TWI447606B - Method for detecting sliversof a billet - Google Patents

Description

Small steel embryo stripping detection method

The invention relates to a method for detecting small steel sheet stripping, in particular to a method for detecting whether an abnormal state of stripping or slag inclusion occurs in a small steel embryo by using a small steel embryo image.

Small steel embryos are the upstream products of rod and wire rods, while rod and wire rods are the main raw materials for the manufacture of screws, nuts, hand tools, etc. Therefore, the quality of small steel embryos will directly affect downstream products such as screw nuts and hand tools. Manufacturing quality and manufacturing costs.

Common defects in small steel embryo production line include corner crack, spongy defect, slag inclusion, burnt stripping and mechanical stripping. Among them, corner crack, sponge defect and slag inclusion are generated from the upstream of small steel embryo production line. Burnt stripping and mechanical stripping are associated with equipment anomalies on small steel embryo production lines. Therefore, if the small steel embryo is produced, the surface of the steel can be immediately detected to ensure that the small steel embryo of poor quality will not be sent downstream.

One aspect of the invention is provided in a method for detecting small steel sheet strips. The small steel embryo stripping detection method uses the image processing method to judge the surface abnormality of the small steel embryo. In addition to detecting the surface abnormality of the mechanical stripping and the slag inclusion, it can also assist the field operator to understand the abnormal condition of the equipment early. .

According to an embodiment of the present invention, in the method for detecting small steel embryo strips, first, a small steel embryo image of a small steel embryo is provided. Next, the small steel embryo shadow The binarization step is performed to obtain a binarized image. This binarized image contains a plurality of foreground objects. Then, a plurality of object area values of these foreground objects are calculated. Next, the object area values are sorted according to the size of the object area values to obtain a sequence of area values. Then, the difference between each two adjacent object area values in the sequence of area values is calculated to obtain at least one area difference of the object area values. Then, from among the at least one area difference, the maximum area difference is selected. Then, at least one larger object area value is determined from the object area values based on the maximum area difference. Next, it is determined whether the sum of the larger object area values is greater than a preset area threshold to provide a first determination result. When the first judgment result is YES, it is determined that the small steel embryo has an abnormal condition of mechanical peeling or slag inclusion.

According to another embodiment of the present invention, in the method for detecting small steel embryo strips, first, a small steel embryo image of a small steel embryo is provided. Next, the small steel embryo image is binarized to obtain a binarized image. The binarized image includes a plurality of foreground objects. Then, calculate the object outline length of the foreground object. Next, the sum of the object outline lengths is divided by the number of foreground objects to obtain an average contour length. Then, it is judged whether the average contour length is greater than the preset contour length threshold to provide the first judgment result. When the first judgment result is YES, it is determined that the small steel embryo has an abnormal condition of mechanical peeling or slag inclusion.

According to still another embodiment of the present invention, in the method for detecting small steel embryo stripping, first, a small steel embryo image of a small steel embryo is provided. Next, a binarization step is performed on the small steel embryo image to obtain a binarized image. The binarized image includes a plurality of foreground objects. Then, the Run-Length Coding (RLC) algorithm is performed on the binarized image to obtain a plurality of Run-Length numbers. This code length encoding algorithm scans the binarized image along the horizontal direction. Then, add the total number of these codes to calculate Calculate the total length of the code. Next, it is determined whether the total code length is less than a preset code length threshold to provide a determination result. When the result of this determination is YES, it is determined that the small steel embryo has an abnormal condition of mechanical peeling or slag inclusion.

According to still another embodiment of the present invention, in the method for detecting small steel embryo stripping, first, a small steel embryo image of a small steel embryo is provided. Next, a binarization step is performed on the small steel embryo image to obtain a binarized image. The binarized image includes a plurality of foreground objects. Then, the code length coding algorithm is performed on the binarized image to obtain a plurality of code lengths. This code length encoding algorithm scans the binarized image along the horizontal direction. Then, the maximum code length is determined from these code lengths. Next, it is determined whether the maximum code length is greater than a preset code length threshold to provide a determination result. When the result of this determination is YES, it is determined that the small steel embryo has an abnormal condition of mechanical peeling or slag inclusion.

