JPH07111406B2 - Surface slab surface defect inspection method - 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 a surface defect of a steel slab, and more particularly to a method for inspecting a surface defect using sparks caused by inclusions on the surface layer generated during scarfing of a steel slab.

[0002]

2. Description of the Related Art As a method for detecting surface defects of non-metallic inclusions in a steel slab, Japanese Patent Laid-Open No. 154346/1984 describes a method of directly processing an analog video signal by a simple threshold judgment or the like. Has been done. However, from an irregularly shaped melting bath that changes moment by moment, the emission time is 0.1
It is extremely difficult to accurately detect a spark caused by inclusions, which is an instantaneous phenomenon of about 0.5 seconds, by distinguishing it from the contour of the melting bath.

[0003]

SUMMARY OF THE INVENTION It is an object of the present invention to provide an inspection method capable of accurately detecting the degree of presence of surface layer inclusions based on the information of the melting bath that changes from moment to moment.

[0004]

DISCLOSURE OF THE INVENTION The present invention has been made in order to advantageously achieve the above object, and its gist is to use a melting bath for billet surface scarfing ablation as an image processing region, The melting bath is imaged by a TV camera during ablation at a minute time interval shorter than the spark emission time, the captured images are sequentially captured in a plurality of image memories, and the images are captured in the image memory at imaging time intervals. Sequentially updating the fusion bath image of and the fusion bath image of an appropriate finite time before, and performing image comparison calculation between the images obtained in this way and within a minute time interval which is the imaging time, and between both images The size of sparks, the position of sparks, and the number of sparks are calculated from the difference in the amount of light, and the results are sequentially stored in another storage device.After the scarfing and ablation is completed, the spark information stored in the storage device is used as the basis. Then sparks and noise Discriminate in a surface defect inspection method of the steel strip, characterized by determining the distribution state of the inclusions.

[0005]

[Operation] The present invention will be described below with reference to the drawings.

FIG. 4 shows a basic image of the melting bath and its spread portion by scarfing. FIG. 4 (a) is an image of a melting bath without sparks, in which the melting bath 11 is generated by fusing,
Sparks appear here when inclusions are present. Sparks do not appear in the enlarged portion 12 of the melting bath due to the pressure of scarfing, and this is a noise region.

Therefore, when a stable melting bath 11 is formed immediately after the start of scarfing and ablation, a mask 13 shown in FIG. 4B is determined as an image processing area, and this is set as a sensitive area, and the other areas are dead zones. The area 14 is set, and only the sensitive zone area is set as the image processing target.

FIG. 4 (c) shows a melting bath 11 which is a sensitive zone region.
4D shows an image in which the spark 15 appears, and FIG. 4D shows an image in which no spark appears. In these images, as will be described later, with respect to the spark 15, which is an instantaneous phenomenon with a light emission time of about 0.1 to 0.5 seconds, a minute time interval shorter than that (for example, 1
Every 30 seconds) continuously input to the image processing apparatus.

FIG. 1 shows this image capturing and spark detection method. First, at least two image memories are prepared, and the image memory 1 is updated every 1/30 seconds in the order of the input fused bath images. Similarly, the image memory 2 is also the image memory 1.
M / 30 seconds (m is 1 or more, preferably 100
(Below) Update to the delayed melting bath image 5.

In this way, the image capture can be performed by, for example, 1 /
By executing the process every 30 seconds, the image memories 1 and 2 in progress of the fusing are sequentially updated with the current fusing bath image and that of m / 30 seconds ago. Spark measurement can be continuously performed by performing image comparison calculation processing on the melt bath images input to the image memories 1 and 2 within 1/30 seconds, which is the melt bath image update time.

Then, as shown in FIG.
If a appears, a light amount difference of sparks is generated between the melting bath images of the image memories 1 and 2, and the image processing is performed using the light amount difference, and the spark position, the time of appearance, the spark area, and the number of sparks are sequentially displayed as spark data. It is held in the storage device 3. The above processing is performed over the entire area of the fusing, and after fusing, the spark discriminating unit 4 discriminates sparks and noises based on the data held in the spark data storage device 3, thereby causing an irregular melting that changes from moment to moment. It is possible to accurately measure instantaneous sparks from the bus and measure the distribution of surface inclusions.

