JPS60196609A - Configuration detecting apparatus - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary of the Invention] The present invention relates to a shape detection device that detects deformation of the leading and trailing ends of a steel material that occurs during rolling, for example, in a hot rolling process of steel.

[Prior art]

The deformation of the tip of the steel material that occurs during hot rolling of steel applies an uneven load to the rolling rolls in subsequent stages, causing damage to the rolls and the production of defective material.

In order to prevent this, conventionally, an operator visually observes the material and operates a cutting machine to cut the material at an appropriate location.

Therefore, the deformation of the tip of the steel material is automatically detected, the deformed part is cut to the minimum extent necessary, and the material is fed to the next roll, eliminating the effects of uneven load and preventing damage to the roll. In the hope of preventing this, the following shape detection device has been proposed as part number 111847-85469.

FIG. 1 shows a block diagram of a conventional device.

The width of a red-hot steel plate 1 flowing through a hot rolling process is measured by an optical system using a lens 2 and a plurality of photoelectric elements 6 arranged in a line to measure the presence or absence of a steel plate on a measurement line 3a. light 7d
The element 6 generates an output proportional to the temperature of the steel plate, and this output is amplified by the amplifier 4 and sent to the analog-to-digital conversion circuit (
(Me) 5 converts each output of each photoelectric element 6 into a digital signal.

That is, the signal of the portion of the photoelectric element O where the image of the steel plate 1 is formed has a logical value of 111, and the signal of the portion of the photoelectric element O where no image of the steel plate 1 is present has a logical value of 101. A logic value of 111 is applied to the width measurement circuit 6, which consists of a counter circuit, for the output of the A/D 5.
By counting the number of outputs, a measurement value corresponding to the width of the steel plate 1 is obtained.

The width measurement value of the steel plate 1 is compared with the width reference value W given from the input terminal 7 by the comparison circuit 8, and when the width of the steel plate 1 is within a certain tolerance value smaller than W, the comparison circuit 8 A cutting signal is given to the cutting machine control device 9, and the tip of the steel plate 1 is cut.

That is, as shown in Fig. 2, if the swing width in the center of the steel plate 1 is W, the position of the plate width of kW (k<1) shown by the chain line at the tip of the steel plate 1 can be detected and cut. .

Conventional equipment is configured as described above, so the steel plate image may have uneven brightness due to scale or water stains generated on the steel plate, or light may not be visible even outside the steel plate image due to light scattering from the background or fine particles in the air. In cases where the image becomes bright, there is a drawback that errors occur when digitizing the steel plate image.

In order to eliminate these drawbacks of the conventional ones, we quantized the analog steel plate image signal with multiple values and stored it in memory.
By spatially differentiating this quantized signal to obtain a steel plate signal at a differential level, and then performing processing such as binarization and peak hold, it can be used as a shape detection device that does not cause errors due to scale and water 9 signals and scattering signals. The one shown in FIG. 8 can be considered. In the figure, the components from the red-hot steel plate 1 to the cutting machine control device 9 correspond to those described in connection with the shape detection device in FIG. 10 is a quantization circuit that quantizes a steel plate image signal obtained in an analog manner with multiple values; 11 is a memory for storing the quantized steel plate image signal; 12 is a circuit for reading out the steel plate image signal stored in the memory 11; A differentiation circuit 14 differentiates the differentiated steel plate image signal with a fixed threshold set in advance.
16 is a peak hold circuit that peak-holds the binarized steel plate Monk in a specific direction; 15 is a drive signal P that is generated every time the steel plate 1 moves a certain distance in the detection area; This is a scanning circuit that receives the signal, scans the photoelectric element group 6a, and sends out an output.

In the apparatus configured as shown in FIG. Each time Vc moves in the direction of the distance arrow, it is scanned and read out once.

This video signal is converted into a digital signal having multiple quantization levels by the quantization circuit 10 via the amplifier 4, and is stored in the memory 11 according to the above-mentioned scanning order. This operation is performed for the number of scans M determined from the detection field of view and scan interval set in advance, and as a result, a video image as shown in FIG. 4 (al) is obtained in the memory 11, for example.

However, as mentioned in the drawbacks of the conventional device, scale or water that is a disturbance component (9 (a)) and scattering (b)
For example, the scanning line part (c)
If the video signal in FIG. 4 (bl, fixed threshold value), (e), etc. is binarized, the error will be large.

