface of a metal strip traveling in a longitudinal direction, and a method for detecting flaws present in a flaw evaluation range by using a detection signal of the image sensor. Multi-level data conversion means for obtaining value data, binary data conversion means for obtaining binary data on flaws present in the flaw evaluation range using the detection signal of the image sensor, and multi-level data conversion means obtained by the multi-level data conversion means First determining means for determining the degree of each of the flaws present in the flaw evaluation range based on the multi-valued data, and present in the flaw evaluation range based on the binary data obtained by the binary data converting means. Second determining means for determining the form of each flaw to be performed; weight value determining means for determining a flaw weight value for each flaw from a flaw weight table based on the determination results of the first and second determining means; The weight A representative flaw determining means for selecting a flaw having a maximum weight value from the weight values determined by the determining means and determining the flaw as a representative flaw, and adding a flaw weight value of each flaw existing in the flaw evaluation range. Weight value calculating means for obtaining the sum, and rating determining means for performing a grade rating from a rating determination table based on the form of the flaw determined by the representative flaw determining means and the weight value sum calculated by the weight value calculating means. It is a metal strip surface inspection apparatus characterized by comprising:

According to the present invention, the multi-value data generating means generates multi-value data using the detection signal of the image sensor,
The binary data generating means generates binary data using the detection signal of the image sensor, and the first determining means determines the degree of each flaw present in the flaw evaluation range based on the obtained multi-value data. Then, the second determining means determines the form of each flaw present in the flaw evaluation range based on the binary data. The weight value determining means determines a flaw weight value for each flaw from the flaw weight table based on the determination results of the first and second determining means, and the representative flaw determining means represents a flaw having the maximum flaw weight value. Determined as flaw. As described above, since the flaw weight value is compared for each flaw and the largest one is determined as the representative flaw,
The representative flaw accurately represents a flaw existing within the flaw evaluation range. In addition, since the rating judging means performs a grade rating from the rating judgment table based on the determined representative flaw form and the total weight value obtained by the weight value calculating means, the grade rating of the flaw evaluation range is sufficiently accurate. Becomes
In particular, since the form of the representative flaw is determined accurately, its rating is accurate.

Further, the flaws present in the flaw evaluation range of the present invention are classified into any of point flaws, linear flaws and planar flaws, and the rating of the rating judgment table is determined by the weight value calculating means. It is characterized in that linear flaws, point flaws and planar flaws are individually set with respect to the total of the weight values.

According to the present invention, the flaws present on the surface of the metal strip are classified into any of point flaws, linear flaws and planar flaws. The linear flaws tend to be inferior to many in spite of their small area, and this tendency also exists to some extent in point flaws, and this tendency is small for planar flaws. For this reason, the rating of the rating table is set individually for the linear flaws, point flaws, and planar flaws with respect to the sum of the weight values by the weight value calculating means, thereby performing a more accurate grade rating. be able to.

[0011]

FIG. 1 is a simplified side view showing a part of an embodiment of the present invention. The metal strip 1 such as stainless steel, which is a long material, is manufactured through several steps. In this embodiment, the strip is brightly annealed and travels in the direction of the arrow 2. In the configuration shown in FIG. 1, an electric signal having a level corresponding to the flaws on both the front and back surfaces of the metal strip 1 is obtained, and thereafter, the electric signal is arithmetically processed as described later, and the type of the flaw, that is, a linear flaw, Point flaws and planar flaws are determined.

A surface inspection means 6 is arranged on the guide roll 5 for inspecting flaws on the front side surface of the metal strip 1, and the other guide roll 4 is provided on the back surface of the metal strip 1. A back surface inspection means 7 is arranged to inspect for flaws.

The axes 8, 9 of the guide rolls 4, 5 are parallel to one another. The metal strip 1 is arranged such that the guide rolls 4, 5 are rotated in directions 10, 10a opposite to each other when viewed from one side in the axial direction of the guide rolls 4, 5, respectively. Wound around. In FIG. 1, the guide roll 4 is rotated clockwise 10 and the guide roll 5 is rotated counterclockwise 1.
Rotated to 1. Although these guide rolls 4 and 5 may be configured to be driven to rotate, they may be configured to be driven as the metal strip 1 travels.