It can be seen from the above description that the small steel piece peeling detection method in the embodiment of the present invention uses the object contour length, the object area value or the total code length of the foreground object to detect the surface of the small steel embryo to judge the small line. Whether the steel embryo has an abnormal state of scale, slag or mechanical peeling.

Please refer to FIG. 1a-1b at the same time, which is a schematic flow chart of a small steel stripping detection method 100 according to an embodiment of the invention. In the small steel stripping detection method 100, an image capturing step 110 is first performed to capture a surface image of a small steel blank on the line. Next, a normalization step 120 is performed to normalize the image of the small steel embryo. Then, a binarization step 130 is performed to binarize the normalized image. In this embodiment, the binarization step 130 uses u-k* σ as the binarization threshold, where u is the normalized image level. Average grayscale value; σ is the grayscale standard deviation; k is a constant (here 2.0).

Please refer to FIG. 1c-1e, which is a schematic diagram showing a binarized image of a small steel blank according to an embodiment of the present invention, wherein the part of the line represents white, and the part of the unlined part represents black. . If the surface of the small steel blank has an abnormal state of mechanical peeling, the binarized image will contain a long strip of foreground object 101, as shown in Fig. 1c. If the surface of the small steel blank has an abnormal state of the scale, the binarized image will contain a plurality of small pieces of the foreground object 101 distributed as a piecemeal, as shown in Fig. 1d. If the surface of the small steel embryo has an abnormal state of slag inclusion, the binarized image will contain a large piece of foreground object 101, as shown in Fig. 1e.

In addition, although the small steel billet detecting method 100 of the present embodiment performs binarization after normalizing the small steel embryo image, the embodiment of the present invention is not limited thereto. In other embodiments of the invention, the normalization step can also be omitted according to the needs of the user.

Returning to Figure 1a, after 130 after the binarization step, an area calculation step 140 is then performed to calculate an object area value for each of the foreground objects in the binarized image. Then, a sorting step 150 is performed to sort the object area values according to the size of the object area values to obtain a sequence of area values. Next, a difference calculation step 160 is performed to calculate a difference between each two adjacent object area values in the sequence of area values to obtain at least one area difference of the object area values. E.g. If there are M foreground objects 101 in the binarized image, the area from large to small is a ₁ , a ₂ , ..., a _M , then the area difference d ₁ , d ₂ , ..., d _{M -1} is a ₁ - a ₂ , a ₂ - a ₃ , ..., a _M-1 - a _M . Then, selection step 170, the difference in these areas, the maximum difference area is selected, for example, d _3. Next, a larger object area value determining step 180 is performed to determine a larger object area value from the object area values.

Please refer to FIG. 1f, which is a flow chart showing a larger object area value determining step 180 according to an embodiment of the present invention. In the larger object area value determining step 180, a starting area value determining step 182 is first performed to determine the starting area value from the two object area values corresponding to the maximum area difference. For example, if the maximum area difference is d ₃ = a ₃ - a ₄ , the area values a ₃ and a ₄ represent the boundary between the large object area value and the small object area value (or the boundary object area value), and the area value a ₃ , the largest of the two boundary object area values, is determined as the starting area value. Then, a decision step 184 is performed to determine a larger object area value based on the starting area value and the maximum area value. In this embodiment, the maximum area value is the largest of the object area values, such as a ₁ , and the larger object area value is included in the sequence of area values from the starting area value a ₃ to the maximum area value a ₁ . Area value, ie a ₁ , a ₂ , a ₃ .

Referring to FIG. 1b, after the larger object area value determining step 180, a determining step 190 is then performed to determine whether the sum of the larger object area values is greater than a predetermined area threshold. In the present embodiment, the preset area threshold is 80% of the total area value of the M foreground objects 101, but the embodiment of the present invention is not limited thereto. When the sum of the larger object area values is greater than 80% of the total area value, it means that the first few foreground objects in the image account for a higher percentage of the total area, and according to this, it is determined that the surface of the small steel embryo has mechanical peeling or slag inclusion abnormality. situation.