Further, in order to discriminate the spark more accurately, it is sufficient to increase the number of fused bath images to be simultaneously compared.

[0013]

FIG. 2 shows an example of an apparatus suitable for carrying out the method of the present invention. A traveling platform 25 is provided in parallel with the traveling direction of the steel slab 21, and a traveling carriage 24 is mounted on the traveling platform 25 so as to be movable in the longitudinal direction of the steel slab 21. Traveling car 24 through the fusing crater 23
And is movably engaged in the width direction of the steel piece 21. The TV camera 22 is arranged above the fusing crater 23, and only the sensitive zone region of the melting bath formed by the fusing crater 23 is imaged. The captured image of the television camera 22 is sent to the image processing device 26 via the signal cable 27.

The fusing crater 23 is capable of fusing the entire surface in the longitudinal direction and the width direction of the steel slab 21. During the fusing, the television camera 22 captures images of the molten bath at 1/30 second intervals, and the captured image is processed by the image processing device 26. The operation of loading the image memories 1 and 2 (FIG. 1) therein is repeated. Image comparison calculation processing is performed between the two image memories 1 and 2 that are updated every 1/30 second for spark measurement. If a spark appears, there will be a difference in the light amount of the spark from the image of the molten bath a minute before, and the singular point (spark 15) on the image can be extracted in the spark data storage device 3, and the spark appearance position, The spark area and number can be detected.

The above operation is continuously performed for one steel piece, and the data in the spark data storage device 3 after the completion of the fusing is discriminated by the spark discriminating section 4 into sparks and noise due to inclusions, so that the steel piece 21 can be obtained. It is possible to accurately detect surface flaws caused by surface layer inclusions.

FIG. 3 shows one piece of low carbon aluminum killed steel (continuous cast material) (width 1300 mm × length 9000 mm).
Fig. 3 (a) shows the result of the inspection at a fusing speed of 12 mpm. The rectangular frame in Fig. 3 (a) shows the fusing region, and the white circles therein show the sparks caused by inclusions. FIG. 3B shows the number of sparks generated in the longitudinal direction of the billet.

[0017]

As described above in detail, the present invention detects the occurrence of sparks from the difference in the light amount of the picked-up image. Therefore, compared to the conventional analog signal evaluation / judgment method, the melting bath The quality of the billet can be determined quickly and accurately because it can be reliably detected in distinction from the contour portion. Further, for example, by comparing the detection result with the billet production conditions, useful information for improving the billet production conditions can be obtained, and the production of defective billets can be prevented in advance.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a spark detection method according to the present invention (an example of inputting with a delay of m = 3).

FIG. 2 is an explanatory diagram showing an example of an apparatus suitable for implementing the present invention.

FIG. 3 is a view showing a result of an inspection in which a surface defect inspection of a steel piece is performed by the method of the present invention, and a memory of a spark appearance position, a spark area, and the number of sparks is discriminated from noise and only sparks are displayed.

FIG. 4 is an explanatory diagram showing a melting bath and an image processing area.

[Explanation of symbols]

 1, 2 Image memory 3 Spark data storage device 4 Spark discriminator 5 Melting bath image group 11 Melting bath 12 Expanded portion of melting bath 13 Mask 14 Dead zone area 15 Spark 21 Steel piece 22 TV camera 23 Welding crater 24 Traveling vehicle 25 Running Frame 26 Image processing device 27 Signal cable

Claims (1)

[Claims] 1. A melting bath for slab surface scarfing welding is used as an image processing area, and the melting bath is imaged by a TV camera during ablation at a minute time interval shorter than a spark emission time, and the captured images are sequentially captured. The image is taken into the image memory, the image is taken into the image memory at the imaging time interval, and the current melt bath image and the melt bath image before an appropriate finite time are sequentially updated, and between the images thus obtained. In addition, image comparisons are calculated within a minute time interval that is the imaging time, the size of sparks, the position of sparks, and the number of sparks are calculated from the difference in the amount of light between both images, and the results are sequentially stored in another storage device. A surface defect inspection of a steel piece, which is characterized by holding, and after discriminating the scarfing, discriminating between a spark and a noise based on spark information held in the storage device and determining a distribution state of inclusions. Method.