JPH0344259B2 - - Google Patents

Description

[Detailed Description of the Invention] (Technical Field) Regarding a method for detecting surface defects in hot metal materials, the technical content described in this specification is to detect defects generated on the surface of the material when detecting surface defects using an optical active method. It is related to improving defect detection accuracy by effectively removing oxide scale and adding innovation to the method of projecting and receiving external light.

(Prior art and its problems) Conventionally, typical methods for detecting surface defects in hot metal materials include: (1) detecting surface defects from processed signals of self-luminescence, such as infrared light, emitted from hot metal materials; (2) the so-called optical active method, in which the sample is irradiated with light from the outside and surface defects are detected from the processed signal of the reflected light; and (3) the two methods described above are used. A known method is to receive reflected light and self-emitted light using the same optical system, separate them, and detect different surface defects from the separately processed signals.

However, all of the conventional methods listed above were affected by the adverse effects of iron oxide generated on the surface of hot metal materials, so until now when detecting surface defects, it was necessary to perform descaling treatment beforehand. was considered essential.

By the way, as such a descaling treatment method, for example, a method using high pressure water has been proposed in Japanese Patent Application Laid-Open No. 100281/1983. However, depending on the material, water spray may induce cracks in some materials, so there is a large risk of creating new defects, and spraying water may also cause a drop in the temperature of the material. This was not desirable from the standpoint of energy conservation.

In addition, JP-A-53-144791 proposed a method of descaling by spraying metal particles onto the slab surface using high-pressure gas. Since the spray is sprayed onto the surface of the slab, the appearance of the surface changes significantly and it also scatters dust, which often impairs the working environment in optical detection methods.

(History of the elucidation of the solution) In view of the current situation, the inventors have conducted intensive research on a new surface defect detection method and have found that (1) As a descaling process, a shot particle is shot onto the surface of the material to be inspected. (2) Although the shot particles leave some marks on the surface of the material to be tested, the negative effects caused by such marks can be avoided by adjusting the light emitting and receiving methods. They discovered that the problem could be effectively solved by doing so, and completed this invention.

(Structure of the Invention) That is, the present invention relates to an improvement of the optical active method, and when detecting surface defects of a moving hot metal material, shot particles are projected onto the detected part of the hot metal material at a density of 36. ~50Kg/ ^m2 , and then immediately irradiate the detected part with light from at least one of the directions parallel to or perpendicular to the running direction of the hot metal material. The reflected light from the detection part is received by a flaw detection camera and photoelectrically converted.Then, the electrical signal obtained is corrected for the influence of shot particle traces, and then binarized to generate a defect signal caused by the surface defect. This is a method for detecting surface defects in hot metal materials, which is characterized by extracting only a defect signal and determining a surface defect based on this defect signal.

This invention will be specifically explained below.

FIG. 1 schematically shows a surface defect detection device suitable for carrying out the present invention, in which numeral 1 is a descaling device, 2 is a flaw detection camera, and 3 is a signal processing device. The present invention uses such a detection device to perform a descaling process by shot blasting, a flaw detection process for observing the descaled surface of the test material, and processing the obtained signals to extract and determine surface defects. It consists of a defect extraction and judgment process.

First, let me explain the descaling process.
The shot blasting force should not be too strong or too weak. This is because if the shot blasting force is too strong, the noise in the background signal will increase and there is a high possibility that defects will be overlooked.On the other hand, if the shot blasting force is too weak, the scale will remain and noise similar to the defect signal will be generated during observation. This is because a signal is generated, leading to an increase in false detections.

Figure 2 shows the results of investigating the relationship between shot grain projection density and scale residual area ratio, and Figure 3 also shows the relationship between projection density and noise ratio, that is, the increase in background signal noise during observation. The results of investigating the relationships are shown below.

According to Figure 2, when the projection density is less than 36Kg/ ^m2 , the residual amount of scale increases, while as shown in Figure 3, when the projection density exceeds 50Kg/ ^m2 , the noise ratio increases. In this invention, the shot density is limited to a range of 36 to 50 kg/m ² .

The specimen material that has been descaled in this way is
The material to be tested is sent to the subsequent flaw detection process, but since the material to be tested is at a high temperature, oxide scale will immediately form on the surface even if it is descaled. FIG. 4 shows the difference in S/N ratio before and after descaling.
According to FIG. 4, it can be seen that descaling improves the S/N ratio by 4 to 5 times, and removes scale, creating an environment that is easy to detect.

Shot blasting also creates slight irregularities on the surface of the slab. In order to eliminate the influence of this unevenness, a low-pass filter was used to reduce noise components. FIG. 5 shows the relationship between the S/N ratio before and after low-pass filtering, and it can be seen that the S/N ratio has been improved. Therefore, it is necessary to observe the detected part immediately after removing the scale, and for this purpose, it is desirable to install the flaw detection camera as close to the descaling device as possible.