Additionally, in other embodiments of the invention, the sum of the larger object area values can be calculated using the equation: k* = argmax(d _k )

Where k* is the starting area value; p is the ratio of the sum of the larger object area values to the total area value.

It can be seen from the above description that the present embodiment uses the area of the foreground object to determine whether the small steel slag has an abnormal state of slag inclusion or mechanical peeling. This is because the binarized image of the scale contains a large number of foreground objects, so The percentage of the first few objects in the total area will be smaller than other abnormal conditions (such as mechanical stripping), as shown in the table in Figure 1g.

Please refer to FIG. 2a-2b, which is a schematic flow chart of a method for detecting small steel sheet stripping according to an embodiment of the present invention. The small steel sheet stripping detection method 200 is similar to the small steel sheet stripping method 100, but the difference is that the small sheet metal stripping detecting method 200 further includes a slag determining step 210. It can be seen from the above description that the small steel sheet stripping detecting method 100 of the embodiment of the present invention can determine that the surface of the small steel blank may have abnormal conditions of mechanical stripping or slag inclusion, but it cannot be further judged that Which of the small steel stripping detection methods 200 can utilize the slag determination step 210 to further determine whether the surface of the small steel blank is an abnormal condition of slag inclusion.

Please refer to FIG. 2c, which is a schematic flow chart of the slag inclusion determining step 210 according to an embodiment of the present invention. In the slag determination step 210, the code length encoding step 211 is first performed to perform a Run-Length Coding (RLC) algorithm on the binarized image to obtain a plurality of Run-Length numbers. In the present embodiment, the code length encoding step 211 scans the binarized image along the vertical direction as shown in Fig. 2d. During the scanning process, each scanning line L corresponds to a code length, this The number of code lengths is related to the number of times the scan line L crosses the boundary of the foreground object. After the code length encoding step 211, a summing step 212 is then performed to add up the code lengths to obtain a total code length. Then, a determining step 213 is performed to determine whether the total code length is greater than a preset number of code length thresholds. When the total length of the code length is greater than the preset threshold of the code length, it represents the abnormal condition of the slag inclusion in the small steel embryo, and vice versa, the abnormal condition of the mechanical peeling of the small steel embryo. In the present embodiment, the preset code length threshold is 20 (one), but the embodiment of the present invention is not limited thereto.

Please refer to FIG. 3 , which is a schematic flow chart of a method for detecting a small steel stripping strip 300 according to an embodiment of the present invention. The small steel sheet stripping detection method 300 is similar to the small steel sheet stripping method 100, but the difference is that the small steel sheet stripping method 300 uses the contour length of the foreground object to determine whether the small steel embryo has An abnormal state of slag inclusion or mechanical peeling occurs.

In the small steel stripping detection method 300, an image capturing step 110, a normalizing step 120, and a binarization step 130 are first performed to obtain a binarized image of a small steel blank. Then, a contour length calculation step 340 is performed to calculate the object contour length of all foreground objects. In the present embodiment, the object profile length of the foreground object is obtained using the Pearson's theorem, but the embodiment of the present invention is not limited thereto. Next, an averaging calculation step 350 is performed to divide the sum of the lengths of all foreground object contours by the number of foreground objects to obtain an average contour length. Then, a determining step 360 is performed to determine whether the average contour length is greater than a preset contour length threshold, and the determination result is provided. In the present embodiment, the preset contour length threshold is 40 (pixel units), but the embodiment of the present invention is not limited thereto. When the judgment result is YES, it is determined that the small steel embryo has an abnormal condition of mechanical peeling or slag inclusion.

It can be seen from the above description that the present embodiment uses the average contour length of the foreground object to determine whether the small steel blank has an abnormal state of slag inclusion or mechanical stripping, because the binarized image of the scale contains a large number of foreground objects. The average profile length will be smaller than other abnormal conditions (such as mechanical stripping). In addition, the small steel stripping detection method 300 may further include the foregoing slag determination step 210 to further determine whether the small steel embryo is abnormally slag or mechanically peeled.