The surface inspection means 6 includes an irradiation lamp 12 and a CC
It is composed of a D camera 14. The irradiation lamp 12 emits light toward the front surface of the metal strip 1, and the light reflected from the surface of the metal strip 1 is received by the CCD camera 14. CCD
The camera 14 has a CCD (charge) as an image sensor.
coupled device) 16 (FIG. 2).
6 obtains an image of the surface of the metal strip 1. The back surface inspection means 7 has substantially the same configuration as the front surface inspection means 6, and includes an irradiation lamp 18 and a CCD camera 20 having a CCD.

The detection signal detected by the CCD 16 is processed as shown in FIG. The CCD 16 has, for example, 2048 bits and detects the entire width of the metal band 1 with 2048 bits. The detection signal of the CCD 16 is first sent to the average value calculating means 22. The average value calculation means 22 averages the detection signals of the CCD 16 and calculates and calculates a noise component. The averaging of this detection signal is CC
One scan by D16 may be used, but it is preferable to average the detection signals of a plurality of scans, for example, about 64 scans in order to improve the accuracy.

The average signal averaged by the average calculating means 22 is sent to the differential calculating means 24. A detection signal from the CCD 16 is supplied to the differential operation means 24. The differential calculating means 24 subtracts the average value obtained by the average value calculating means 22 from the detected value of the CCD 16 and performs the arithmetic processing in this manner, thereby obtaining a flaw corresponding to the flaw existing on the surface of the metal strip 1. A detection signal is generated.

The flaw detection signal generated by the differential operation means 24 is sent to a defect extraction means 28. The defect extracting means 28 is provided with a level setting means 30. The level setting means 30 sets a cutting level to be determined as a flaw. By changing the level value of the cutting level, a determination level for determining a flaw is determined. Can be adjusted. The defect extraction means 28 cuts the flaw detection signal at the extraction level set by the level setting means 30, and the obtained extraction signal is stored in the storage means 32 as multivalued data relating to the flaw. Similarly, the cut-out signal is sent to the binary data conversion means 26, converted into binary data by the binary data conversion means 26, and the converted binary data is stored in the storage means 32.

The average value calculating means 22, the differential calculating means 24, the binary data converting means 26 and the defect extracting means 28 for processing the detection signal of the CCD 16 are constituted by a microcomputer, for example. The multi-level data conversion means is configured. Then, the binary data conversion means 26 generates, for example, binary data as shown in FIG. 3, and the multi-value data conversion means generates, for example, multi-value data as shown in FIG. Note that CC
The conversion of the detection signal of D16 into binary data and multivalue data is performed, for example, for each scan by the CCD 16.

In this embodiment, a grade evaluation unit range (hereinafter referred to as a “flaw evaluation range”) for performing rating by inspection of the metal strip 1 is, for example, assuming that the width W is the entire width of the metal strip 1 and the length is H is set to 1 m, and this 1 m range is scanned, for example, by 1024 scans. Then, for each of the flaw evaluation ranges, quantization, profile creation, feature extraction, and flaw type flaw form determination are performed to determine a grade rating.

FIG. 5 is a simplified block diagram showing a signal processing circuit 40 realized by a microcomputer, and FIG. 6 is a flowchart for explaining a schematic operation of the signal processing circuit 40. Referring to FIG. 5, storage means 32 for storing binary data and multi-valued data is connected to signal processing circuit 40. The signal processing circuit 40 includes a first determination unit 41 that determines the degree of the flaw based on the multi-valued data, a second determination unit 42 that determines the form (type) of the flaw based on the binary data, And a flaw type / degree determining means 43 for determining the type and degree of the flaw based on the determination results of the second determining means 41 and 42. The table storage means 44 includes a flaw weight table 45 for determining a weight value, a rating determination table 46 for determining a grade rating of the metal strip 1, and each flaw determined by the flaw type / degree determination means 43. , And a weight value memory 47 for storing the weight values. The flaw type / degree determination means 43
The weight value is read from the flaw weight table 45 based on the determined type and degree of the flaw, and the read weight value of each flaw is stored in the weight memory 47.