In this apparatus, a differentiation circuit 12 two-dimensionally differentiates such a warm level video image and converts it into a steep level of signal change, thereby obtaining a steepness signal shown in FIG. 4 (cl). In other words, as is clear from the video signal shown in Fig. 4 fbl, the change in the signal at the edge part of the steel plate 1 is large and the change in the scattering part is small, so it is easy to distinguish between them. As shown, the scale or water level signal has a wide range of steepness levels that make it impossible to distinguish between the two mentioned above, but these are different from steel plate 1.
It only occurs within. Next, this steepness signal is binarized by the binarization circuit 14 at a certain threshold level (to), and then the peak hold circuit 16 peak-holds this binarized signal, so that the signal as shown in FIG. It is possible to extract the width signal of a steel plate that is not affected by scale or water stains.

The operation of generating a cutting signal based on this width signal is similar to that of the conventional device.

In the above device, the binarization circuit 14 binarizes the steepness signal using a fixed threshold value, but if the temperature change of the steel plate 1 is large, it may be difficult to set the threshold value.

FIG. 5 is a block diagram of a device that can be considered to solve this problem. First, a peak hold circuit 16 performs peak hold processing on the steepened signal processed by the differentiating circuit 12. As a result, the video signal shown as fat in FIG. 6 is partially stored in the memory 11 as a peak-held signal as shown in FIG. The direction of is set toward the center of the steel plate as shown by the arrows (to) and (su) in FIG. 14, calculate the concentration histogram. Figure 6 (cl is an example of the concentration histogram signal obtained in this way, with the steel plate part 4 and the scattering part) can be seen as the peak of the histogram signal, so these two 2 with the 1st shift of the valley part v1 between the peaks as the threshold
(d) z-value conversion (only the width signal of the steel plate can be extracted from the number i. According to this embodiment, a floating z-value conversion method is used in which a threshold value V1 is set according to the temperature of the steel plate 1. Therefore, the steel plate image can be extracted stably.The cutting signal is obtained from the width signal in the same way as the conventional apparatus shown in FIG.

In the apparatus described above, in the two-dimensional differential processing, differential abnormal values may occur depending on the shape and temperature distribution of the target steel plate.

Now, the method of two-dimensional differentiation will be explained by taking as an example a case where wall-to-wall differentiation processing is performed in which 8.times.8 pixels of the memory are considered as one unit. FIG. 7 shows an 8×8 pixel memory configuration in which each memory is labeled. Generally, the differential value E' of the median value E of this memory is expressed by the following formula.

E'=f (A+nB+C)~(G+nH+I))+
((A+nl)10G) to (C+nF+I)1 where n is a weighting coefficient for the pixel.

It can be seen that the differential value E' increases as the difference between pixels B and H increases when considered in the Y-axis direction.

Next, in order to consider the state in which the tip of the steel plate is stored in the memory 11, a part of the mesori is expressed in FIG. 8.

Scattered light exists at the edge of the red-hot steel plate indicated by diagonal lines in FIG. 8, making it difficult to distinguish between the edge of the steel plate and the space.
Temperature data is stored in memory. In this state, if we spatially differentiate this memory, as explained in Fig. 7, 8x3 strokes are required, so differential calculations are possible from the second line, and 1 Differential processing is not performed on the row, and the original data remains. Normally, when steel plate temperature data is stored from the second line in FIG. 8, the data in the first line is normally 0 (zero).

The differential results of FIG. 8 are shown in FIG. The Δ mark is the differential value between the 0 data in the first line of Figure 8 and the steel plate data in the eighth line of Figure 8, and the steel plate edge is emphasized and expressed similarly. The marks are the scattered light in the second row of Figure 9 and the eighth row.
This is the differential value with respect to the steel sheet data in the fourth row of the figure, and this also emphasizes the copper sheet edge. However, the area represented by the X mark is the scattered light on the 8th line of Figure 8.
This is a differential image with respect to the 0 data in the first row of FIG. 8. If the scattered light is higher than a certain level, the differential result between the two becomes a large value, and it is expressed as if it were emphasized as a steel plate edge.

A11B, NO, and 2 in Fig. 8, which are originally the ends of the steel plate.

The contours to E and C are 411 mouths 1°C and 1. 2', E1. Although it should be expressed as 1 (the line moves up one line by the differential process), the differential value of the X mark in FIG. 9 is high, so a. E 1° 1 is determined to be the steel plate outline. As a result, there was a drawback that the binarization accuracy deteriorated.

[Summary of the invention]

This invention was made in order to eliminate the above-mentioned drawbacks, and provides a shape detection device equipped with correction means before and after differential processing in order to correct differential abnormal values and obtain a highly accurate binary image of a steel plate. We are aiming to provide the following.

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to the drawings. In FIG. 10, a pre-differentiation correction processing circuit 21 and a post-differentiation correction processing circuit 22 are added before and after the differentiation circuit 12, respectively, and the other configurations are the same as in FIG. 8. Furthermore, a pre-differentiation correction processing circuit 21 and a post-differentiation correction processing circuit 22 may be added to the structure shown in FIG. 5 as well.