Please refer to FIG. 4 , which is a schematic flow chart of a method for detecting a small steel stripping strip 400 according to an embodiment of the present invention. The small steel sheet stripping detection method 400 is similar to the small steel sheet stripping method 100, but the difference is that the small steel sheet stripping method 300 uses the number of code lengths of the binarized image to judge the small steel. Whether the embryo has an abnormal state of slag inclusion or mechanical peeling.

In the small steel stripping detection method 400, the image capturing step 110, the normalizing step 120, and the binarization step 130 are first performed to obtain a binarized image of the small steel blank. Then, a code length encoding step 411 is performed to perform a code length encoding algorithm on the binarized image to obtain a plurality of code length numbers. In the present embodiment, the code length encoding step 411 scans the binarized image along the horizontal direction as shown in FIG. 4a. During the scanning process, each scanning line L corresponds to a code length number, which is related to the number of times the scanning line L crosses the boundary of the foreground object. After the code length encoding step 411, a summing step 412 is then performed to add the total number of code lengths to obtain a total code length. Then, a determining step 413 is performed to determine whether the total code length is less than a preset number of code length thresholds. When the total length of the code length is less than the preset threshold of the number of code lengths, it means that the small steel blank has abnormal conditions of mechanical peeling or slag inclusion. In this embodiment, the preset code length threshold is 70 (pieces), However, embodiments of the invention are not limited thereto.

It can be seen from the above description that the present embodiment uses the number of code lengths of the binarized image to determine whether the small steel embryo has an abnormal state of slag inclusion or mechanical stripping, because the binarized image of the scale contains a large amount of foreground. For objects, when the horizontal scanning line sweeps over the binary image of the scale, the total length of the code will be much larger than the total length of other abnormal states (such as mechanical stripping), as shown in the table in Figure 4b. In addition, the small steel stripping detection method 400 may further include the foregoing slag determination step 210 to further determine whether the abnormal condition of the small steel blank is slag or mechanical peeling.

Please refer to FIG. 5 , which is a schematic flow chart of a small steel stripping detection method 500 according to an embodiment of the present invention. The small steel sheet stripping detection method 500 is similar to the small steel sheet stripping method 400, but the difference is that the small steel sheet stripping detection method 500 uses the code length of the binarized image to judge the small steel. Whether the embryo has an abnormal state of slag inclusion or mechanical peeling.

In the small steel stripping detection method 500, an image capturing step 110, a normalizing step 120, and a binarization step 130 are first performed to obtain a binarized image of a small steel blank. Then, a code length encoding step 511 is performed to perform a code length encoding algorithm on the binarized image to obtain a plurality of code lengths. In this embodiment, the code length encoding step 511 scans the binarized image along the horizontal direction to obtain a plurality of code lengths, and the length of each code length represents the distance swept by the scan line before passing the edge of the object. (or pixel). Next, a decision step 512 is performed to determine the maximum code length from these code lengths. Then, a determining step 513 is performed to determine whether the maximum code length is greater than a preset code length threshold. When the maximum code length is greater than the preset code length threshold, it represents the abnormal condition of mechanical stripping or slag inclusion in the small steel blank. In this embodiment, The preset code length threshold is 30 (pixel units), but embodiments of the present invention are not limited thereto.

It can be seen from the above description that the present embodiment uses the code length of the binarized image to determine whether the small steel embryo has an abnormal state of slag inclusion or mechanical peeling, because the binarized image of the scale contains a large amount of foreground. When the horizontal scan line sweeps over the binary image of the scale, the code length is usually no longer than the code length of other abnormal states (such as mechanical peeling), as shown in the table in Figure 5a. In addition, the small steel stripping detection method 500 may further include the foregoing slag determination step 210 to further determine whether the small steel embryo is abnormally slag or mechanically peeled.

