KR100290637B1 - Apparatus and method for accurately spraying powder into mold using wide-angle lens - Google Patents

Abstract

PURPOSE: An apparatus and a method for accurately spraying powder into a mold using wide-angle lens are provided to automatically spray powder onto a part where powder is insufficient by analyzing the surface of molten metal of continuous process using the wide-angle lens. CONSTITUTION: The apparatus for accurately spraying powder into a mold using wide-angle lens comprises a CCD camera(70) which is installed at the lower part of a support(60) positioned between the lower part of a tundish(10) and the upper part of a mold so that the CCD camera(70) is protected through a housing, and on which a wide-angle lens is mounted; an image taker(90) correcting distorted images of a wide-angle lens(80) after receiving image signals of the surface of molten metal analyzed by the wide-angle lens(80); an image processor(100) processing the image signals into an image in which a brightness of the surface of molten metal is strikingly displayed after receiving the image signals processed and transmitted by the image taker(90); and a computer(110) performing control administration by transmitting information on the position of a dark part where a powder injection amount is insufficient to a powder sprayer(50) so as to accurately spray powder after automatically analyzing the brightness of the surface of molten metal processed by the image processor(100).

Description

Mold powder precision spraying device and method using wide angle lens

The present invention is to analyze the difference between the contrast of the surface of the playing process using a wide-angle lens camera through the image and to recognize the surface of the lack of powder automatically by the difference in the contrast of the wide-angle lens that can spray as much powder It relates to a mold powder precision spreading apparatus using the same and a method thereof.

In general, the working environment of the continuous casting process in the steelmaking site, skilled workers reside in the mold powder inlet, the injection of the mold powder in the mold mold 40, the start of the powder spreader 50, the specimen collection operation, slug pool ( Pool depth, mold mold 40 wall slug bear generation and removal work, casting speed control, abnormal situation measures, etc. are performed.

However, the casting environment is a high noise due to the powder in the powder, the noise due to the vertical vibration of the mold mold 40 and the noise in the field, and the high noise due to the heat of the molten steel, such as poor working environment.

In order to improve the working environment, an automatic analysis system of the noodles is constructed.

Such a system has been mentioned in the patent application No. 89-17706, but it was not possible to determine the exact position of the part where the powder input is insufficient because it does not consider the distortion caused by the use of the wide-angle lens 80 in the CCD camera 70.

The present invention has been invented to solve the above problems and disadvantages, to correct the distortion caused by using a wide-angle lens to accurately grasp the position of the portion where the powder input is insufficient to accurately spray the wide-angle lens camera The purpose is to provide a mold powder precision spray system using.

The present invention having the above object distinguishes the state on the molten surface into a roll and a dark portion during actual use through the field of view of an image sensor manufactured to subdivide the molten surface portion into a set area, thereby determining the state of the molten surface according to the present invention. After detecting the abnormality of the hot water surface, it is determined that the mold powder is insufficient and the detected powder shortage and one or more of the melted surface abnormality conditions are used to control the casting speed, control the flow rate of the working gas in the injection nozzle, and the mold. One or two or more of the molten steel flow rate control and the powder spraying control to be injected in order to remove the above molten surface to produce a high quality molten steel, between the lower part of the tundish car and the upper part of the mold Mounts a wide-angle lens that is installed under the housing to be protected by the housing CCD camera, an image acquirer for receiving the image signal of the melt surface analyzed by the wide-angle lens, and correcting the distorted image of the wide-angle lens, and an image signal that is processed and transmitted by the image acquirer, and the brightness and contrast of the melt surface. The image processor and the image processor to process the image of the marked appearance, and automatically analyzes the light and dark state of the molten surface processed by the image processor to accurately recognize the position of the dark portion of the powder input is insufficient, and then to the dark position of the insufficient powder input It is characterized in that it comprises a computer for performing a control stroke to transmit the information to the powder spreader to precisely spray the powder.

1 is a perspective view showing a mold powder precision spraying apparatus using a wide-angle lens camera of the present invention.

2 is a block diagram illustrating the procedure of the distortion correction step of the wide-angle lens according to the present invention.

3 is a block diagram illustrating a procedure for finding a light center through the above steps.

4 is a block diagram illustrating a method for obtaining a function for converting the resolution in the vertical and horizontal directions to be equal to the resolution in the horizontal direction.

5 is a block diagram for explaining a step of obtaining a conversion function for correcting lens distortion of the present invention.

6 is a block diagram illustrating a procedure of performing a calibration on an input image after obtaining a calibration function of the present invention.

7 is a flowchart illustrating a method of analyzing an image of a mold melt surface of the present invention.