The table storage means 44 further includes a representative flaw determining means 48, a weight value calculating means 49 and a rating judging means 50. The representative flaw determining means 48 compares the weight values of the flaws present in the flaw evaluation range and determines the flaw having the largest weight value as the representative flaw. The weight value calculating means 49 adds the weight values stored in the weight value memory 47, that is, the weight values of all the flaws present in the flaw evaluation range, and calculates the sum of them. Further, the rating determining means 50 performs a rating from the rating determination table 46 based on the determined type of the representative flaw and the obtained sum of the weight values.

The grading by the signal processing circuit 40 and the table storage means 44 is performed according to the flowchart shown in FIG. First, in step S-1, multi-value data is read from the storage means 32. That is, for each flaw present in the flaw evaluation range, its area and image density are determined, and based on these areas and image density, the first
Determination means 41 determines the degree of flaw. In the metal strip 1, the image density is low when there is a flaw, and the image density also decreases as the degree of the flaw becomes worse. Utilizing this, the image density per unit area is obtained by dividing the image density of each flaw by its area, and the degree of the flaw is determined based on the degree of the image density per unit area.

In this embodiment, as shown in the flaw weight table of Table 2 (stored in the table storage means 44), determination is made in ten stages from "0" to "9". It should be noted that the degree of this determination may be less than 10 levels depending on the type of the metal strip 1 to be evaluated, and may be set to 11 or more levels for more accurate determination.

[0024]

[Table 1]

After step S-1, step S-2 is performed. In step S-2, the binary data is read from the storage means 32, and then in step S-3, the binary data is read out.
The type and degree of the flaw are determined based on the value data and the multi-value data. For the read binary data, a distribution state of the presence / absence of a flaw in the flaw evaluation range is created, and this distribution state is created, for example, as shown in FIG. In FIG. 7, a black portion indicates a portion where a flaw exists, and is binary “1” in binary data. The type and size of the flaw are, for each flaw shown in FIG. 7, the length of the flaw (the length in the traveling direction 2), the width of the flaw (the length in the scanning direction), the area of the flaw, and the length of the flaw. The width ratio (length / width) is determined, and the form and size of the flaw are determined using these as parameters. In this embodiment, as shown in the flaw weight table of Table 2, flaw types, ie, flaw types, are classified into three types: linear flaws, point flaws, and planar flaws. Are classified into three types of sizes, "small", "medium" and "large". The types and sizes of the flaws may be other than those described above.

When the size of the flaw and the type of the flaw are determined, the process proceeds to step S-4, and a weight value is determined for each flaw existing in the flaw evaluation range. When determining this weight value,
The flaw weight table shown in Table 2 is used. For example, when the existing flaw is determined to be “medium” of a linear flaw (or a point flaw, a planar flaw) and the degree is determined to be “3”, The flaw type / degree determining means 43 reads “6” (or “6”, “5”) as a weight value from the flaw weight table, and the read weight value is stored in the weight value memory 47.

When the weight values for all the flaws in the flaw evaluation range are determined, the process proceeds to step S-5, where the representative flaw is determined. The representative flaw determining means 48 includes a weight value memory 47
Is selected from all the weight values stored in the table, and the type of the flaw having the maximum weight value is determined as a representative flaw (any of a linear flaw, a point flaw, and a planar flaw). The representative flaw is a flaw representing a plurality of flaws existing in the flaw evaluation range, and is used as a flaw type at the time of grading.

Next, proceeding to step S-6, the weight value calculating means 49 adds the weight values of all the flaws stored in the weight value memory 47 to obtain a total weight value. Then, step S
At -7, the grading is performed on the flaw evaluation range using the rating determination table shown in Table 3 (stored in the table storage means 44). The rating determining means 50 performs a rating from the rating determination table based on the selected representative flaws and the calculated sum of the weight values. For example, the representative flaw is a linear flaw (or a point flaw, a planar flaw), and the weight value sum Φ
Is "43", it is determined to be grade "D". In addition,
The arithmetic output of the signal processing circuit 40 is visually displayed on a visual display means 51 such as a cathode ray tube or a liquid crystal, or printed on a recording paper by a printer 52.