While the invention has been described above in terms of several embodiments, it is not intended to limit the scope of the invention, and the invention may be practiced in various embodiments without departing from the spirit and scope of the invention. The scope of protection of the present invention is defined by the scope of the appended claims.

100‧‧‧Small steel embryo stripping detection method

101‧‧‧ Prospect objects

110‧‧‧Image capture steps

120‧‧‧ formalization steps

130‧‧‧ Binarization steps

140‧‧‧ Area calculation steps

150‧‧‧Sorting steps

160‧‧‧Difference calculation steps

170‧‧‧Selection steps

180‧‧‧ Large object area value determination steps

182‧‧‧ starting area value decision step

184‧‧‧Decision steps

190‧‧‧ Judgment steps

200‧‧‧Small steel embryo stripping detection method

210‧‧‧Slag slag determination step

211‧‧‧ code length coding step

212‧‧‧Additional steps

213‧‧‧ Judgment steps

300‧‧‧Small steel embryo stripping detection method

340‧‧‧Contour length calculation steps

350‧‧‧Average calculation steps

360‧‧‧ Judgment steps

400‧‧‧Small steel embryo stripping detection method

411‧‧‧ code length coding step

412‧‧‧Additional steps

413‧‧‧ judgment steps

500‧‧‧Small steel embryo stripping detection method

511‧‧‧ code length coding step

512‧‧‧Decision steps

513‧‧‧ judgment steps

L‧‧‧ scan line

The above and other objects, features and advantages of the present invention will become more <RTIgt; A schematic flow chart of a method for detecting a small steel sheet strip according to an embodiment of the present invention.

The 1c-1e diagram is a schematic diagram showing a binarized image of a small steel embryo according to an embodiment of the present invention.

Figure 1f is a diagram showing the larger object area value according to an embodiment of the present invention. A schematic diagram of the process of the steps.

The 1g figure shows the percentage of the total area of several objects before the binarized image of various abnormal states according to an embodiment of the present invention.

2a-2b is a schematic flow chart showing a method for detecting small steel sheet stripping according to an embodiment of the present invention.

2c is a schematic flow chart showing the slag determination step according to an embodiment of the present invention.

FIG. 2d is a schematic diagram showing scanning along the vertical direction according to the code length encoding step of the embodiment of the present invention.

FIG. 3 is a schematic flow chart showing a method for detecting a small steel sheet peeling strip according to an embodiment of the present invention.

FIG. 4 is a schematic flow chart showing a method for detecting a small steel piece peeling strip according to an embodiment of the present invention.

Figure 4a is a schematic diagram showing scanning along the horizontal direction in accordance with the code length encoding step of the embodiment of the present invention.

Figure 4b is a diagram showing the total code length of the binarized image of various abnormal states according to an embodiment of the present invention.

FIG. 5 is a schematic flow chart showing a method for detecting a small steel piece peeling strip according to an embodiment of the present invention.

Figure 5a is a diagram showing the maximum code length of a binarized image of various abnormal states according to an embodiment of the present invention.

300‧‧‧Small steel embryo stripping detection method

110‧‧‧Image capture steps

120‧‧‧ formalization steps

130‧‧‧ Binarization steps

340‧‧‧Contour length calculation steps

350‧‧‧Average calculation steps

360‧‧‧ Judgment steps

Claims (10)