* Explanation of symbols for main parts of the drawings

10: tundish car 20: cassette

30: immersion nozzle 40: mold mold

50: powder spreader 60: support

70: CCD camera 80: wide-angle lens

90: image acquirer 100: image processor

110: computer 120: camera housing

130: left floor monitor 140: right floor monitor

According to the present invention, a CCD having a wide-angle lens 80 having a camera housing 120 for protecting the tundish from high temperatures and dust in the workplace is provided between the lower portion of the tundish car 10 and the upper portion of the mold mold 40. The camera 70 is installed, which is supported by the support 60, and has an image acquirer 90 for acquiring an image of the water surface and an image processor 100 equipped with a DSP for processing image processing in real time. And a computer 110 as a terminal device.

Referring to the operation of the present invention configured as described above is as follows. In order to secure a wide field of view such as a human visual system, a CCD camera 70 equipped with a wide-angle lens 80 may be used. This is because a standard lens is used for high accuracy, but the visible area is not so large.

The use of the wide-angle lens 80 has the advantage of ensuring a wide field of view while the central portion of the lens has a disadvantage that the resolution is high and gradually decrease toward the outer portion of the lens.

This resolution is remarkable enough to be seen by the eye, and it is necessary to convert the multivariate resolution into a uniform resolution by eliminating the distorted image.

First, the image obtained by using the wide-angle lens 80 has a lower resolution as the distance from the optical center, that is, the radius increases. Looking closely, the actual distance should appear as a linear straight line as the radius changes. Therefore, it is similar to the nonlinear exponential function.

In other words, when converting the radius r 'of the distorted coordinate system to r of the coordinate system without distortion, r = α e ^{(β r')} may be represented and expressed as a series as follows.

r = α (1 + β r '+ β ₂ r' ² + β ₃ r ' ³ + β ₄ r' ⁴ +...

Where α and β all correspond to the coefficient of r '

r = g ₀ + g ₁ r '+ g ₂ r' ² + g ₃ r ' ³ + g ₄ r' ⁴ +... …

Since the angle has little distortion, the angle θ 'at the multilateral resolution can be replaced with the angle θ at the uniform resolution. In other words

θ = θ ′

In this way, the multi-sided resolution image of the camera equipped with the wide-angle lens can be transferred to the uniform resolution image and can be easily corrected to the real coordinate system in the image coordinate system.

The method for obtaining the coefficient of such a transform will be described in detail as follows.

First of all, prior to the conversion, the image obtained by the CCD camera 70 equipped with the wide-angle lens 80 is converted into the input image, and the image to be converted to the uniform resolution is defined as the output image, and the rectangular coordinate system of the input image is defined as (x ″). , y ″), the spherical coordinate system is defined as (r ″, θ ″), and the rectangular coordinate system of the output image is defined as (x, y).

First of all, to find the transformation coefficient, first find the optical center where the optical axis meets the focal plane and find out the origin of the spherical coordinate system.

The optical center can be obtained in the following way.

First, in order to obtain the optical center, the flat plate should be placed in front of the CCD camera 70 and the CCD camera 70, and the flat plate should be placed vertically with the CCD camera 70 to obtain the optical center with high accuracy. The correction error of the distorted image is also reduced as much as it is obtained.

A picture drawn with several concentric circles is fixed on the lattice plate on the plate on which the lattice is drawn in the vertical and horizontal directions at uniform intervals, and then the image is acquired by the CCD camera 70.

At this time, the center point of the concentric circles is fixed to coincide with any intersection point of the grid. The center point of each concentric circle is obtained and the CCD camera 70 is moved up, down, left and right little by little so as to coincide with the center point.

If the center points of the concentric circles coincide, the camera and the flat plate are assumed to be vertical, and the center point at that time becomes the light center.

In the same way as above, the optical center can be obtained with high precision. The following describes how to obtain the conversion coefficient.

After the alignment of the flat panel with the CCD camera 70 is completed, the image is acquired. Then, the intersection point of the horizontal and vertical grids is selected as the input point in the input image to obtain the rectangular coordinate values (x, y) of the input points. Find the spherical coordinates (r, θ) as the origin.

That is, r ″ = sqrt (x ″ ² + y ″ ² ), θ ″ = arc tan (y ″ / x ″)

The radius of the intersection between the grid in the vertical direction relative to the optical center to the horizontal resolution of the direction and the vertical direction of the image is unlikely because the first to correct the discrepancy resolution d _{v1, v2} d, d ‥‥‥‥ _vN and horizontal Obtain the radii d _h1 , d _h2 , ‥‥‥ d _hN of the lattice intersection points in the direction, and use the following matrix equation to find the coefficients of the calibration function that make the vertical resolution equal to the horizontal resolution. dv).

d _h = C (d _v ) = c ₀ + c ₁ d _v + c ₂ d _v ² + c ₃ d _v ³ + c ₄ d _v ⁴

Using the C (d _v ) obtained in this way, all coordinate values of the input point are transformed to define the converted input points as input points.

This completes the description of the input point. Next, the description of the output point remains. The description of the output point is described in the vertical and horizontal uniform intervals based on the optical center.