The details of this part will be explained below. Pre-differentiation correction processing circuit 2
1 indicates that after the signal quantized by the quantization circuit 10 is stored in the memory 11, it is stored in the memory 11 with the same value as the scattered light in the second row, as shown in FIG. 11 where the steel plate is shaded. Forcibly fills the memory in the first line. If the differential processing is performed in this state, the differential processing between the 0 data in the first row and the scattered light will be eliminated.

The results are shown in FIG. The ◎ mark is the differential value between the scattered light caught in the 1st line of Figure 11 and the steel plate data on the 08th line of Figure 1. Since the temperature difference between the two is high, the differential value becomes high, and the steel plate edge is expressing. The mark ■ is the differential value of the scattered light in the first line of Figure 11 and the scattered light in the third line of Figure 11, which were forcibly written, but the difference between the two is the scattered light. The differential value is low. Therefore, 411 1° C1 in FIG. 12 is surrounded by high differential values. 21. B, H1 can be extracted as lines corresponding to the steel plate contour.

When converting the final z-value into histogram processing based on the differential processing results, the scattered light that was forcibly written to the first line of the memory remains as it is without being differentiated, so the histogram It is necessary to make sure that it does not affect the processing as noise. For this purpose, after the differentiation process, the post-differentiation correction processing circuit 22 changes all the data in the first row to ○.
Change it to a mark. By doing so, correct binarization processing can be performed even in the histogram processing after the differentiation processing.

In the above embodiment, the explanation has been made using hardware such as a differentiating circuit, a correction processing circuit, a peak hold circuit, a binarization circuit, etc., but it may also be executed in software by a computer capable of processing the same functions. Furthermore, as an imaging means for the steel plate signal, a function of adjusting the amount of light received by various filters, an automatic multimeter υ machine, etc. may be added.

〔Effect of the invention〕

As described above, according to the present invention, a correction processing circuit is added before and after the differential processing in order to prevent differential abnormal values from occurring, so a highly accurate steel plate z-valued image can be extracted, and a highly accurate z-valued image can be extracted. This has the effect of providing a shape detection device that can secure a width signal and accurately output a comparison result with a prescribed width value.

[Brief explanation of drawings]

Figure 1 is a block diagram showing the configuration of the conventional device, Figure 2 is an explanatory diagram showing the shape of the tip of the steel plate to explain its operation,
FIG. 8 is a block diagram showing the configuration of the device that is the basis of this invention, FIG. 4 is a signal waveform diagram of each part of FIG. 3, and FIG. Block diagram, 6th
The figure is a signal waveform diagram of Figure 5, Figure 7 is a diagram of the principle of two-dimensional differential processing, Figure 8 is a diagram of the memory storage state before differentiation, Figure 9 is a diagram of the memory storage state after differentiation processing, and Figure 1O. FIG. 11 is a block diagram showing the configuration of an embodiment of the present invention. FIG. 12 is a memory state diagram before and after differentiation when the correction according to the present invention is applied. 1... Steel plate, 2... Lens, 6... Photoelectric element group,
4... Amplifier, 6... Width measurement circuit, 8... Comparison circuit, 9... Cutting machine control device, 10... Quantization circuit,
11...Memory, 12...Differentiating circuit, 16...Peak hold circuit, 14...z value conversion circuit, 15...
Scanning circuit, 21... Pre-differentiation correction processing circuit, 22...
Post-differentiation correction processing circuit. In addition, in the figures, the same reference numerals indicate the same or equivalent parts. Patent applicant Mitsubishi Gunki Co., Ltd. (2 others) Figure 1 a Figure 2 Figure 4 Figure 6 Figure 7 Figure 8 Figure 9

Claims (1)

[Claims] (11) A detection means for forming a steel plate image on a plurality of photoelectric element groups and generating a scanning electric signal corresponding to the temperature of the steel plate; A scanning means for controlling according to the traveling distance, a quantization means for multileveling the detection signal obtained by the detection means, and a two-dimensional quantization signal for converting the quantized signal stored in the memory means for storing the multilevel signal. a differentiating means for differentiating the output signal of the differentiating means; a width measuring means for licking the plate width of each scanning portion of the steel plate from a binarizing means and a peak hold means; and a comparison means for comparing and outputting the result, the shape detection device comprising a post-differentiation correction means for correcting a differential abnormal value after differentiation processing of the differentiation means. (2) The shape detection device according to claim 1, further comprising a pre-differentiation correction means for correcting the differential abnormal value before the differentiation processing of the differentiation means. a peak hold means for peak holding a signal; a binarization means for determining an a-degree histogram threshold level of the peak hold signal and binarizing the peak hold signal; and each scanning section of the steel plate from the binarized signal. 3. The shape detecting device according to claim 1, further comprising a width measuring means for measuring the plate width.