Figure 6 shows the results of investigating the relationship between the projection angle of external light and the noise ratio, and Figure 7 also shows the relationship between the projection angle and the S/N ratio (defect signal amount relative to the noise signal amount). The results of investigating the relationships are shown below.

As is clear from both figures, the optimal projection angle is 50° to 60°.

Next, Figure 8 shows that not only vertical cracks and horizontal cracks but also baldness, pockmarks, and blowholes are targeted.The light projection angle is kept constant at 60 degrees, and the light projection direction is set in the direction of travel of the material to be inspected. The following is a summary of the results of an investigation into the relationship between the light projection direction and the S/N ratio when flaw detection was performed with various changes in the direction. The light projection direction was set at 0° perpendicular to the running direction of the test material, and at 90° parallel to the running direction.

As is clear from the figure, it is optimal to project light in two directions perpendicular to 0° and 90°. In the case of projecting light from only one direction, it is preferable to project the light from one of the directions parallel to or perpendicular to the traveling direction of the test material. From an industrial standpoint, any one of the light projection methods may be selected in consideration of cost/performance.

The reflected light from the detected part obtained by irradiating external light in this way is received by a flaw detection camera, photoelectrically converted, and then sent to the defect extraction and determination process, where it is binarized and Surface defects are determined based on the results, but since traces of shot grains are slightly shiny, it is necessary to perform preprocessing as shown in Figure 9 prior to binarization processing to correct the electrical signal. is essential.

That is, the photoelectrically converted electrical signal received by the flaw detection camera is subjected to blanking removal and amplitude correction. This is done in order to ensure the S/N ratio in the portion where the amount of light decreases when some light projection unevenness occurs, and to facilitate binarization. Next, in order to perform the subsequent gradation conversion, average value correction is performed to bring the background signal to a constant level, noise is removed, and then gradation conversion is performed. By performing such gradation conversion, the signal of the defective portion is emphasized to make it easier to perform a binary comparison, and then a binarization process is performed, and a surface defect is determined based on the result.

FIGS. 10a and 10b show a comparison of electrical signals after photoelectric conversion when no descaling is performed and when descaling is performed at a projection density of 40 Kg/m ² .

As is clear from the figure, the noise signal is significantly reduced by performing the descaling process.

Next, FIG. 11 shows the transition of the signal shape in the angular stages up to the binarization processing of the electrical signal after photoelectric conversion obtained in this way, in the order of processing.

It can be seen that by applying the signal processing according to the present invention, only defects originating from surface defects are effectively extracted.

(Effects of the Invention) Thus, according to the present invention, when detecting surface defects using the optical active method, it is possible to effectively eliminate the negative effects caused by oxide scale that have been conventionally attempted.
A marked improvement in detection accuracy can be achieved.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram of a surface defect detection device suitable for carrying out the present invention, FIG. 2 is a graph showing the relationship between shot particle projection density and scale remaining area ratio, and FIG. , also a graph showing the relationship between shot grain projection density and noise ratio, Figure 4 is a graph showing the difference in S/N ratio before and after descaling, and Figure 5 is a graph showing the S/N ratio before and after low-pass filtering. Figure 6 is a graph showing the relationship between the projection angle of external light and the noise ratio, and Figure 7 is a graph showing the relationship between the projection angle of external light and S/N.
Figure 8 is a graph showing the relationship between the projection direction of external light and the S/N ratio. Figure 9 is a graph showing the relationship between the direction of external light projection and the S/N ratio. Figure 9 is the pre-processing prior to the binarization process of the photoelectric conversion signal. The flowchart shown in Figure 10a and b is as follows:
Figure 11 is a diagram comparing electrical signals after photoelectric conversion without and with descaling processing, and shows the signal from the electrical signal immediately after photoelectric conversion to the binarized signal. FIG. 3 is a diagram showing the transition of a waveform in the order of processing.

Claims (1)

[Claims] 1. When detecting surface defects of a moving hot metal material, shot particles are projected onto the detected part of the hot metal material under conditions of a projection density of 36 to 50 kg/ ^m2 , and then the hot metal material is immediately removed. The detected part is irradiated with light from at least one of the directions parallel or perpendicular to the running direction of the material, and the reflected light from the detected part is received by a flaw detection camera and photoelectrically converted, and then After correcting the influence of the shot grain traces on the obtained electrical signal, it is binarized to extract only the defect signal caused by the surface defect, and the surface defect is determined based on this defect signal. A method for detecting surface defects in hot metal materials.