A method for detecting small steel embryo stripping comprises: providing a small steel embryo image of a small steel embryo; performing a binarization step on the small steel embryo image to obtain a binarized image, wherein the binarization The image includes a plurality of foreground objects; calculating a plurality of object area values of the foreground objects; sorting the object area values according to the size of the object area values to obtain an area value sequence; calculating each of the area value sequences a difference between the two adjacent object area values to obtain at least one area difference value of the object area values; and from the at least one area difference value, selecting a maximum area difference value; according to the maximum area difference value Determining at least one larger object area value from the object area values; determining whether the sum of the at least one larger object area values is greater than a predetermined area threshold to provide a first determination result; and when the first When the judgment result is YES, it is determined that the small steel embryo has an abnormal condition of mechanical peeling or slag inclusion. The small steel sheet stripping detecting method according to claim 1, wherein the preset area threshold is 80% of the sum of the area values of the objects. The method for detecting a small steel piece peeling strip according to claim 1, wherein the step of determining the at least one larger object area value comprises: determining from the two boundary object area values corresponding to the maximum area difference value. a starting area value, wherein a difference between the two boundary object area values is the maximum area difference, and the starting area value is the highest of the two boundary object area values And determining, according to the initial area value and a maximum area value, the at least one larger object area value, wherein the maximum area value is the largest of the object area values, the at least one larger object area value The area value included in the sequence of area values from the starting area value to the maximum area value. The method for detecting a small steel stripping strip according to the first aspect of the patent application includes: performing a Run-Length Coding (RLC) algorithm on the binarized image to obtain a plurality of code lengths. a (Run-Length) number, wherein the code length encoding algorithm scans the binarized image along a vertical direction; summing the number of code lengths to obtain a total code length; determining the code length Whether the total amount is greater than a predetermined code length threshold to provide a second determination result; and when the first determination result is yes and the second determination result is no, determining that the small steel embryo is mechanically peeled off An abnormal condition, and when both the first judgment result and the second result are YES, it is determined that the small steel embryo has an abnormal condition of slag inclusion. The method for detecting a small steel sheet peeling strip according to claim 1, wherein the preset code length threshold is 20 pixel units. A small steel embryo stripping detection method comprises: providing a small steel embryo image of a small steel embryo; performing a binarization step on the small steel embryo image to obtain a binarization An image, wherein the binarized image comprises a plurality of foreground objects; calculating a plurality of object contour lengths of the foreground objects; dividing the sum of the contour lengths of the objects by the number of the foreground objects to obtain an average contour length Determining whether the average contour length is greater than a predetermined contour length threshold to provide a first determination result; and when the first determination result is YES, determining an abnormal condition of the mechanical peeling or slag inclusion of the small steel blank . The small steel piece peeling detection method according to claim 6, wherein the preset contour length threshold is 40 pixel units. The method for detecting small steel stripping strips according to item 6 of the patent application scope further comprises: performing a Run-Length Coding (RLC) algorithm on the binarized image to obtain a plurality of code lengths. (Run-Length), wherein the code length encoding algorithm scans the binarized image along a vertical direction; summing the number of code lengths to calculate a total code length; determining the code Whether the total amount is greater than a predetermined code length threshold to provide a second determination result; and when the first determination result is yes and the second determination result is no, determining that the small steel embryo is mechanically peeled off The abnormal condition, and when both the first judgment result and the second result are YES, it is determined that the small steel embryo has an abnormal condition of slag inclusion. A method for detecting small steel embryo stripping comprises: providing a small steel embryo image of a small steel embryo; performing a binarization step on the small steel embryo image to obtain a binarized image, wherein the binarization The image includes a plurality of foreground objects; a Run-Length Coding (RLC) algorithm is performed on the binarized image to obtain a plurality of code lengths (Run-Length), wherein the code length encoding algorithm is Scanning the binarized image along the horizontal direction; summing the number of code lengths to calculate a total code length; determining whether the total code length is less than a preset code length threshold, Providing a judgment result; and when the judgment result is YES, determining an abnormal condition of the mechanical peeling or slag inclusion of the small steel embryo. A method for detecting small steel embryo stripping comprises: providing a small steel embryo image of a small steel embryo; performing a binarization step on the small steel embryo image to obtain a binarized image, wherein the binarization The image includes a plurality of foreground objects; performing a code length encoding algorithm on the binarized image to obtain a plurality of code lengths, wherein the code length encoding algorithm scans the binarized image along a horizontal direction; Determining a maximum code length from the code lengths; determining whether the maximum code length is greater than a predetermined code length threshold to provide a determination result; and when the determination result is yes, determining the small steel embryo appearance mechanism Abnormal condition of stripping or slag inclusion.