As described above, when the input point and the output point are described, the radius of the spherical coordinate system r ₁ ′, r ₂ ′ ‥‥‥‥‥ r _N ′ and the radius of the spherical coordinate system r ₁ , r ₂ ‥‥‥ For r _N , the transform function is expressed in matrix form:

θ = θ ′

When the matrix form is expressed as R '= RG, the matrix G can be obtained as follows.

G = (R ^T R) ^-1 R ^T R ′

The conversion coefficients g ₀ , g ₁ , g ₂ , g ₃ , g ₄ were obtained.

r = G (r ′) = g ₀ + g ₁ r ′ + g ₂ r ′ ² + g ₃ r ′ ³ + g ₄ r ′ ⁴

Is determined.

Thus, the function of distortion correction can be obtained, and these two functions can be applied to each input image to obtain the output image correcting the distortion.

The process of correcting distortion using the correction function is as follows. The following process is repeated for each pixel of the input image.

A) A new rectangular coordinate value (x ', y') is obtained by applying a function of correcting the rectangular coordinate value (x ″, y ″) of the pixel in the input image so that the vertical resolution is equal to the horizontal resolution.

B) The rectangular coordinate values (x ', y') of the pixel obtained in (a) are converted into spherical coordinate values (r ', θ') having the optical center as the origin.

C) Spherical coordinate of the output image using the conversion function r = G (r ′) = g ₀ + g ₁ r ′ + g ₂ r ′ ² + g ₃ r ′ ³ + g ₄ r ′ ⁴ , θ = θ ′ Obtain the values (r, θ).

D) The spherical coordinate value (r, θ) of the output image is converted to the rectangular coordinate value (x, y).

E) The content of pixels existing between the rectangular coordinates (x, y) of the input image and between the rectangular coordinates (x ″, y ″) of the input image is replaced.

The action of the system configuration is that the image of the left and right sides is first sent to each of the left and right monitors 130 and 140 so that the operator visually checks each of the images of the surface, which is connected to the image acquirer 90. do.

The system first corrects the distortion of the image by using the wide-angle lens 80 and undergoes binarization for fast processing of the image and reduction of information.

Next, the screen of the analysis target is divided from the area 0 to the area N, and the histogram and the histogram ratio of each area are calculated and stored in the memory. When the size of the frame is smaller than the set size, the above operation is repeated. The histogram is larger than the threshold and the area of the histogram smaller than the threshold is determined to be a lack of powder in the bath surface, and the powder spreader 50 transmits information on the lack of powder to spray the powder. In Fig. 7, the above algorithm is shown in a flowchart. This operation is performed alternately with the image of the left tap surface, and automatic analysis of the tap surface is performed.

The present invention as described above has the effect of producing a good quality product by correcting the precise location of the portion of the insufficient powder input by correcting the distortion caused by using a wide-angle lens.

Claims (2)

After visually identifying the state of the melted surface by detecting the state of the melted surface according to the roster and the dark part through the field of view of the image sensor designed to subdivide the melted surface into the set area, Determination of the lack of mold powder and the detection of powder shortage and the occurrence of one or more melt surface abnormalities in the occurrence position, controlling the casting speed, controlling the flow rate of working gas in the injection nozzle, controlling the flow rate of the molten steel and the powder injected into the mold. In any one or more of the spraying control to eliminate the above molten surface to produce a high quality molten steel, the support is located between the lower portion of the tundish car 10 and the upper portion of the mold ( A CCD camera 70 equipped with a wide-angle lens 80 installed under the housing 60 so as to be protected by the housing 120; The image acquirer 90 receives the image signal of the molten surface analyzed by the wide angle lens 80 and corrects the distorted image of the wide angle lens 80, and the image acquirer 90 processes the image signal. The image processor 100 receives the image signal and processes the image of the melt surface of the melt surface, and the intensity of the powder is automatically analyzed by analyzing the contrast state of the melt surface processed by the image processor 100. It is configured to include a computer 110 that performs a control stroke so that after accurately recognizing the insufficient position of the dark portion, the information about the position of the insufficient portion of the powder is delivered to the powder spreader 50 to precisely spread the powder Mold powder precision spraying device using a wide-angle lens. A new image is obtained by applying a function of correcting the rectangular coordinate values (x ″, y ″) and the vertical resolution of the pixel in the input image to be equal to the horizontal resolution by using the image acquirer 90 of the structure of claim 1. Obtaining rectangular coordinate values (x ', y'); Converting rectangular coordinate values (x ', y') into spherical coordinate values (r ', θ') with the optical center as the origin; Conversion function r = G (r ′) = g ₀ + g ₁ r ′ + g ₂ r ′ ² + g ₃ r ′ ³ , θ = θ ′ Obtaining a; Converting a spherical coordinate value (r, θ) of the output image into a rectangular coordinate value (x, y); Replacing the content of the pixel present in the rectangular coordinate values (x ″, y ″) of the input image with the rectangular coordinate values (x, y) of the output image; taking a transform to correct the distorted image of the wide-angle lens 80 A method for correcting a distorted image input by a wide-angle lens, characterized by obtaining a function.