[0029]

[Table 2]

As described above, the weight value for each flaw existing in the flaw evaluation range is determined, and the largest weight value is selected as the representative flaw. Therefore, in the rating judgment using the rating judgment table, A grade judgment result that matches the judgment of the inspector is obtained. Further, in the rating judgment table, as shown in Table 3, the weight value sum
Grades are individually set for the linear flaws, point flaws, and planar flaws, so that a judgment result consistent with the judgment of the inspector is obtained. For example, when the total weight value Φ is “47”, the representative flaw is determined to be a point flaw or a planar flaw, and is determined to be grade “D”.
Is determined.

[0031]

According to the metal band surface inspection apparatus of the present invention, the flaw having the maximum flaw weight value among the flaws existing in the flaw evaluation range is selected as a representative flaw for rating judgment. Since the rating judgment is performed using the representative flaws and the sum of the weight values of the flaws existing in the flaw evaluation range, it is possible to perform a grade judgment that matches the judgment result of the inspector.

Further, according to the metal strip surface inspection apparatus of the present invention, in the rating judgment table used for grading, linear flaws, point flaws and planar flaws are determined with respect to the total weight value. Since the grades are individually set, it is possible to perform a grade judgment that is more consistent with the judgment result of the inspector.

[Brief description of the drawings]

FIG. 1 is a diagram schematically showing a part of a metal strip transport system equipped with an embodiment of a surface inspection apparatus according to the present invention.

FIG. 2 is a block diagram showing a processing system for processing signals from a CCD of a CCD camera of the surface inspection apparatus of FIG. 1;

FIG. 3 is a diagram for explaining binary data.

FIG. 4 is a diagram for explaining multi-value data.

FIG. 5 is a block diagram schematically showing a processing circuit of the surface inspection apparatus and a configuration related thereto.

6 is a flowchart showing an operation flow of rating judgment by the processing circuit of FIG. 5;

FIG. 7 is a diagram showing a distribution of flaws extracted from the binary data of FIG. 3;

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 metal strip 6,7 surface inspection means 14,20 CCD camera 16 CCD 26 binary data conversion means 40 signal processing circuit 41 first determination means 42 second determination means 43 flaw type / degree determination means 44 table storage means 45 Flaw weight table 46 Rating judgment table 47 Weight value memory 48 Representative flaw determination means 49 Weight value calculation means 50 Rating judgment means

Claims (2)

[Claims] An image sensor for obtaining an image of a surface of a metal strip running in a longitudinal direction, and multi-valued data for obtaining multi-valued data on a flaw existing in a flaw evaluation range using a detection signal of the image sensor. Means, binary data conversion means for obtaining binary data relating to the flaws present in the flaw evaluation range using the detection signal of the image sensor, and flaws based on the multi-value data obtained by the multi-value data conversion means. First determining means for determining the degree of each flaw present in the evaluation range; and determining the form of each flaw present in the flaw evaluation range based on the binary data obtained by the binary data converting means. Second determining means, weight value determining means for determining a flaw weight value for each flaw from a flaw weight table based on the determination results of the first and second determining means, A representative flaw determining means for selecting a flaw having a maximum weight value from the obtained weight values and determining the flaw as a representative flaw; and a weight for adding a flaw weight value of each flaw existing within the flaw evaluation range to obtain a sum. Value calculating means, and rating determining means for performing a rating from a rating determining table based on the form of the flaw determined by the representative flaw determining means and the sum of the weight values calculated by the weight value calculating means. Characteristic metal surface inspection equipment. 2. The flaw existing in the flaw evaluation range is classified into one of a point flaw, a linear flaw, and a plane flaw, and the rating of the rating judgment table is determined by the weight value calculating means. The metal strip surface inspection apparatus according to claim 1, wherein linear flaws, point flaws, and planar flaws are individually set for the total value.